From 5c9084e9783c28f121b0b4217bd49fd81aca2b23 Mon Sep 17 00:00:00 2001 From: tejass Date: Thu, 25 Jun 2026 22:12:44 +0200 Subject: [PATCH 01/10] Organizing GUI task wise --- app.py | 374 +------------- examples/tutorial_image_segmentation.ipynb | 2 +- perceptionmetrics/models/torch_detection.py | 5 +- tabs/dataset_viewer.py | 313 +----------- tabs/evaluator.py | 502 +------------------ tabs/inference.py | 151 +----- tabs/tasks/__init__.py | 1 + tabs/tasks/image_detection/__init__.py | 26 + tabs/tasks/image_detection/dataset_viewer.py | 288 +++++++++++ tabs/tasks/image_detection/evaluator.py | 447 +++++++++++++++++ tabs/tasks/image_detection/inference.py | 132 +++++ tabs/tasks/image_detection/sidebar.py | 329 ++++++++++++ 12 files changed, 1264 insertions(+), 1306 deletions(-) create mode 100644 tabs/tasks/__init__.py create mode 100644 tabs/tasks/image_detection/__init__.py create mode 100644 tabs/tasks/image_detection/dataset_viewer.py create mode 100644 tabs/tasks/image_detection/evaluator.py create mode 100644 tabs/tasks/image_detection/inference.py create mode 100644 tabs/tasks/image_detection/sidebar.py diff --git a/app.py b/app.py index c0ec8dfe..cbdbf3c9 100644 --- a/app.py +++ b/app.py @@ -1,13 +1,10 @@ import streamlit as st + +from perceptionmetrics.utils.torch import get_device_info from tabs.dataset_viewer import dataset_viewer_tab -from tabs.inference import inference_tab from tabs.evaluator import evaluator_tab -from perceptionmetrics.utils.gui import browse_folder -from perceptionmetrics.utils.torch import get_device_info - - -def browse_dataset_path(): - st.session_state.dataset_path = browse_folder() +from tabs.inference import inference_tab +from tabs.tasks.image_detection import render_image_detection_sidebar st.set_page_config(page_title="PerceptionMetrics", layout="wide") @@ -20,368 +17,37 @@ def browse_dataset_path(): best_device, available_devices = get_device_info() -# Initialize commonly used session state keys +# Shared state +st.session_state.setdefault("task", "Image Detection") st.session_state.setdefault("dataset_path", "") -st.session_state.setdefault("dataset_type", "YOLO") st.session_state.setdefault("split", "test") +st.session_state.setdefault("device", best_device) + +# Image detection state +st.session_state.setdefault("dataset_type", "YOLO") st.session_state.setdefault("config_option", "Manual Configuration") st.session_state.setdefault("confidence_threshold", 0.5) st.session_state.setdefault("nms_threshold", 0.5) st.session_state.setdefault("max_detections", 100) -st.session_state.setdefault("device", best_device) st.session_state.setdefault("batch_size", 1) st.session_state.setdefault("evaluation_step", 5) st.session_state.setdefault("detection_model", None) st.session_state.setdefault("detection_model_loaded", False) -# Sidebar: Dataset Inputs with st.sidebar: - with st.expander("Dataset Inputs", expanded=True): - # First row: Type and Split - col1, col2 = st.columns(2) - with col1: - st.selectbox( - "Type", - ["COCO", "YOLO"], - key="dataset_type", - ) - with col2: - st.selectbox( - "Split", - ["train", "val", "test"], - key="split", - ) - - # Second row: Path and Browse button - col1, col2 = st.columns([3, 1]) - with col1: - st.text_input("Dataset Folder", key="dataset_path") - with col2: - st.markdown( - "
", unsafe_allow_html=True - ) - st.button("Browse", on_click=browse_dataset_path) - - # Additional input for YOLO config file - if st.session_state.get("dataset_type", "COCO") == "YOLO": - st.file_uploader( - "Dataset Configuration (.yaml)", - type=["yaml"], - key="dataset_config_file", - help="Upload a YAML dataset configuration file.", - ) - - with st.expander("Model Inputs", expanded=False): - st.file_uploader( - "Model File (.pt, .onnx, .h5, .pb, .pth, .torchscript)", - type=["pt", "onnx", "h5", "pb", "pth", "torchscript"], - key="model_file", - help="Upload your trained model file.", - max_upload_size=1024, # MB - ) - st.file_uploader( - "Ontology File (.json)", - type=["json"], - key="ontology_file", - help="Upload a JSON file with class labels.", - ) - st.radio( - "Configuration Method:", - ["Manual Configuration", "Upload Config File"], - key="config_option", - horizontal=True, - ) - if ( - st.session_state.get("config_option", "Manual Configuration") - == "Upload Config File" - ): - st.file_uploader( - "Configuration File (.json)", - type=["json"], - key="config_file", - help="Upload a JSON configuration file.", - ) - else: - col1, col2 = st.columns(2) - with col1: - st.slider( - "Confidence Threshold", - min_value=0.0, - max_value=1.0, - step=0.01, - key="confidence_threshold", - help="Minimum confidence score for detections", - ) - st.slider( - "NMS Threshold", - min_value=0.0, - max_value=1.0, - step=0.01, - key="nms_threshold", - help="Non-maximum suppression threshold", - ) - st.number_input( - "Max Detections/Image", - min_value=1, - max_value=1000, - step=1, - key="max_detections", - ) - with col2: - st.selectbox( - "Device", - available_devices, - key="device", - ) - st.selectbox( - "Model Format", - ["torchvision", "YOLO"], - index=( - 0 - if st.session_state.get("model_format", "torchvision") - == "torchvision" - else 1 - ), - key="model_format", - ) - st.number_input( - "Batch Size", - min_value=1, - max_value=256, - step=1, - key="batch_size", - ) - st.number_input( - "Evaluation Step", - min_value=0, - max_value=1000, - step=1, - key="evaluation_step", - help="Update UI with intermediate metrics every N images (0 = disable intermediate updates)", - ) - - st.write("---") - st.write("**Image Size Configuration**") - - # Resize Logic - enable_resize = st.checkbox( - "Enable Resize", value=True, key="enable_resize" - ) - - if enable_resize: - resize_strategy = st.radio( - "Resize Strategy", - ["Fixed Dimensions", "Min Side"], - key="resize_strategy", - horizontal=True, - label_visibility="collapsed", - ) - - if resize_strategy == "Fixed Dimensions": - c1, c2 = st.columns(2) - with c1: - st.number_input( - "Image Resize Height", - min_value=1, - max_value=4096, - value=640, - step=1, - key="resize_height", - help="Height to resize images for inference", - ) - with c2: - st.number_input( - "Image Resize Width", - min_value=1, - max_value=4096, - value=640, - step=1, - key="resize_width", - help="Width to resize images for inference", - ) - else: - st.number_input( - "Min Side", - min_value=1, - max_value=4096, - value=640, - step=1, - key="min_side", - help="Minimum size of the shorter side of the image", - ) - - # Crop Logic - enable_crop = st.checkbox("Enable Center Crop", key="enable_crop") - - if enable_crop: - c1, c2 = st.columns(2) - with c1: - st.number_input( - "Crop Height", - min_value=1, - max_value=4096, - value=640, - step=1, - key="crop_height", - help="Center crop height", - ) - with c2: - st.number_input( - "Crop Width", - min_value=1, - max_value=4096, - value=640, - step=1, - key="crop_width", - help="Center crop width", - ) - - # Load model action in sidebar - from perceptionmetrics.models.torch_detection import TorchImageDetectionModel - import json, tempfile - - load_model_btn = st.button( - "Load Model", - type="primary", - width="stretch", - help="Load and save the model for use in the Inference tab", - key="sidebar_load_model_btn", - ) - - if load_model_btn: - model_file = st.session_state.get("model_file") - ontology_file = st.session_state.get("ontology_file") - config_option = st.session_state.get( - "config_option", "Manual Configuration" - ) - config_file = ( - st.session_state.get("config_file") - if config_option == "Upload Config File" - else None - ) - - # Prepare configuration - config_data = None - config_path = None - try: - if config_option == "Upload Config File": - if config_file is not None: - config_data = json.load(config_file) - with tempfile.NamedTemporaryFile( - delete=False, suffix=".json", mode="w" - ) as tmp_cfg: - json.dump(config_data, tmp_cfg) - config_path = tmp_cfg.name - else: - st.error("Please upload a configuration file") - else: - confidence_threshold = float( - st.session_state.get("confidence_threshold", 0.5) - ) - nms_threshold = float(st.session_state.get("nms_threshold", 0.5)) - max_detections = int(st.session_state.get("max_detections", 100)) - device = st.session_state.get("device", "cpu") - batch_size = int(st.session_state.get("batch_size", 1)) - evaluation_step = int(st.session_state.get("evaluation_step", 5)) - model_format = st.session_state.get("model_format", "torchvision") - - # Resize Logic extraction - enable_resize = st.session_state.get("enable_resize", True) - resize_cfg = None - if enable_resize: - resize_strategy = st.session_state.get( - "resize_strategy", "Fixed Dimensions" - ) - if resize_strategy == "Fixed Dimensions": - resize_height = int( - st.session_state.get("resize_height", 640) - ) - resize_width = int( - st.session_state.get("resize_width", 640) - ) - resize_cfg = { - "height": resize_height, - "width": resize_width, - } - else: - min_side = int(st.session_state.get("min_side", 640)) - resize_cfg = {"min_side": min_side} - - config_data = { - "confidence_threshold": confidence_threshold, - "nms_threshold": nms_threshold, - "max_detections_per_image": max_detections, - "device": device, - "batch_size": batch_size, - "evaluation_step": evaluation_step, - "model_format": model_format.lower(), - } - if resize_cfg is not None: - config_data["resize"] = resize_cfg - - if enable_crop: - crop_height = int(st.session_state.get("crop_height", 640)) - crop_width = int(st.session_state.get("crop_width", 640)) - crop_cfg = {"height": crop_height, "width": crop_width} - config_data["crop"] = crop_cfg - - with tempfile.NamedTemporaryFile( - delete=False, suffix=".json", mode="w" - ) as tmp_cfg: - json.dump(config_data, tmp_cfg) - config_path = tmp_cfg.name - except Exception as e: - st.error(f"Failed to prepare configuration: {e}") - config_path = None - - if model_file is None: - st.error("Please upload a model file") - elif config_path is None: - st.error("Please provide a valid model configuration") - elif ontology_file is None: - st.error("Please upload an ontology file") - else: - with st.spinner("Loading model..."): - # Persist ontology to temp file - try: - ontology_data = json.load(ontology_file) - with tempfile.NamedTemporaryFile( - delete=False, suffix=".json", mode="w" - ) as tmp_ont: - json.dump(ontology_data, tmp_ont) - ontology_path = tmp_ont.name - except Exception as e: - st.error(f"Failed to load ontology: {e}") - ontology_path = None + task = st.selectbox( + "Task", + ["Image Detection"], + key="task", + help="Only image detection is currently wired in the GUI.", + ) - # Persist model to temp file - try: - with tempfile.NamedTemporaryFile( - delete=False, suffix=".pt", mode="wb" - ) as tmp_model: - tmp_model.write(model_file.read()) - model_temp_path = tmp_model.name - except Exception as e: - st.error(f"Failed to save model file: {e}") - model_temp_path = None + if task == "Image Detection": + render_image_detection_sidebar(available_devices) + else: + st.error(f"Unsupported task: {task}") - if ontology_path and model_temp_path: - try: - model = TorchImageDetectionModel( - model=model_temp_path, - model_cfg=config_path, - ontology_fname=ontology_path, - device=st.session_state.get("device", "cpu"), - ) - st.session_state.detection_model = model - st.session_state.detection_model_loaded = True - st.success("Model loaded and saved for inference") - except Exception as e: - st.session_state.detection_model = None - st.session_state.detection_model_loaded = False - st.error(f"Failed to load model: {e}") -# Main content area with horizontal tabs tab1, tab2, tab3 = st.tabs(["Dataset Viewer", "Inference", "Evaluator"]) with tab1: diff --git a/examples/tutorial_image_segmentation.ipynb b/examples/tutorial_image_segmentation.ipynb index 6de4877b..a31eadf3 100644 --- a/examples/tutorial_image_segmentation.ipynb +++ b/examples/tutorial_image_segmentation.ipynb @@ -1135,7 +1135,7 @@ "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", - "version": "3.10.19" + "version": "3.11.14" } }, "nbformat": 4, diff --git a/perceptionmetrics/models/torch_detection.py b/perceptionmetrics/models/torch_detection.py index 52efd705..96e71fe4 100644 --- a/perceptionmetrics/models/torch_detection.py +++ b/perceptionmetrics/models/torch_detection.py @@ -9,7 +9,10 @@ import torch from torch.utils.data import DataLoader, Dataset from torchvision.transforms import v2 as transforms -from torchvision import tv_tensors +try: + from torchvision import tv_tensors +except ImportError: + from torchvision import datapoints as tv_tensors from tqdm.auto import tqdm from perceptionmetrics.datasets import detection as detection_dataset diff --git a/tabs/dataset_viewer.py b/tabs/dataset_viewer.py index 85077129..c24cc89e 100644 --- a/tabs/dataset_viewer.py +++ b/tabs/dataset_viewer.py @@ -1,316 +1,13 @@ import streamlit as st -import os -from streamlit_image_select import image_select - -from perceptionmetrics.datasets.coco import find_img_dir_and_ann_file def dataset_viewer_tab(): - import tempfile - from perceptionmetrics.datasets.coco import CocoDataset - from perceptionmetrics.datasets.yolo import YOLODataset - import numpy as np - from PIL import Image - from supervision.draw.color import ColorPalette - from supervision.detection.annotate import BoxAnnotator - from supervision.detection.core import Detections - - # Get inputs from session state - dataset_path = st.session_state.get("dataset_path", "") - dataset_type = st.session_state.get("dataset_type", "COCO").lower() - split = st.session_state.get("split", "val") - - # Header row only - st.header("Dataset Viewer") - - if not dataset_path or not os.path.isdir(dataset_path): - st.warning("⚠️ Please select a valid dataset folder.") - return - - # Setup paths and pagination - if dataset_type == "coco": - try: - img_dir, ann_file = find_img_dir_and_ann_file( - dataset_path=dataset_path, split=split - ) - except FileNotFoundError: - st.warning("Dataset files not found. Check path and split.") - return - - elif dataset_type == "yolo": - dataset_config_file = st.session_state.get("dataset_config_file", None) - img_dir = os.path.join(dataset_path, f"images/{split}") - if not os.path.isdir(img_dir): - st.warning("Image directory not found.") - return - if dataset_config_file is None: - st.warning("Dataset configuration file not found. Please upload it.") - return - else: - st.error("Unsupported dataset type.") - return - - # Pagination and search row - nav_col1, nav_col2, nav_col3, nav_col4 = st.columns([1, 1, 2, 1.5]) - with nav_col1: - pass # Placeholder for "< page" button, to be added later in the code - with nav_col2: - pass # Placeholder for ">" button, to be added later in the code - with nav_col3: - pass # Placeholder for page info, to be added later in the code - with nav_col4: - # Move the button up by reducing the margin and decrease button size with custom CSS - st.markdown( - """ - - """, - unsafe_allow_html=True, - ) - st.markdown("
", unsafe_allow_html=True) - - # Load dataset - dataset_key = f"{dataset_path}_{split}" - if dataset_key not in st.session_state: - try: - if dataset_type == "coco": - st.session_state[dataset_key] = CocoDataset( - annotation_file=ann_file, - image_dir=img_dir, - split=split, - ) - elif dataset_type == "yolo": - if dataset_config_file is not None: - # Save uploaded config file to a temporary location - with tempfile.NamedTemporaryFile( - delete=False, suffix=".yaml" - ) as tmp: - tmp.write(dataset_config_file.read()) - tmp_path = tmp.name - - # Load YOLO dataset - yolo_dataset = YOLODataset(tmp_path, dataset_path) - st.session_state["full_dataset_df"] = yolo_dataset.dataset + task = st.session_state.get("task", "Image Detection") - # Filter dataset for the selected split - yolo_dataset.dataset = yolo_dataset.dataset[ - yolo_dataset.dataset["split"] == split - ].reset_index(drop=True) - st.session_state[dataset_key] = yolo_dataset + if task == "Image Detection": + from tabs.tasks.image_detection import render_image_detection_viewer - os.unlink(tmp_path) # Clean up temp file - else: - st.warning( - "Dataset configuration file not found. Please upload it." - ) - return - else: - st.error("Unsupported dataset type.") - return - - except Exception as e: - st.error(f"Failed to load dataset: {e}") - return - else: - # Ensure cached dataset has the correct split; if not, rebuild it - try: - cached_ds = st.session_state[dataset_key] - cached_split = getattr(cached_ds, "split", None) - if cached_split != split: - if dataset_type == "coco": - st.session_state[dataset_key] = CocoDataset( - annotation_file=ann_file, - image_dir=img_dir, - split=split, - ) - elif dataset_type == "yolo": - yolo_dataset = st.session_state[dataset_key] - yolo_dataset.dataset = st.session_state["full_dataset_df"][ - st.session_state["full_dataset_df"]["split"] == split - ].reset_index(drop=True) - st.session_state[dataset_key] = yolo_dataset - else: - st.error("Unsupported dataset type.") - return - except Exception: - pass - dataset = st.session_state[dataset_key] - - # Get image files - image_files = [ - f for f in os.listdir(img_dir) if f.lower().endswith((".jpg", ".jpeg", ".png")) - ] - if not image_files: - st.warning("No images found.") + render_image_detection_viewer() return - # Pagination - IMAGES_PER_PAGE = 12 - _, total_pages = ( - len(image_files), - (len(image_files) + IMAGES_PER_PAGE - 1) // IMAGES_PER_PAGE, - ) - page_key = f"image_page_{dataset_path}_{split}" - - if page_key not in st.session_state: - st.session_state[page_key] = 0 - current_page = max(0, min(st.session_state[page_key], total_pages - 1)) - st.session_state[page_key] = current_page - - start_idx = current_page * IMAGES_PER_PAGE - sample_images = image_files[start_idx : start_idx + IMAGES_PER_PAGE] - image_paths = [os.path.join(img_dir, img_name) for img_name in sample_images] - - # Navigation - col1, col2, col3, col4 = st.columns([0.5, 9.5, 0.5, 0.5]) - with col1: - if st.button("⟨", key="prev_page_btn", disabled=(current_page == 0)): - st.session_state[page_key] = current_page - 1 - st.rerun() - with col2: - st.markdown( - f"
Page {current_page + 1} of {total_pages}
", - unsafe_allow_html=True, - ) - with col3: - if st.button( - "⟩", key="next_page_btn", disabled=(current_page >= total_pages - 1) - ): - st.session_state[page_key] = current_page + 1 - st.rerun() - with col4: - if st.button( - "🔍", - key="search_icon_btn", - help="Search for an image by name", - disabled=not (dataset_path and os.path.isdir(dataset_path)), - ): - st.session_state["show_search_dropdown"] = True - - # Search dropdown - if st.session_state.get("show_search_dropdown", False): - col1, col2, col3 = st.columns([4, 1, 1]) - with col1: - selected_img = st.selectbox( - "Search image:", options=image_files, key="search_image" - ) - with col2: - st.markdown( - "
", unsafe_allow_html=True - ) - if st.button("Go to image", key="go_to_image_btn"): - new_page = image_files.index(selected_img) // IMAGES_PER_PAGE - st.session_state[page_key] = new_page - st.session_state[ - f"img_select_all_{dataset_path}_{split}_{new_page}" - ] = (image_files.index(selected_img) % IMAGES_PER_PAGE) - st.session_state["show_search_dropdown"] = False - st.rerun() - with col3: - st.markdown( - "
", unsafe_allow_html=True - ) - if st.button("Cancel", key="cancel_search_btn"): - st.session_state["show_search_dropdown"] = False - st.rerun() - - caption_len_limit = 17 - captions = [ - ( - (name[:caption_len_limit] + "..." + name[-3:]) - if len(name) > caption_len_limit - else name - ) - for name in sample_images - ] - - # Image grid - img_select_key = f"img_select_all_{dataset_path}_{split}_{current_page}" - img_select_index = st.session_state.get(img_select_key) - if img_select_index is None or not isinstance(img_select_index, int): - img_select_index = 0 - selected_img_path = ( - image_select( - label="", - images=image_paths, - captions=captions, - use_container_width=False, - key=img_select_key, - index=img_select_index, - ) - if image_paths - else None - ) - - # Display selected image with annotations - if selected_img_path: - selected_img_name = os.path.basename(selected_img_path) - try: - img = Image.open(selected_img_path).convert("RGB") - img_np = np.array(img) - - if dataset_type == "yolo": - ann_row = dataset.dataset[ - dataset.dataset["image"].str.endswith(selected_img_name) - ] - else: - ann_row = dataset.dataset[dataset.dataset["image"] == selected_img_name] - - if not ann_row.empty: - annotation_id = ann_row.iloc[0]["annotation"] - if dataset_type == "yolo": - annotation_id = os.path.join(dataset.dataset_dir, annotation_id) - - boxes, category_indices = dataset.read_annotation(annotation_id) - - # Get class names from ontology - ontology = getattr(dataset, "ontology", None) - if ontology is None and hasattr(dataset.dataset, "attrs"): - ontology = dataset.dataset.attrs.get("ontology", None) - - if ontology: - catid_to_name = {v["idx"]: k for k, v in ontology.items()} - class_names = [ - catid_to_name.get(cat_id, str(cat_id)) - for cat_id in category_indices - ] - else: - class_names = [str(cat_id) for cat_id in category_indices] - - # Annotate image - palette = ColorPalette.default() - detections = Detections( - xyxy=np.array(boxes), class_id=np.array(category_indices) - ) - annotator = BoxAnnotator( - color=palette, text_scale=0.7, text_thickness=1, text_padding=2 - ) - annotated_img = annotator.annotate( - scene=img_np, detections=detections, labels=class_names - ) - - # Resize for display - annotated_pil = Image.fromarray(annotated_img) - try: - resample = getattr(Image, "Resampling", Image).LANCZOS - except AttributeError: - resample = Image.LANCZOS - annotated_pil.thumbnail((500, 500), resample) - st.image(annotated_pil, width="content") - else: - st.warning("No annotation found for this image.") - except Exception as e: - st.error(f"Error displaying image: {e}") - else: - st.info("Select an image to view with annotations.") + st.error(f"Unsupported task for dataset viewer: {task}") diff --git a/tabs/evaluator.py b/tabs/evaluator.py index 2e5fcac8..bb6da0e0 100644 --- a/tabs/evaluator.py +++ b/tabs/evaluator.py @@ -1,505 +1,13 @@ import streamlit as st -import os -import tempfile -import json -from perceptionmetrics.datasets.coco import CocoDataset - - -from perceptionmetrics.utils.gui import browse_folder -from perceptionmetrics.datasets.coco import find_img_dir_and_ann_file - - -def browse_predictions_outdir(): - folder = browse_folder() - if folder: - st.session_state.predictions_outdir = folder def evaluator_tab(): - st.header("Evaluator") - st.markdown("Evaluate your model on the loaded dataset using PerceptionMetrics.") - - # Check if we have the required objects from sidebar inputs - dataset_available = False - model_available = False - dataset = None - model = None - - # Check for dataset from sidebar inputs - dataset_path = st.session_state.get("dataset_path", "") - dataset_type = st.session_state.get("dataset_type", "Coco") - split = st.session_state.get("split", "val") - - # Try to get existing dataset from session state first - dataset_key = f"{dataset_path}_{split}" - if dataset_key in st.session_state: - dataset = st.session_state[dataset_key] - dataset_available = True - st.success( - f"✅ Dataset loaded: {dataset_path} ({split} split) - {len(dataset.dataset)} samples" - ) - elif dataset_path and os.path.isdir(dataset_path): - try: - if dataset_type.lower() == "coco": - img_dir, ann_file = find_img_dir_and_ann_file( - dataset_path=dataset_path, split=split - ) - - if os.path.isdir(img_dir) and os.path.isfile(ann_file): - st.session_state[dataset_key] = CocoDataset( - annotation_file=ann_file, image_dir=img_dir, split=split - ) - # Make filenames global - this is crucial for evaluation - st.session_state[dataset_key].make_fname_global() - dataset = st.session_state[dataset_key] - dataset_available = True - st.success( - f"✅ Dataset loaded: {dataset_path} ({split} split) - {len(dataset.dataset)} samples" - ) - else: - st.warning( - "⚠️ Dataset files not found. Please check the dataset path and split in the sidebar." - ) - else: - st.warning( - "⚠️ Only COCO datasets are currently supported for evaluation." - ) - except Exception as e: - st.error(f"❌ Error loading dataset: {e}") - else: - st.warning( - "⚠️ No dataset path provided. Please set the dataset path in the sidebar." - ) - - # Check for model from sidebar (loaded via Load Model button) - if ( - "detection_model" in st.session_state - and st.session_state.detection_model is not None - ): - model = st.session_state.detection_model - model_available = True - st.success("✅ Model loaded and ready for evaluation") - else: - st.warning( - "⚠️ No model loaded. Please load a model using the 'Load Model' button in the sidebar." - ) - - # Evaluation configuration - st.markdown("### Evaluation Configuration") - - save_predictions = st.checkbox( - "Save Predictions", - value=False, - help="Save individual predictions and metrics per sample", - ) - - save_visualizations = st.checkbox( - "Save Visualizations", - value=False, - help="Save visualized qualitative results (Image + GT + Preds)", - ) - - predictions_outdir_input = None - if save_predictions or save_visualizations: - col1, col2 = st.columns([3, 1]) - with col1: - st.text_input("Predictions Output Directory", key="predictions_outdir") - with col2: - st.markdown( - "
", unsafe_allow_html=True - ) - st.button( - "Browse", on_click=browse_predictions_outdir, key="browse_preds_outdir" - ) - - predictions_outdir_input = st.session_state.get("predictions_outdir") - - ontology_translation = st.file_uploader( - "Ontology Translation (Optional)", - type=["json"], - help="JSON file for translating between dataset and model ontologies", - ) - - # Disable Run Evaluation button if output dir is missing - output_dir_required = save_predictions or save_visualizations - output_dir_missing = output_dir_required and not ( - predictions_outdir_input and predictions_outdir_input.strip() - ) - - if output_dir_missing: - st.warning( - "⚠️ Please provide a Predictions Output Directory to enable evaluation " - "when 'Save Predictions' or 'Save Visualizations' is turned on." - ) - - # Run evaluation button - if st.button( - "🚀 Run Evaluation", - type="primary", - disabled=not (dataset_available and model_available) or output_dir_missing, - ): - if not dataset_available or not model_available: - st.error( - "Please ensure both dataset and model are loaded before running evaluation." - ) - return - - # Prepare evaluation - with st.spinner("Running evaluation..."): - try: - # Validate dataset and model - if len(dataset.dataset) == 0: - st.error( - "Dataset has no samples. Please check the dataset configuration." - ) - return - - if not hasattr(model, "model_cfg") or model.model_cfg is None: - st.error( - "Model configuration is missing. Please reload the model in the Inference tab." - ) - return - - # Handle ontology translation if provided - ontology_translation_path = None - if ontology_translation is not None: - with tempfile.NamedTemporaryFile( - delete=False, suffix=".json", mode="w" - ) as tmp_trans: - json.dump(json.load(ontology_translation), tmp_trans) - ontology_translation_path = tmp_trans.name - - # Prepare predictions output directory if needed - predictions_outdir = None - if save_predictions or save_visualizations: - if predictions_outdir_input and predictions_outdir_input.strip(): - predictions_outdir = predictions_outdir_input.strip() - os.makedirs(predictions_outdir, exist_ok=True) - else: - predictions_outdir = tempfile.mkdtemp( - prefix="eval_predictions_" - ) - - # Create progress bar for evaluation - progress_bar = st.progress(0) - status_text = st.empty() - - # Create placeholders for intermediate metrics that will be updated in place - intermediate_metrics_placeholder = st.empty() - intermediate_table_placeholder = st.empty() - - def progress_callback(processed, total): - """Progress callback for Streamlit UI""" - try: - progress = processed / total if total > 0 else 0 - progress_bar.progress(progress) - status_text.text( - f"Processing: {processed}/{total} images ({progress:.1%})" - ) - except Exception as e: - st.error(f"Progress callback error: {e}") - - def metrics_callback(metrics_df, processed, total): - """Metrics callback for intermediate results display""" - try: - # Update the metrics placeholder with current summary metrics - if "mean" in metrics_df.columns: - mean_metrics = metrics_df["mean"] - - with intermediate_metrics_placeholder.container(): - st.markdown( - f"#### 📊 Intermediate Results (after {processed} images)" - ) - - col1, col2, col3 = st.columns(3) - with col1: - st.metric("mAP", f"{mean_metrics.get('AP', 0):.3f}") - with col2: - st.metric( - "Mean Precision", - f"{mean_metrics.get('Precision', 0):.3f}", - ) - with col3: - st.metric( - "Mean Recall", - f"{mean_metrics.get('Recall', 0):.3f}", - ) - - # Update the table placeholder with current per-class results - per_class_results = ( - metrics_df.drop(columns=["mean"]) - if "mean" in metrics_df.columns - else metrics_df - ) - per_class_results = per_class_results.drop( - ["AUC-PR", "mAP@[0.5:0.95]"], errors="ignore" - ) - - # Round for display - display_df = per_class_results.copy() - numeric_columns = display_df.select_dtypes( - include=["float64", "int64"] - ).columns - for col in numeric_columns: - if col in display_df.columns: - display_df[col] = display_df[col].round(3) - - with intermediate_table_placeholder.container(): - st.markdown("#### Per-Class Metrics (Intermediate)") - st.dataframe(display_df, width="stretch") - - except Exception as e: - st.error(f"Metrics callback error: {e}") - - # Run evaluation with progress tracking - # Use full dataset for evaluation - - try: - # Use the full dataset for evaluation - results = model.eval( - dataset=dataset, - split=split, - ontology_translation=ontology_translation_path, - predictions_outdir=predictions_outdir, - results_per_sample=save_predictions, - save_visualizations=save_visualizations, - progress_callback=progress_callback, - metrics_callback=metrics_callback, - ) - except Exception as e: - st.error(f"Error in model.eval(): {e}") - return - - # Results ready - - # Clear progress elements and intermediate results - progress_bar.empty() - status_text.empty() - intermediate_metrics_placeholder.empty() - intermediate_table_placeholder.empty() - - # Store results in session state - st.session_state["evaluation_results"] = results - st.session_state["evaluation_config"] = { - "split": split, - "predictions_saved": save_predictions, - "visualizations_saved": save_visualizations, - } + task = st.session_state.get("task", "Image Detection") - st.success("✅ Evaluation completed successfully!") + if task == "Image Detection": + from tabs.tasks.image_detection import render_image_detection_evaluator - except Exception as e: - st.error(f"❌ Evaluation failed: {e}") - import traceback - - st.code(traceback.format_exc()) - - # Display results (either from current evaluation or previous) - if "evaluation_results" in st.session_state: - display_evaluation_results(st.session_state["evaluation_results"]) - - -def display_evaluation_results(results): - """Display evaluation results in a comprehensive format""" - - if results is None: - st.warning("No evaluation results to display.") + render_image_detection_evaluator() return - # Handle new results format (dictionary with metrics_df and metrics_factory) - if isinstance(results, dict): - metrics_df = results.get("metrics_df") - metrics_factory = results.get("metrics_factory") - else: - # Fallback for old format - metrics_df = results - metrics_factory = None - - if metrics_df is None or metrics_df.empty: - st.warning("No evaluation results to display.") - return - - # Display summary metrics - st.markdown("#### Summary Metrics") - - # Get mean metrics - mean is a column - if "mean" in metrics_df.columns: - mean_metrics = metrics_df["mean"] - - col1, col2, col3, col4, col5 = st.columns(5) - with col1: - st.metric("mAP", f"{mean_metrics.get('AP', 0):.3f}") - with col2: - st.metric("Mean Precision", f"{mean_metrics.get('Precision', 0):.3f}") - with col3: - st.metric("Mean Recall", f"{mean_metrics.get('Recall', 0):.3f}") - with col4: - coco_map = mean_metrics.get("mAP@[0.5:0.95]", 0) - st.metric("mAP@[0.5:0.95]", f"{coco_map:.3f}") - with col5: - auc_pr = mean_metrics.get("AUC-PR", 0) - st.metric("AUC-PR", f"{auc_pr:.3f}") - - # Display per-class metrics first - st.markdown("#### Per-Class Metrics") - - # Filter out the 'mean' column for per-class display - per_class_results = ( - metrics_df.drop(columns=["mean"]) - if "mean" in metrics_df.columns - else metrics_df - ) - - # Remove overall metrics rows (AUC-PR and mAP@[0.5:0.95]) from per-class display - per_class_results = per_class_results.drop( - ["AUC-PR", "mAP@[0.5:0.95]"], errors="ignore" - ) - - # Create a more readable display - display_df = per_class_results.copy() - - # Round numeric columns for better display - numeric_columns = display_df.select_dtypes(include=["float64", "int64"]).columns - for col in numeric_columns: - if col in display_df.columns: - display_df[col] = display_df[col].round(3) - - st.dataframe(display_df, width="stretch") - - # Now display Precision-Recall Curve - if metrics_factory is not None: - st.markdown("#### Precision-Recall Curve") - - try: - # Get the precision-recall curve data - curve_data = metrics_factory.get_overall_precision_recall_curve() - auc_pr = metrics_factory.compute_auc_pr() - - # Create the plot using streamlit's plotly integration - import plotly.graph_objects as go - - # Create the precision-recall curve - fig = go.Figure() - - # Add the curve - fig.add_trace( - go.Scatter( - x=curve_data["recall"], - y=curve_data["precision"], - mode="lines", - name="Precision-Recall Curve", - line=dict(color="blue", width=2), - fill="tonexty", - fillcolor="rgba(0, 0, 255, 0.1)", - ) - ) - - # Add AUC-PR annotation - fig.add_annotation( - x=0.6, - y=0.2, - text=f"AUC-PR: {auc_pr:.3f}", - showarrow=False, - font=dict(size=12), - bgcolor="white", - bordercolor="black", - borderwidth=1, - ) - - # Update layout - fig.update_layout( - # title='Overall Precision-Recall Curve', - xaxis_title="Recall", - yaxis_title="Precision", - xaxis=dict(range=[0, 1]), - yaxis=dict(range=[0, 1]), - showlegend=True, - height=500, - ) - - # Add grid - fig.update_xaxes(showgrid=True, gridwidth=1, gridcolor="lightgray") - fig.update_yaxes(showgrid=True, gridwidth=1, gridcolor="lightgray") - - st.plotly_chart(fig, width="stretch") - - except Exception as e: - st.error(f"Error plotting precision-recall curve: {e}") - st.info("Precision-recall curve data not available.") - - # Download results - st.markdown("#### Download Results") - - # Convert to CSV for download - csv = metrics_df.to_csv(index=True) - st.download_button( - label="📥 Download per class metrics", - data=csv, - file_name="evaluation_results.csv", - mime="text/csv", - ) - try: - curve_data = ( - metrics_factory.get_overall_precision_recall_curve() - if metrics_factory is not None - else None - ) - if curve_data is not None: - import pandas as pd - - pr_points_df = pd.DataFrame( - {"recall": curve_data["recall"], "precision": curve_data["precision"]} - ) - pr_csv = pr_points_df.to_csv(index=False) - st.download_button( - label="📈 Download precision-recall points", - data=pr_csv, - file_name="precision_recall_points.csv", - mime="text/csv", - ) - else: - st.write("No precision-recall data available.") - except Exception as e: - st.write(f"Error preparing precision-recall points: {e}") - - # Show detailed statistics - with st.expander("📊 Detailed Statistics"): - st.markdown("**Results Shape:**") - st.write(f"Rows: {metrics_df.shape[0]}, Columns: {metrics_df.shape[1]}") - - st.markdown("**Available Metrics:**") - st.write(list(metrics_df.columns)) - - st.markdown("**Class Names:**") - st.write( - list(metrics_df.index) if len(metrics_df.index) > 0 else "No classes found" - ) - - st.markdown("**DataFrame Info:**") - st.write("Index:", metrics_df.index.tolist()) - st.write("Columns:", metrics_df.columns.tolist()) - - st.markdown("**Sample Data:**") - st.dataframe(metrics_df.head(), width="stretch") - - if "evaluation_config" in st.session_state: - st.markdown("**Evaluation Configuration:**") - config = st.session_state["evaluation_config"] - for key, value in config.items(): - st.write(f"- {key}: {value}") - - # Show precision-recall curve data if available - if metrics_factory is not None: - st.markdown("**Precision-Recall Curve Data:**") - try: - curve_data = metrics_factory.get_overall_precision_recall_curve() - st.write(f"Number of points: {len(curve_data['precision'])}") - st.write( - f"Precision range: {min(curve_data['precision']):.3f} - {max(curve_data['precision']):.3f}" - ) - st.write( - f"Recall range: {min(curve_data['recall']):.3f} - {max(curve_data['recall']):.3f}" - ) - st.write(f"AUC-PR: {metrics_factory.compute_auc_pr():.3f}") - except Exception as e: - st.write(f"Error accessing curve data: {e}") + st.error(f"Unsupported task for evaluator: {task}") diff --git a/tabs/inference.py b/tabs/inference.py index 20589a1d..4932d1d6 100644 --- a/tabs/inference.py +++ b/tabs/inference.py @@ -1,152 +1,13 @@ -from typing import Optional - import streamlit as st -import json -from PIL import Image -try: - import torch -except ImportError: - raise ImportError( - "PyTorch is required for GUI-based inference and evaluation. " - ) - - -def draw_detections(image: Image, predictions: dict, label_map: Optional[dict] = None): - """Draw color-coded bounding boxes and labels on the image using supervision. - - :param image: PIL Image - :type image: Image.Image - :param predictions: dict with 'boxes', 'labels', 'scores' (torch tensors) - :type predictions: dict - :param label_map: dict mapping label indices to class names (optional) - :type label_map: dict - :return: np.ndarray with detections drawn (for st.image) - :rtype: np.ndarray - """ - from perceptionmetrics.utils import image as ui - - boxes = predictions.get("boxes", torch.empty(0)).cpu().numpy() - class_ids = predictions.get("labels", torch.empty(0)).cpu().numpy().astype(int) - - scores_tensor = predictions.get("scores") - if scores_tensor is not None and len(scores_tensor) > 0: - scores = scores_tensor.cpu().numpy() - else: - scores = None - - if label_map: - class_names = [label_map.get(int(label), str(label)) for label in class_ids] - else: - class_names = [str(label) for label in class_ids] - - return ui.draw_detections( - image=image, - boxes=boxes, - class_ids=class_ids, - class_names=class_names, - scores=scores, - ) def inference_tab(): - st.header("Model Inference") - st.markdown("Select an image and run inference using the loaded model.") - - # Check if a model has been loaded and saved in session - if ( - "detection_model" not in st.session_state - or st.session_state.detection_model is None - ): - st.warning("⚠️ Load a model from the sidebar to start inference") - return + task = st.session_state.get("task", "Image Detection") - st.success("Model loaded and saved. You can now select an image.") + if task == "Image Detection": + from tabs.tasks.image_detection import render_image_detection_inference - # Image picker in the tab - image_file = st.file_uploader( - "Choose an image", - type=["jpg", "jpeg", "png"], - key="inference_image_file", - help="Upload an image to run inference", - ) - - if image_file is not None: - with st.spinner("Running inference..."): - try: - image = Image.open(image_file).convert("RGB") - predictions, sample_tensor = st.session_state.detection_model.predict( - image, return_sample=True - ) - from torchvision.transforms import v2 as transforms - - img_to_draw = transforms.ToPILImage()(sample_tensor[0]) - label_map = getattr( - st.session_state.detection_model, "idx_to_class_name", None - ) - result_img = draw_detections(img_to_draw, predictions, label_map) - - st.markdown("#### Detection Results") - st.image(result_img, caption="Detection Results", width="stretch") - - # Display detection statistics - if ( - predictions.get("scores") is not None - and len(predictions["scores"]) > 0 - ): - st.markdown("#### Detection Statistics") - col1, col2, col3 = st.columns(3) - with col1: - st.metric("Total Detections", len(predictions["scores"])) - with col2: - avg_confidence = float(predictions["scores"].mean()) - st.metric("Avg Confidence", f"{avg_confidence:.3f}") - with col3: - max_confidence = float(predictions["scores"].max()) - st.metric("Max Confidence", f"{max_confidence:.3f}") - - # Display and download detection results - st.markdown("#### Detection Details") - - # Convert predictions to JSON format - detection_results = [] - boxes = predictions.get("boxes", torch.empty(0)).cpu().numpy() - labels = predictions.get("labels", torch.empty(0)).cpu().numpy() - scores = predictions.get("scores", torch.empty(0)).cpu().numpy() - - for i in range(len(scores)): - class_name = ( - label_map.get(int(labels[i]), f"class_{labels[i]}") - if label_map - else f"class_{labels[i]}" - ) - detection_results.append( - { - "detection_id": i, - "class_id": int(labels[i]), - "class_name": class_name, - "confidence": float(scores[i]), - "bbox": { - "x1": float(boxes[i][0]), - "y1": float(boxes[i][1]), - "x2": float(boxes[i][2]), - "y2": float(boxes[i][3]), - }, - "bbox_xyxy": boxes[i].tolist(), - } - ) - - with st.expander(" View Detection Results (JSON)", expanded=False): - st.json(detection_results) + render_image_detection_inference() + return - json_str = json.dumps(detection_results, indent=2) - st.download_button( - label="Download Detection Results as JSON", - data=json_str, - file_name="detection_results.json", - mime="application/json", - help="Download the detection results as a JSON file", - ) - else: - st.info("No detections found in the image.") - except Exception as e: - st.error(f"Failed to run inference: {e}") + st.error(f"Unsupported task for inference: {task}") diff --git a/tabs/tasks/__init__.py b/tabs/tasks/__init__.py new file mode 100644 index 00000000..8b137891 --- /dev/null +++ b/tabs/tasks/__init__.py @@ -0,0 +1 @@ + diff --git a/tabs/tasks/image_detection/__init__.py b/tabs/tasks/image_detection/__init__.py new file mode 100644 index 00000000..07cd5ef9 --- /dev/null +++ b/tabs/tasks/image_detection/__init__.py @@ -0,0 +1,26 @@ +__all__ = [ + "render_image_detection_sidebar", + "render_image_detection_viewer", + "render_image_detection_inference", + "render_image_detection_evaluator", +] + + +def __getattr__(name): + if name == "render_image_detection_sidebar": + from tabs.tasks.image_detection.sidebar import render_image_detection_sidebar + + return render_image_detection_sidebar + if name == "render_image_detection_viewer": + from tabs.tasks.image_detection.dataset_viewer import render_image_detection_viewer + + return render_image_detection_viewer + if name == "render_image_detection_inference": + from tabs.tasks.image_detection.inference import render_image_detection_inference + + return render_image_detection_inference + if name == "render_image_detection_evaluator": + from tabs.tasks.image_detection.evaluator import render_image_detection_evaluator + + return render_image_detection_evaluator + raise AttributeError(f"module {__name__!r} has no attribute {name!r}") diff --git a/tabs/tasks/image_detection/dataset_viewer.py b/tabs/tasks/image_detection/dataset_viewer.py new file mode 100644 index 00000000..d198b714 --- /dev/null +++ b/tabs/tasks/image_detection/dataset_viewer.py @@ -0,0 +1,288 @@ +import os +import tempfile + +import streamlit as st +from PIL import Image + +from perceptionmetrics.datasets.coco import CocoDataset, find_img_dir_and_ann_file + + +def render_image_detection_viewer(): + import numpy as np + from perceptionmetrics.datasets.yolo import YOLODataset + from streamlit_image_select import image_select + from supervision.detection.annotate import BoxAnnotator + from supervision.detection.core import Detections + from supervision.draw.color import ColorPalette + + dataset_path = st.session_state.get("dataset_path", "") + dataset_type = st.session_state.get("dataset_type", "COCO").lower() + split = st.session_state.get("split", "val") + + st.header("Dataset Viewer") + + if not dataset_path or not os.path.isdir(dataset_path): + st.warning("Please select a valid dataset folder.") + return + + if dataset_type == "coco": + try: + img_dir, ann_file = find_img_dir_and_ann_file( + dataset_path=dataset_path, split=split + ) + except FileNotFoundError: + st.warning("Dataset files not found. Check path and split.") + return + elif dataset_type == "yolo": + dataset_config_file = st.session_state.get("dataset_config_file", None) + img_dir = os.path.join(dataset_path, f"images/{split}") + if not os.path.isdir(img_dir): + st.warning("Image directory not found.") + return + if dataset_config_file is None: + st.warning("Dataset configuration file not found. Please upload it.") + return + else: + st.error("Unsupported dataset type.") + return + + _render_detection_viewer_style() + dataset = _load_detection_dataset(dataset_path, dataset_type, split, img_dir, ann_file if dataset_type == "coco" else None) + if dataset is None: + return + + image_files = [ + f for f in os.listdir(img_dir) if f.lower().endswith((".jpg", ".jpeg", ".png")) + ] + if not image_files: + st.warning("No images found.") + return + + image_paths, sample_images, current_page = _render_image_grid_navigation( + dataset_path, split, image_files, img_dir + ) + captions = [_short_caption(name) for name in sample_images] + + img_select_key = f"img_select_all_{dataset_path}_{split}_{current_page}" + img_select_index = st.session_state.get(img_select_key) + if img_select_index is None or not isinstance(img_select_index, int): + img_select_index = 0 + selected_img_path = ( + image_select( + label="", + images=image_paths, + captions=captions, + use_container_width=False, + key=img_select_key, + index=img_select_index, + ) + if image_paths + else None + ) + + if selected_img_path: + selected_img_name = os.path.basename(selected_img_path) + try: + img = Image.open(selected_img_path).convert("RGB") + img_np = np.array(img) + + if dataset_type == "yolo": + ann_row = dataset.dataset[ + dataset.dataset["image"].str.endswith(selected_img_name) + ] + else: + ann_row = dataset.dataset[dataset.dataset["image"] == selected_img_name] + + if not ann_row.empty: + annotation_id = ann_row.iloc[0]["annotation"] + if dataset_type == "yolo": + annotation_id = os.path.join(dataset.dataset_dir, annotation_id) + + boxes, category_indices = dataset.read_annotation(annotation_id) + class_names = _class_names_from_ontology(dataset, category_indices) + + palette = ColorPalette.default() + detections = Detections( + xyxy=np.array(boxes), class_id=np.array(category_indices) + ) + annotator = BoxAnnotator( + color=palette, text_scale=0.7, text_thickness=1, text_padding=2 + ) + annotated_img = annotator.annotate( + scene=img_np, detections=detections, labels=class_names + ) + + annotated_pil = Image.fromarray(annotated_img) + resample = getattr(getattr(Image, "Resampling", Image), "LANCZOS") + annotated_pil.thumbnail((500, 500), resample) + st.image(annotated_pil, width="content") + else: + st.warning("No annotation found for this image.") + except Exception as e: + st.error(f"Error displaying image: {e}") + else: + st.info("Select an image to view with annotations.") + +def _render_detection_viewer_style(): + nav_col1, nav_col2, nav_col3, nav_col4 = st.columns([1, 1, 2, 1.5]) + with nav_col1: + pass + with nav_col2: + pass + with nav_col3: + pass + with nav_col4: + st.markdown( + """ + + """, + unsafe_allow_html=True, + ) + st.markdown("
", unsafe_allow_html=True) + +def _load_detection_dataset(dataset_path, dataset_type, split, img_dir, ann_file): + from perceptionmetrics.datasets.yolo import YOLODataset + + dataset_key = f"{dataset_path}_{split}" + if dataset_key not in st.session_state: + try: + if dataset_type == "coco": + st.session_state[dataset_key] = CocoDataset( + annotation_file=ann_file, + image_dir=img_dir, + split=split, + ) + elif dataset_type == "yolo": + dataset_config_file = st.session_state.get("dataset_config_file", None) + if dataset_config_file is None: + st.warning("Dataset configuration file not found. Please upload it.") + return None + with tempfile.NamedTemporaryFile(delete=False, suffix=".yaml") as tmp: + tmp.write(dataset_config_file.read()) + tmp_path = tmp.name + + yolo_dataset = YOLODataset(tmp_path, dataset_path) + st.session_state["full_dataset_df"] = yolo_dataset.dataset + yolo_dataset.dataset = yolo_dataset.dataset[ + yolo_dataset.dataset["split"] == split + ].reset_index(drop=True) + st.session_state[dataset_key] = yolo_dataset + os.unlink(tmp_path) + else: + st.error("Unsupported dataset type.") + return None + except Exception as e: + st.error(f"Failed to load dataset: {e}") + return None + else: + try: + cached_ds = st.session_state[dataset_key] + cached_split = getattr(cached_ds, "split", None) + if cached_split != split: + if dataset_type == "coco": + st.session_state[dataset_key] = CocoDataset( + annotation_file=ann_file, + image_dir=img_dir, + split=split, + ) + elif dataset_type == "yolo": + yolo_dataset = st.session_state[dataset_key] + yolo_dataset.dataset = st.session_state["full_dataset_df"][ + st.session_state["full_dataset_df"]["split"] == split + ].reset_index(drop=True) + st.session_state[dataset_key] = yolo_dataset + except Exception: + pass + return st.session_state[dataset_key] + +def _render_image_grid_navigation(dataset_path, split, image_files, img_dir): + images_per_page = 12 + total_pages = (len(image_files) + images_per_page - 1) // images_per_page + page_key = f"image_page_{dataset_path}_{split}" + + if page_key not in st.session_state: + st.session_state[page_key] = 0 + current_page = max(0, min(st.session_state[page_key], total_pages - 1)) + st.session_state[page_key] = current_page + + start_idx = current_page * images_per_page + sample_images = image_files[start_idx : start_idx + images_per_page] + image_paths = [os.path.join(img_dir, img_name) for img_name in sample_images] + + col1, col2, col3, col4 = st.columns([0.5, 9.5, 0.5, 0.5]) + with col1: + if st.button("⟨", key="prev_page_btn", disabled=(current_page == 0)): + st.session_state[page_key] = current_page - 1 + st.rerun() + with col2: + st.markdown( + f"
Page {current_page + 1} of {total_pages}
", + unsafe_allow_html=True, + ) + with col3: + if st.button( + "⟩", key="next_page_btn", disabled=(current_page >= total_pages - 1) + ): + st.session_state[page_key] = current_page + 1 + st.rerun() + with col4: + if st.button( + "🔍", + key="search_icon_btn", + help="Search for an image by name", + disabled=not (dataset_path and os.path.isdir(dataset_path)), + ): + st.session_state["show_search_dropdown"] = True + + if st.session_state.get("show_search_dropdown", False): + _render_image_search_dropdown(image_files, images_per_page, page_key, dataset_path, split) + + return image_paths, sample_images, current_page + +def _render_image_search_dropdown(image_files, images_per_page, page_key, dataset_path, split): + col1, col2, col3 = st.columns([4, 1, 1]) + with col1: + selected_img = st.selectbox("Search image:", options=image_files, key="search_image") + with col2: + st.markdown("
", unsafe_allow_html=True) + if st.button("Go to image", key="go_to_image_btn"): + new_page = image_files.index(selected_img) // images_per_page + st.session_state[page_key] = new_page + st.session_state[f"img_select_all_{dataset_path}_{split}_{new_page}"] = ( + image_files.index(selected_img) % images_per_page + ) + st.session_state["show_search_dropdown"] = False + st.rerun() + with col3: + st.markdown("
", unsafe_allow_html=True) + if st.button("Cancel", key="cancel_search_btn"): + st.session_state["show_search_dropdown"] = False + st.rerun() + +def _short_caption(name: str) -> str: + caption_len_limit = 17 + if len(name) > caption_len_limit: + return name[:caption_len_limit] + "..." + name[-3:] + return name + +def _class_names_from_ontology(dataset, category_indices): + ontology = getattr(dataset, "ontology", None) + if ontology is None and hasattr(dataset.dataset, "attrs"): + ontology = dataset.dataset.attrs.get("ontology", None) + + if ontology: + catid_to_name = {v["idx"]: k for k, v in ontology.items()} + return [catid_to_name.get(cat_id, str(cat_id)) for cat_id in category_indices] + return [str(cat_id) for cat_id in category_indices] + diff --git a/tabs/tasks/image_detection/evaluator.py b/tabs/tasks/image_detection/evaluator.py new file mode 100644 index 00000000..9554ed3d --- /dev/null +++ b/tabs/tasks/image_detection/evaluator.py @@ -0,0 +1,447 @@ +import os +import tempfile + +import streamlit as st + +from perceptionmetrics.datasets.coco import CocoDataset, find_img_dir_and_ann_file +from perceptionmetrics.utils.gui import browse_folder +from tabs.tasks.image_detection.sidebar import _uploaded_json_to_tempfile + + +def browse_predictions_outdir(): + folder = browse_folder() + if folder: + st.session_state.predictions_outdir = folder + + +def render_image_detection_evaluator(): + st.header("Evaluator") + st.markdown("Evaluate your model on the loaded dataset using PerceptionMetrics.") + + dataset_available, model_available, dataset, model = _get_eval_dataset_and_model() + + st.markdown("### Evaluation Configuration") + + save_predictions = st.checkbox( + "Save Predictions", + value=False, + help="Save individual predictions and metrics per sample", + ) + + save_visualizations = st.checkbox( + "Save Visualizations", + value=False, + help="Save visualized qualitative results (Image + GT + Preds)", + ) + + predictions_outdir_input = None + if save_predictions or save_visualizations: + col1, col2 = st.columns([3, 1]) + with col1: + st.text_input("Predictions Output Directory", key="predictions_outdir") + with col2: + st.markdown( + "
", + unsafe_allow_html=True, + ) + st.button( + "Browse", on_click=browse_predictions_outdir, key="browse_preds_outdir" + ) + predictions_outdir_input = st.session_state.get("predictions_outdir") + + ontology_translation = st.file_uploader( + "Ontology Translation (Optional)", + type=["json"], + help="JSON file for translating between dataset and model ontologies", + ) + + output_dir_required = save_predictions or save_visualizations + output_dir_missing = output_dir_required and not ( + predictions_outdir_input and predictions_outdir_input.strip() + ) + + if output_dir_missing: + st.warning( + "Please provide a Predictions Output Directory to enable evaluation " + "when 'Save Predictions' or 'Save Visualizations' is turned on." + ) + + if st.button( + "Run Evaluation", + type="primary", + disabled=not (dataset_available and model_available) or output_dir_missing, + ): + _run_detection_evaluation( + dataset, + model, + ontology_translation, + save_predictions, + save_visualizations, + predictions_outdir_input, + ) + + if "evaluation_results" in st.session_state: + display_evaluation_results(st.session_state["evaluation_results"]) + +def _get_eval_dataset_and_model(): + dataset_available = False + model_available = False + dataset = None + model = None + + dataset_path = st.session_state.get("dataset_path", "") + dataset_type = st.session_state.get("dataset_type", "Coco") + split = st.session_state.get("split", "val") + dataset_key = f"{dataset_path}_{split}" + + if dataset_key in st.session_state: + dataset = st.session_state[dataset_key] + dataset_available = True + st.success( + f"Dataset loaded: {dataset_path} ({split} split) - {len(dataset.dataset)} samples" + ) + elif dataset_path and os.path.isdir(dataset_path): + try: + if dataset_type.lower() == "coco": + img_dir, ann_file = find_img_dir_and_ann_file( + dataset_path=dataset_path, split=split + ) + + if os.path.isdir(img_dir) and os.path.isfile(ann_file): + st.session_state[dataset_key] = CocoDataset( + annotation_file=ann_file, image_dir=img_dir, split=split + ) + st.session_state[dataset_key].make_fname_global() + dataset = st.session_state[dataset_key] + dataset_available = True + st.success( + f"Dataset loaded: {dataset_path} ({split} split) - {len(dataset.dataset)} samples" + ) + else: + st.warning( + "Dataset files not found. Please check the dataset path and split in the sidebar." + ) + else: + st.warning("Only COCO datasets are currently supported for evaluation.") + except Exception as e: + st.error(f"Error loading dataset: {e}") + else: + st.warning("No dataset path provided. Please set the dataset path in the sidebar.") + + if ( + "detection_model" in st.session_state + and st.session_state.detection_model is not None + ): + model = st.session_state.detection_model + model_available = True + st.success("Model loaded and ready for evaluation") + else: + st.warning("No model loaded. Please load a model using the sidebar.") + + return dataset_available, model_available, dataset, model + +def _run_detection_evaluation( + dataset, + model, + ontology_translation, + save_predictions, + save_visualizations, + predictions_outdir_input, +): + split = st.session_state.get("split", "val") + with st.spinner("Running evaluation..."): + try: + if len(dataset.dataset) == 0: + st.error("Dataset has no samples. Please check the dataset configuration.") + return + + if not hasattr(model, "model_cfg") or model.model_cfg is None: + st.error("Model configuration is missing. Please reload the model.") + return + + ontology_translation_path = None + if ontology_translation is not None: + ontology_translation_path = _uploaded_json_to_tempfile(ontology_translation) + + predictions_outdir = None + if save_predictions or save_visualizations: + if predictions_outdir_input and predictions_outdir_input.strip(): + predictions_outdir = predictions_outdir_input.strip() + os.makedirs(predictions_outdir, exist_ok=True) + else: + predictions_outdir = tempfile.mkdtemp(prefix="eval_predictions_") + + progress_bar = st.progress(0) + status_text = st.empty() + intermediate_metrics_placeholder = st.empty() + intermediate_table_placeholder = st.empty() + + def progress_callback(processed, total): + try: + progress = processed / total if total > 0 else 0 + progress_bar.progress(progress) + status_text.text( + f"Processing: {processed}/{total} images ({progress:.1%})" + ) + except Exception as e: + st.error(f"Progress callback error: {e}") + + def metrics_callback(metrics_df, processed, total): + _render_intermediate_detection_metrics( + metrics_df, + processed, + intermediate_metrics_placeholder, + intermediate_table_placeholder, + ) + + try: + results = model.eval( + dataset=dataset, + split=split, + ontology_translation=ontology_translation_path, + predictions_outdir=predictions_outdir, + results_per_sample=save_predictions, + save_visualizations=save_visualizations, + progress_callback=progress_callback, + metrics_callback=metrics_callback, + ) + except Exception as e: + st.error(f"Error in model.eval(): {e}") + return + + progress_bar.empty() + status_text.empty() + intermediate_metrics_placeholder.empty() + intermediate_table_placeholder.empty() + + st.session_state["evaluation_results"] = results + st.session_state["evaluation_config"] = { + "split": split, + "predictions_saved": save_predictions, + "visualizations_saved": save_visualizations, + } + + st.success("Evaluation completed successfully!") + except Exception as e: + st.error(f"Evaluation failed: {e}") + import traceback + + st.code(traceback.format_exc()) + +def _render_intermediate_detection_metrics( + metrics_df, + processed, + intermediate_metrics_placeholder, + intermediate_table_placeholder, +): + try: + if "mean" in metrics_df.columns: + mean_metrics = metrics_df["mean"] + + with intermediate_metrics_placeholder.container(): + st.markdown(f"#### Intermediate Results (after {processed} images)") + col1, col2, col3 = st.columns(3) + with col1: + st.metric("mAP", f"{mean_metrics.get('AP', 0):.3f}") + with col2: + st.metric( + "Mean Precision", + f"{mean_metrics.get('Precision', 0):.3f}", + ) + with col3: + st.metric("Mean Recall", f"{mean_metrics.get('Recall', 0):.3f}") + + per_class_results = ( + metrics_df.drop(columns=["mean"]) + if "mean" in metrics_df.columns + else metrics_df + ) + per_class_results = per_class_results.drop( + ["AUC-PR", "mAP@[0.5:0.95]"], errors="ignore" + ) + + display_df = per_class_results.copy() + numeric_columns = display_df.select_dtypes(include=["float64", "int64"]).columns + for col in numeric_columns: + if col in display_df.columns: + display_df[col] = display_df[col].round(3) + + with intermediate_table_placeholder.container(): + st.markdown("#### Per-Class Metrics (Intermediate)") + st.dataframe(display_df, width="stretch") + except Exception as e: + st.error(f"Metrics callback error: {e}") + +def display_evaluation_results(results): + if results is None: + st.warning("No evaluation results to display.") + return + + if isinstance(results, dict): + metrics_df = results.get("metrics_df") + metrics_factory = results.get("metrics_factory") + else: + metrics_df = results + metrics_factory = None + + if metrics_df is None or metrics_df.empty: + st.warning("No evaluation results to display.") + return + + st.markdown("#### Summary Metrics") + + if "mean" in metrics_df.columns: + mean_metrics = metrics_df["mean"] + col1, col2, col3, col4, col5 = st.columns(5) + with col1: + st.metric("mAP", f"{mean_metrics.get('AP', 0):.3f}") + with col2: + st.metric("Mean Precision", f"{mean_metrics.get('Precision', 0):.3f}") + with col3: + st.metric("Mean Recall", f"{mean_metrics.get('Recall', 0):.3f}") + with col4: + coco_map = mean_metrics.get("mAP@[0.5:0.95]", 0) + st.metric("mAP@[0.5:0.95]", f"{coco_map:.3f}") + with col5: + auc_pr = mean_metrics.get("AUC-PR", 0) + st.metric("AUC-PR", f"{auc_pr:.3f}") + + st.markdown("#### Per-Class Metrics") + per_class_results = ( + metrics_df.drop(columns=["mean"]) + if "mean" in metrics_df.columns + else metrics_df + ) + per_class_results = per_class_results.drop( + ["AUC-PR", "mAP@[0.5:0.95]"], errors="ignore" + ) + + display_df = per_class_results.copy() + numeric_columns = display_df.select_dtypes(include=["float64", "int64"]).columns + for col in numeric_columns: + if col in display_df.columns: + display_df[col] = display_df[col].round(3) + + st.dataframe(display_df, width="stretch") + + if metrics_factory is not None: + _render_precision_recall_curve(metrics_factory) + + st.markdown("#### Download Results") + st.download_button( + label="Download per class metrics", + data=metrics_df.to_csv(index=True), + file_name="evaluation_results.csv", + mime="text/csv", + ) + _render_pr_points_download(metrics_factory) + _render_detailed_detection_statistics(metrics_df, metrics_factory) + +def _render_precision_recall_curve(metrics_factory): + st.markdown("#### Precision-Recall Curve") + try: + import plotly.graph_objects as go + + curve_data = metrics_factory.get_overall_precision_recall_curve() + auc_pr = metrics_factory.compute_auc_pr() + fig = go.Figure() + fig.add_trace( + go.Scatter( + x=curve_data["recall"], + y=curve_data["precision"], + mode="lines", + name="Precision-Recall Curve", + line=dict(color="blue", width=2), + fill="tonexty", + fillcolor="rgba(0, 0, 255, 0.1)", + ) + ) + fig.add_annotation( + x=0.6, + y=0.2, + text=f"AUC-PR: {auc_pr:.3f}", + showarrow=False, + font=dict(size=12), + bgcolor="white", + bordercolor="black", + borderwidth=1, + ) + fig.update_layout( + xaxis_title="Recall", + yaxis_title="Precision", + xaxis=dict(range=[0, 1]), + yaxis=dict(range=[0, 1]), + showlegend=True, + height=500, + ) + fig.update_xaxes(showgrid=True, gridwidth=1, gridcolor="lightgray") + fig.update_yaxes(showgrid=True, gridwidth=1, gridcolor="lightgray") + st.plotly_chart(fig, width="stretch") + except Exception as e: + st.error(f"Error plotting precision-recall curve: {e}") + st.info("Precision-recall curve data not available.") + +def _render_pr_points_download(metrics_factory): + try: + curve_data = ( + metrics_factory.get_overall_precision_recall_curve() + if metrics_factory is not None + else None + ) + if curve_data is not None: + import pandas as pd + + pr_points_df = pd.DataFrame( + {"recall": curve_data["recall"], "precision": curve_data["precision"]} + ) + st.download_button( + label="Download precision-recall points", + data=pr_points_df.to_csv(index=False), + file_name="precision_recall_points.csv", + mime="text/csv", + ) + else: + st.write("No precision-recall data available.") + except Exception as e: + st.write(f"Error preparing precision-recall points: {e}") + +def _render_detailed_detection_statistics(metrics_df, metrics_factory): + with st.expander("Detailed Statistics"): + st.markdown("**Results Shape:**") + st.write(f"Rows: {metrics_df.shape[0]}, Columns: {metrics_df.shape[1]}") + + st.markdown("**Available Metrics:**") + st.write(list(metrics_df.columns)) + + st.markdown("**Class Names:**") + st.write( + list(metrics_df.index) if len(metrics_df.index) > 0 else "No classes found" + ) + + st.markdown("**DataFrame Info:**") + st.write("Index:", metrics_df.index.tolist()) + st.write("Columns:", metrics_df.columns.tolist()) + + st.markdown("**Sample Data:**") + st.dataframe(metrics_df.head(), width="stretch") + + if "evaluation_config" in st.session_state: + st.markdown("**Evaluation Configuration:**") + config = st.session_state["evaluation_config"] + for key, value in config.items(): + st.write(f"- {key}: {value}") + + if metrics_factory is not None: + st.markdown("**Precision-Recall Curve Data:**") + try: + curve_data = metrics_factory.get_overall_precision_recall_curve() + st.write(f"Number of points: {len(curve_data['precision'])}") + st.write( + f"Precision range: {min(curve_data['precision']):.3f} - {max(curve_data['precision']):.3f}" + ) + st.write( + f"Recall range: {min(curve_data['recall']):.3f} - {max(curve_data['recall']):.3f}" + ) + st.write(f"AUC-PR: {metrics_factory.compute_auc_pr():.3f}") + except Exception as e: + st.write(f"Error accessing curve data: {e}") diff --git a/tabs/tasks/image_detection/inference.py b/tabs/tasks/image_detection/inference.py new file mode 100644 index 00000000..f7a290e1 --- /dev/null +++ b/tabs/tasks/image_detection/inference.py @@ -0,0 +1,132 @@ +import json +from typing import Optional + +import streamlit as st +from PIL import Image + + +def draw_detections(image: Image.Image, predictions: dict, label_map: Optional[dict] = None): + import torch + from perceptionmetrics.utils import image as ui + + boxes = predictions.get("boxes", torch.empty(0)).cpu().numpy() + class_ids = predictions.get("labels", torch.empty(0)).cpu().numpy().astype(int) + + scores_tensor = predictions.get("scores") + scores = scores_tensor.cpu().numpy() if scores_tensor is not None and len(scores_tensor) > 0 else None + + if label_map: + class_names = [label_map.get(int(label), str(label)) for label in class_ids] + else: + class_names = [str(label) for label in class_ids] + + return ui.draw_detections( + image=image, + boxes=boxes, + class_ids=class_ids, + class_names=class_names, + scores=scores, + ) + +def render_image_detection_inference(): + import torch + + st.header("Model Inference") + st.markdown("Select an image and run inference using the loaded model.") + + if ( + "detection_model" not in st.session_state + or st.session_state.detection_model is None + ): + st.warning("Load a model from the sidebar to start inference") + return + + st.success("Model loaded and saved. You can now select an image.") + + image_file = st.file_uploader( + "Choose an image", + type=["jpg", "jpeg", "png"], + key="inference_image_file", + help="Upload an image to run inference", + ) + + if image_file is not None: + with st.spinner("Running inference..."): + try: + image = Image.open(image_file).convert("RGB") + predictions, sample_tensor = st.session_state.detection_model.predict( + image, return_sample=True + ) + from torchvision.transforms import v2 as transforms + + img_to_draw = transforms.ToPILImage()(sample_tensor[0]) + label_map = getattr( + st.session_state.detection_model, "idx_to_class_name", None + ) + result_img = draw_detections(img_to_draw, predictions, label_map) + + st.markdown("#### Detection Results") + st.image(result_img, caption="Detection Results", width="stretch") + + if ( + predictions.get("scores") is not None + and len(predictions["scores"]) > 0 + ): + st.markdown("#### Detection Statistics") + col1, col2, col3 = st.columns(3) + with col1: + st.metric("Total Detections", len(predictions["scores"])) + with col2: + avg_confidence = float(predictions["scores"].mean()) + st.metric("Avg Confidence", f"{avg_confidence:.3f}") + with col3: + max_confidence = float(predictions["scores"].max()) + st.metric("Max Confidence", f"{max_confidence:.3f}") + + st.markdown("#### Detection Details") + detection_results = _detection_results_to_json(predictions, label_map, torch) + + with st.expander(" View Detection Results (JSON)", expanded=False): + st.json(detection_results) + + st.download_button( + label="Download Detection Results as JSON", + data=json.dumps(detection_results, indent=2), + file_name="detection_results.json", + mime="application/json", + help="Download the detection results as a JSON file", + ) + else: + st.info("No detections found in the image.") + except Exception as e: + st.error(f"Failed to run inference: {e}") + +def _detection_results_to_json(predictions, label_map, torch_module): + detection_results = [] + boxes = predictions.get("boxes", torch_module.empty(0)).cpu().numpy() + labels = predictions.get("labels", torch_module.empty(0)).cpu().numpy() + scores = predictions.get("scores", torch_module.empty(0)).cpu().numpy() + + for i in range(len(scores)): + class_name = ( + label_map.get(int(labels[i]), f"class_{labels[i]}") + if label_map + else f"class_{labels[i]}" + ) + detection_results.append( + { + "detection_id": i, + "class_id": int(labels[i]), + "class_name": class_name, + "confidence": float(scores[i]), + "bbox": { + "x1": float(boxes[i][0]), + "y1": float(boxes[i][1]), + "x2": float(boxes[i][2]), + "y2": float(boxes[i][3]), + }, + "bbox_xyxy": boxes[i].tolist(), + } + ) + return detection_results + diff --git a/tabs/tasks/image_detection/sidebar.py b/tabs/tasks/image_detection/sidebar.py new file mode 100644 index 00000000..f7bcfaab --- /dev/null +++ b/tabs/tasks/image_detection/sidebar.py @@ -0,0 +1,329 @@ +import json +import os +import tempfile +from typing import Optional + +import streamlit as st + +from perceptionmetrics.utils.gui import browse_folder + + +def browse_dataset_path(): + st.session_state.dataset_path = browse_folder() + + +def render_image_detection_sidebar(available_devices): + with st.expander("Image Detection Dataset", expanded=True): + col1, col2 = st.columns(2) + with col1: + st.selectbox("Type", ["COCO", "YOLO"], key="dataset_type") + with col2: + st.selectbox("Split", ["train", "val", "test"], key="split") + + col1, col2 = st.columns([3, 1]) + with col1: + st.text_input("Dataset Folder", key="dataset_path") + with col2: + st.markdown( + "
", + unsafe_allow_html=True, + ) + st.button("Browse", on_click=browse_dataset_path) + + if st.session_state.get("dataset_type", "COCO") == "YOLO": + st.file_uploader( + "Dataset Configuration (.yaml)", + type=["yaml"], + key="dataset_config_file", + help="Upload a YAML dataset configuration file.", + ) + + with st.expander("Image Detection Model", expanded=False): + st.file_uploader( + "Model File (.pt, .onnx, .h5, .pb, .pth, .torchscript)", + type=["pt", "onnx", "h5", "pb", "pth", "torchscript"], + key="model_file", + help="Upload your trained model file.", + max_upload_size=1024, + ) + st.file_uploader( + "Ontology File (.json)", + type=["json"], + key="ontology_file", + help="Upload a JSON file with class labels.", + ) + st.radio( + "Configuration Method:", + ["Manual Configuration", "Upload Config File"], + key="config_option", + horizontal=True, + ) + if ( + st.session_state.get("config_option", "Manual Configuration") + == "Upload Config File" + ): + st.file_uploader( + "Configuration File (.json)", + type=["json"], + key="config_file", + help="Upload a JSON configuration file.", + ) + else: + _render_manual_detection_model_config(available_devices) + + if st.button( + "Load Model", + type="primary", + width="stretch", + help="Load and save the model for use in the Inference tab", + key="sidebar_load_model_btn", + ): + load_image_detection_model() + +def _render_manual_detection_model_config(available_devices): + col1, col2 = st.columns(2) + with col1: + st.slider( + "Confidence Threshold", + min_value=0.0, + max_value=1.0, + step=0.01, + key="confidence_threshold", + help="Minimum confidence score for detections", + ) + st.slider( + "NMS Threshold", + min_value=0.0, + max_value=1.0, + step=0.01, + key="nms_threshold", + help="Non-maximum suppression threshold", + ) + st.number_input( + "Max Detections/Image", + min_value=1, + max_value=1000, + step=1, + key="max_detections", + ) + with col2: + st.selectbox("Device", available_devices, key="device") + st.selectbox( + "Model Format", + ["torchvision", "YOLO"], + index=( + 0 + if st.session_state.get("model_format", "torchvision") + == "torchvision" + else 1 + ), + key="model_format", + ) + st.number_input( + "Batch Size", + min_value=1, + max_value=256, + step=1, + key="batch_size", + ) + st.number_input( + "Evaluation Step", + min_value=0, + max_value=1000, + step=1, + key="evaluation_step", + help="Update UI with intermediate metrics every N images (0 = disable intermediate updates)", + ) + + st.write("---") + st.write("**Image Size Configuration**") + + enable_resize = st.checkbox("Enable Resize", value=True, key="enable_resize") + + if enable_resize: + resize_strategy = st.radio( + "Resize Strategy", + ["Fixed Dimensions", "Min Side"], + key="resize_strategy", + horizontal=True, + label_visibility="collapsed", + ) + + if resize_strategy == "Fixed Dimensions": + c1, c2 = st.columns(2) + with c1: + st.number_input( + "Image Resize Height", + min_value=1, + max_value=4096, + value=640, + step=1, + key="resize_height", + help="Height to resize images for inference", + ) + with c2: + st.number_input( + "Image Resize Width", + min_value=1, + max_value=4096, + value=640, + step=1, + key="resize_width", + help="Width to resize images for inference", + ) + else: + st.number_input( + "Min Side", + min_value=1, + max_value=4096, + value=640, + step=1, + key="min_side", + help="Minimum size of the shorter side of the image", + ) + + enable_crop = st.checkbox("Enable Center Crop", key="enable_crop") + + if enable_crop: + c1, c2 = st.columns(2) + with c1: + st.number_input( + "Crop Height", + min_value=1, + max_value=4096, + value=640, + step=1, + key="crop_height", + help="Center crop height", + ) + with c2: + st.number_input( + "Crop Width", + min_value=1, + max_value=4096, + value=640, + step=1, + key="crop_width", + help="Center crop width", + ) + +def load_image_detection_model(): + from perceptionmetrics.models.torch_detection import TorchImageDetectionModel + + model_file = st.session_state.get("model_file") + ontology_file = st.session_state.get("ontology_file") + config_path = _write_detection_config() + + if model_file is None: + st.error("Please upload a model file") + elif config_path is None: + st.error("Please provide a valid model configuration") + elif ontology_file is None: + st.error("Please upload an ontology file") + else: + with st.spinner("Loading model..."): + ontology_path = _uploaded_json_to_tempfile(ontology_file) + model_temp_path = _uploaded_model_to_tempfile(model_file) + + if ontology_path and model_temp_path: + try: + model = TorchImageDetectionModel( + model=model_temp_path, + model_cfg=config_path, + ontology_fname=ontology_path, + device=st.session_state.get("device", "cpu"), + ) + st.session_state.detection_model = model + st.session_state.detection_model_loaded = True + st.success("Model loaded and saved for inference") + except Exception as e: + st.session_state.detection_model = None + st.session_state.detection_model_loaded = False + st.error(f"Failed to load model: {e}") + +def _write_detection_config() -> Optional[str]: + config_option = st.session_state.get("config_option", "Manual Configuration") + config_file = ( + st.session_state.get("config_file") + if config_option == "Upload Config File" + else None + ) + + try: + if config_option == "Upload Config File": + if config_file is None: + st.error("Please upload a configuration file") + return None + config_data = json.load(config_file) + else: + config_data = _manual_detection_config() + + with tempfile.NamedTemporaryFile( + delete=False, suffix=".json", mode="w" + ) as tmp_cfg: + json.dump(config_data, tmp_cfg) + return tmp_cfg.name + except Exception as e: + st.error(f"Failed to prepare configuration: {e}") + return None + +def _manual_detection_config() -> dict: + resize_cfg = None + if st.session_state.get("enable_resize", True): + resize_strategy = st.session_state.get("resize_strategy", "Fixed Dimensions") + if resize_strategy == "Fixed Dimensions": + resize_cfg = { + "height": int(st.session_state.get("resize_height", 640)), + "width": int(st.session_state.get("resize_width", 640)), + } + else: + resize_cfg = {"min_side": int(st.session_state.get("min_side", 640))} + + config_data = { + "confidence_threshold": float( + st.session_state.get("confidence_threshold", 0.5) + ), + "nms_threshold": float(st.session_state.get("nms_threshold", 0.5)), + "max_detections_per_image": int( + st.session_state.get("max_detections", 100) + ), + "device": st.session_state.get("device", "cpu"), + "batch_size": int(st.session_state.get("batch_size", 1)), + "evaluation_step": int(st.session_state.get("evaluation_step", 5)), + "model_format": st.session_state.get("model_format", "torchvision").lower(), + } + if resize_cfg is not None: + config_data["resize"] = resize_cfg + + if st.session_state.get("enable_crop", False): + config_data["crop"] = { + "height": int(st.session_state.get("crop_height", 640)), + "width": int(st.session_state.get("crop_width", 640)), + } + + return config_data + +def _uploaded_json_to_tempfile(uploaded_file) -> Optional[str]: + try: + data = json.load(uploaded_file) + with tempfile.NamedTemporaryFile( + delete=False, suffix=".json", mode="w" + ) as tmp_file: + json.dump(data, tmp_file) + return tmp_file.name + except Exception as e: + st.error(f"Failed to load JSON file: {e}") + return None + +def _uploaded_model_to_tempfile(uploaded_file) -> Optional[str]: + try: + suffix = os.path.splitext(uploaded_file.name)[1] or ".pt" + with tempfile.NamedTemporaryFile( + delete=False, suffix=suffix, mode="wb" + ) as tmp_model: + tmp_model.write(uploaded_file.read()) + return tmp_model.name + except Exception as e: + st.error(f"Failed to save model file: {e}") + return None + From 2ab3e71621dfcf66eba651045b5e5dd67897e807 Mon Sep 17 00:00:00 2001 From: tejass Date: Fri, 26 Jun 2026 15:56:13 +0200 Subject: [PATCH 02/10] Add image segmentation dataset viewer --- app.py | 16 +- tabs/dataset_viewer.py | 12 +- tabs/evaluator.py | 15 +- tabs/inference.py | 15 +- tabs/tasks/__init__.py | 1 - tabs/tasks/image_detection/__init__.py | 26 - tabs/tasks/image_detection/dataset_viewer.py | 326 +++++----- tabs/tasks/image_detection/evaluator.py | 558 ++++++++++-------- tabs/tasks/image_detection/inference.py | 98 +-- tabs/tasks/image_detection/sidebar.py | 8 +- .../image_segmentation/dataset_viewer.py | 175 ++++++ tabs/tasks/image_segmentation/evaluator.py | 2 + tabs/tasks/image_segmentation/inference.py | 2 + tabs/tasks/image_segmentation/sidebar.py | 73 +++ .../lidar_segmentation/dataset_viewer.py | 2 + tabs/tasks/lidar_segmentation/evaluator.py | 2 + tabs/tasks/lidar_segmentation/inference.py | 2 + tabs/tasks/lidar_segmentation/sidebar.py | 2 + 18 files changed, 853 insertions(+), 482 deletions(-) delete mode 100644 tabs/tasks/__init__.py delete mode 100644 tabs/tasks/image_detection/__init__.py create mode 100644 tabs/tasks/image_segmentation/dataset_viewer.py create mode 100644 tabs/tasks/image_segmentation/evaluator.py create mode 100644 tabs/tasks/image_segmentation/inference.py create mode 100644 tabs/tasks/image_segmentation/sidebar.py create mode 100644 tabs/tasks/lidar_segmentation/dataset_viewer.py create mode 100644 tabs/tasks/lidar_segmentation/evaluator.py create mode 100644 tabs/tasks/lidar_segmentation/inference.py create mode 100644 tabs/tasks/lidar_segmentation/sidebar.py diff --git a/app.py b/app.py index cbdbf3c9..9728ae32 100644 --- a/app.py +++ b/app.py @@ -4,9 +4,9 @@ from tabs.dataset_viewer import dataset_viewer_tab from tabs.evaluator import evaluator_tab from tabs.inference import inference_tab -from tabs.tasks.image_detection import render_image_detection_sidebar - - +from tabs.tasks.image_detection.sidebar import render_image_detection_sidebar +from tabs.tasks.image_segmentation.sidebar import render_image_segmentation_sidebar +from tabs.tasks.lidar_segmentation.sidebar import render_lidar_segmentation_sidebar st.set_page_config(page_title="PerceptionMetrics", layout="wide") PAGES = { @@ -37,15 +37,21 @@ with st.sidebar: task = st.selectbox( "Task", - ["Image Detection"], + ["Image Detection", "Image Segmentation", "Lidar Segmentation"], key="task", - help="Only image detection is currently wired in the GUI.", + help="Image segmentation is currently a placeholder.", ) if task == "Image Detection": render_image_detection_sidebar(available_devices) + elif task == "Image Segmentation": + render_image_segmentation_sidebar(available_devices) + elif task == "Lidar Segmentation": + render_lidar_segmentation_sidebar(available_devices) else: st.error(f"Unsupported task: {task}") + + tab1, tab2, tab3 = st.tabs(["Dataset Viewer", "Inference", "Evaluator"]) diff --git a/tabs/dataset_viewer.py b/tabs/dataset_viewer.py index c24cc89e..f5338e12 100644 --- a/tabs/dataset_viewer.py +++ b/tabs/dataset_viewer.py @@ -1,13 +1,21 @@ import streamlit as st +from tabs.tasks.image_detection.dataset_viewer import render_image_detection_viewer +from tabs.tasks.image_segmentation.dataset_viewer import render_image_segmentation_viewer +from tabs.tasks.lidar_segmentation.dataset_viewer import render_lidar_segmentation_viewer def dataset_viewer_tab(): task = st.session_state.get("task", "Image Detection") if task == "Image Detection": - from tabs.tasks.image_detection import render_image_detection_viewer - render_image_detection_viewer() return + if task == "Image Segmentation": + render_image_segmentation_viewer() + return + if task == "Lidar Segmentation": + render_lidar_segmentation_viewer() + return + st.error(f"Unsupported task for dataset viewer: {task}") diff --git a/tabs/evaluator.py b/tabs/evaluator.py index bb6da0e0..e3565d7f 100644 --- a/tabs/evaluator.py +++ b/tabs/evaluator.py @@ -1,13 +1,22 @@ import streamlit as st - +from tabs.tasks.image_detection.evaluator import render_image_detection_evaluator +from tabs.tasks.image_segmentation.evaluator import render_image_segmentation_evaluator +from tabs.tasks.lidar_segmentation.evaluator import render_lidar_segmentation_evaluator def evaluator_tab(): task = st.session_state.get("task", "Image Detection") if task == "Image Detection": - from tabs.tasks.image_detection import render_image_detection_evaluator - render_image_detection_evaluator() return + + + if task == "Image Segmentation": + render_image_segmentation_evaluator() + return + + if task == "Lidar Segmentation": + render_lidar_segmentation_evaluator() + return st.error(f"Unsupported task for evaluator: {task}") diff --git a/tabs/inference.py b/tabs/inference.py index 4932d1d6..8e406f16 100644 --- a/tabs/inference.py +++ b/tabs/inference.py @@ -1,13 +1,22 @@ import streamlit as st - +from tabs.tasks.image_detection.inference import render_image_detection_inference +from tabs.tasks.image_segmentation.inference import render_image_segmentation_inference +from tabs.tasks.lidar_segmentation.inference import render_lidar_segmentation_inference def inference_tab(): task = st.session_state.get("task", "Image Detection") if task == "Image Detection": - from tabs.tasks.image_detection import render_image_detection_inference - render_image_detection_inference() return + + if task == "Image Segmentation": + render_image_segmentation_inference() + return + + if task == "Lidar Segmentation": + render_lidar_segmentation_inference() + return + st.error(f"Unsupported task for inference: {task}") diff --git a/tabs/tasks/__init__.py b/tabs/tasks/__init__.py deleted file mode 100644 index 8b137891..00000000 --- a/tabs/tasks/__init__.py +++ /dev/null @@ -1 +0,0 @@ - diff --git a/tabs/tasks/image_detection/__init__.py b/tabs/tasks/image_detection/__init__.py deleted file mode 100644 index 07cd5ef9..00000000 --- a/tabs/tasks/image_detection/__init__.py +++ /dev/null @@ -1,26 +0,0 @@ -__all__ = [ - "render_image_detection_sidebar", - "render_image_detection_viewer", - "render_image_detection_inference", - "render_image_detection_evaluator", -] - - -def __getattr__(name): - if name == "render_image_detection_sidebar": - from tabs.tasks.image_detection.sidebar import render_image_detection_sidebar - - return render_image_detection_sidebar - if name == "render_image_detection_viewer": - from tabs.tasks.image_detection.dataset_viewer import render_image_detection_viewer - - return render_image_detection_viewer - if name == "render_image_detection_inference": - from tabs.tasks.image_detection.inference import render_image_detection_inference - - return render_image_detection_inference - if name == "render_image_detection_evaluator": - from tabs.tasks.image_detection.evaluator import render_image_detection_evaluator - - return render_image_detection_evaluator - raise AttributeError(f"module {__name__!r} has no attribute {name!r}") diff --git a/tabs/tasks/image_detection/dataset_viewer.py b/tabs/tasks/image_detection/dataset_viewer.py index d198b714..c3a65e84 100644 --- a/tabs/tasks/image_detection/dataset_viewer.py +++ b/tabs/tasks/image_detection/dataset_viewer.py @@ -1,30 +1,33 @@ -import os -import tempfile - import streamlit as st -from PIL import Image +import os +from streamlit_image_select import image_select -from perceptionmetrics.datasets.coco import CocoDataset, find_img_dir_and_ann_file +from perceptionmetrics.datasets.coco import find_img_dir_and_ann_file def render_image_detection_viewer(): - import numpy as np + import tempfile + from perceptionmetrics.datasets.coco import CocoDataset from perceptionmetrics.datasets.yolo import YOLODataset - from streamlit_image_select import image_select + import numpy as np + from PIL import Image + from supervision.draw.color import ColorPalette from supervision.detection.annotate import BoxAnnotator from supervision.detection.core import Detections - from supervision.draw.color import ColorPalette + # Get inputs from session state dataset_path = st.session_state.get("dataset_path", "") dataset_type = st.session_state.get("dataset_type", "COCO").lower() split = st.session_state.get("split", "val") + # Header row only st.header("Dataset Viewer") if not dataset_path or not os.path.isdir(dataset_path): - st.warning("Please select a valid dataset folder.") + st.warning("⚠️ Please select a valid dataset folder.") return + # Setup paths and pagination if dataset_type == "coco": try: img_dir, ann_file = find_img_dir_and_ann_file( @@ -33,6 +36,7 @@ def render_image_detection_viewer(): except FileNotFoundError: st.warning("Dataset files not found. Check path and split.") return + elif dataset_type == "yolo": dataset_config_file = st.session_state.get("dataset_config_file", None) img_dir = os.path.join(dataset_path, f"images/{split}") @@ -46,92 +50,16 @@ def render_image_detection_viewer(): st.error("Unsupported dataset type.") return - _render_detection_viewer_style() - dataset = _load_detection_dataset(dataset_path, dataset_type, split, img_dir, ann_file if dataset_type == "coco" else None) - if dataset is None: - return - - image_files = [ - f for f in os.listdir(img_dir) if f.lower().endswith((".jpg", ".jpeg", ".png")) - ] - if not image_files: - st.warning("No images found.") - return - - image_paths, sample_images, current_page = _render_image_grid_navigation( - dataset_path, split, image_files, img_dir - ) - captions = [_short_caption(name) for name in sample_images] - - img_select_key = f"img_select_all_{dataset_path}_{split}_{current_page}" - img_select_index = st.session_state.get(img_select_key) - if img_select_index is None or not isinstance(img_select_index, int): - img_select_index = 0 - selected_img_path = ( - image_select( - label="", - images=image_paths, - captions=captions, - use_container_width=False, - key=img_select_key, - index=img_select_index, - ) - if image_paths - else None - ) - - if selected_img_path: - selected_img_name = os.path.basename(selected_img_path) - try: - img = Image.open(selected_img_path).convert("RGB") - img_np = np.array(img) - - if dataset_type == "yolo": - ann_row = dataset.dataset[ - dataset.dataset["image"].str.endswith(selected_img_name) - ] - else: - ann_row = dataset.dataset[dataset.dataset["image"] == selected_img_name] - - if not ann_row.empty: - annotation_id = ann_row.iloc[0]["annotation"] - if dataset_type == "yolo": - annotation_id = os.path.join(dataset.dataset_dir, annotation_id) - - boxes, category_indices = dataset.read_annotation(annotation_id) - class_names = _class_names_from_ontology(dataset, category_indices) - - palette = ColorPalette.default() - detections = Detections( - xyxy=np.array(boxes), class_id=np.array(category_indices) - ) - annotator = BoxAnnotator( - color=palette, text_scale=0.7, text_thickness=1, text_padding=2 - ) - annotated_img = annotator.annotate( - scene=img_np, detections=detections, labels=class_names - ) - - annotated_pil = Image.fromarray(annotated_img) - resample = getattr(getattr(Image, "Resampling", Image), "LANCZOS") - annotated_pil.thumbnail((500, 500), resample) - st.image(annotated_pil, width="content") - else: - st.warning("No annotation found for this image.") - except Exception as e: - st.error(f"Error displaying image: {e}") - else: - st.info("Select an image to view with annotations.") - -def _render_detection_viewer_style(): + # Pagination and search row nav_col1, nav_col2, nav_col3, nav_col4 = st.columns([1, 1, 2, 1.5]) with nav_col1: - pass + pass # Placeholder for "< page" button, to be added later in the code with nav_col2: - pass + pass # Placeholder for ">" button, to be added later in the code with nav_col3: - pass + pass # Placeholder for page info, to be added later in the code with nav_col4: + # Move the button up by reducing the margin and decrease button size with custom CSS st.markdown( """ - """, - unsafe_allow_html=True, - ) - st.markdown("
", unsafe_allow_html=True) - # Load dataset dataset_key = f"{dataset_path}_{split}" if dataset_key not in st.session_state: @@ -155,102 +125,13 @@ def render_image_detection_viewer(): st.warning("No images found.") return - # Pagination - IMAGES_PER_PAGE = 12 - _, total_pages = ( - len(image_files), - (len(image_files) + IMAGES_PER_PAGE - 1) // IMAGES_PER_PAGE, - ) - page_key = f"image_page_{dataset_path}_{split}" - - if page_key not in st.session_state: - st.session_state[page_key] = 0 - current_page = max(0, min(st.session_state[page_key], total_pages - 1)) - st.session_state[page_key] = current_page - - start_idx = current_page * IMAGES_PER_PAGE - sample_images = image_files[start_idx : start_idx + IMAGES_PER_PAGE] - image_paths = [os.path.join(img_dir, img_name) for img_name in sample_images] - - # Navigation - col1, col2, col3, col4 = st.columns([0.5, 9.5, 0.5, 0.5]) - with col1: - if st.button("⟨", key="prev_page_btn", disabled=(current_page == 0)): - st.session_state[page_key] = current_page - 1 - st.rerun() - with col2: - st.markdown( - f"
Page {current_page + 1} of {total_pages}
", - unsafe_allow_html=True, - ) - with col3: - if st.button( - "⟩", key="next_page_btn", disabled=(current_page >= total_pages - 1) - ): - st.session_state[page_key] = current_page + 1 - st.rerun() - with col4: - if st.button( - "🔍", - key="search_icon_btn", - help="Search for an image by name", - disabled=not (dataset_path and os.path.isdir(dataset_path)), - ): - st.session_state["show_search_dropdown"] = True - - # Search dropdown - if st.session_state.get("show_search_dropdown", False): - col1, col2, col3 = st.columns([4, 1, 1]) - with col1: - selected_img = st.selectbox( - "Search image:", options=image_files, key="search_image" - ) - with col2: - st.markdown( - "
", unsafe_allow_html=True - ) - if st.button("Go to image", key="go_to_image_btn"): - new_page = image_files.index(selected_img) // IMAGES_PER_PAGE - st.session_state[page_key] = new_page - st.session_state[ - f"img_select_all_{dataset_path}_{split}_{new_page}" - ] = (image_files.index(selected_img) % IMAGES_PER_PAGE) - st.session_state["show_search_dropdown"] = False - st.rerun() - with col3: - st.markdown( - "
", unsafe_allow_html=True - ) - if st.button("Cancel", key="cancel_search_btn"): - st.session_state["show_search_dropdown"] = False - st.rerun() - - caption_len_limit = 17 - captions = [ - ( - (name[:caption_len_limit] + "..." + name[-3:]) - if len(name) > caption_len_limit - else name - ) - for name in sample_images - ] - - # Image grid - img_select_key = f"img_select_all_{dataset_path}_{split}_{current_page}" - img_select_index = st.session_state.get(img_select_key) - if img_select_index is None or not isinstance(img_select_index, int): - img_select_index = 0 - selected_img_path = ( - image_select( - label="", - images=image_paths, - captions=captions, - use_container_width=False, - key=img_select_key, - index=img_select_index, - ) - if image_paths - else None + image_paths = [os.path.join(img_dir, img_name) for img_name in image_files] + selected_img_path, _ = render_image_grid( + item_names=image_files, + image_paths=image_paths, + state_prefix="image_detection", + context=f"{dataset_path}_{split}", + search_label="image", ) # Display selected image with annotations diff --git a/tabs/tasks/image_detection/evaluator.py b/tabs/tasks/image_detection/evaluator.py index 67a533fc..e4e4fdf0 100644 --- a/tabs/tasks/image_detection/evaluator.py +++ b/tabs/tasks/image_detection/evaluator.py @@ -5,7 +5,7 @@ from perceptionmetrics.datasets.coco import CocoDataset -from perceptionmetrics.utils.gui import browse_folder +from tabs.tasks.utils import browse_folder from perceptionmetrics.datasets.coco import find_img_dir_and_ann_file diff --git a/tabs/tasks/image_detection/sidebar.py b/tabs/tasks/image_detection/sidebar.py index 85efea37..d20e5124 100644 --- a/tabs/tasks/image_detection/sidebar.py +++ b/tabs/tasks/image_detection/sidebar.py @@ -5,7 +5,7 @@ import streamlit as st -from perceptionmetrics.utils.gui import browse_folder +from tabs.tasks.utils import browse_folder def browse_dataset_path(): @@ -326,4 +326,3 @@ def _uploaded_model_to_tempfile(uploaded_file) -> Optional[str]: except Exception as e: st.error(f"Failed to save model file: {e}") return None - diff --git a/tabs/tasks/image_segmentation/dataset_viewer.py b/tabs/tasks/image_segmentation/dataset_viewer.py index f2109a22..23f0e93f 100644 --- a/tabs/tasks/image_segmentation/dataset_viewer.py +++ b/tabs/tasks/image_segmentation/dataset_viewer.py @@ -6,6 +6,7 @@ from PIL import Image from perceptionmetrics.datasets.cityscapes import CityscapesImageSegmentationDataset +from tabs.tasks.utils import render_image_grid def render_image_segmentation_viewer(): @@ -39,22 +40,28 @@ def _render_cityscapes_viewer(): st.warning(f"No Cityscapes samples found for split '{split}'.") return - col1, col2 = st.columns([3, 1]) - with col1: - sample_name = st.selectbox( - "Sample", - split_df.index.tolist(), - key=f"cityscapes_segmentation_sample_{split}", - ) - with col2: - mask_opacity = st.slider( - "Mask Opacity", - min_value=0.0, - max_value=1.0, - value=0.45, - step=0.05, - key="cityscapes_segmentation_mask_opacity", - ) + sample_names = split_df.index.tolist() + image_paths = split_df["image"].tolist() + selected_img_path, sample_name = render_image_grid( + item_names=sample_names, + image_paths=image_paths, + state_prefix="cityscapes_segmentation", + context=f"{dataset_path}_{split}", + search_label="sample", + ) + + if not selected_img_path: + st.info("Select an image to view the ground truth mask.") + return + + mask_opacity = st.slider( + "Mask Opacity", + min_value=0.0, + max_value=1.0, + value=0.45, + step=0.05, + key="cityscapes_segmentation_mask_opacity", + ) row = split_df.loc[sample_name] image_fname = row["image"] diff --git a/tabs/tasks/image_segmentation/evaluator.py b/tabs/tasks/image_segmentation/evaluator.py index e749ef29..6221dcab 100644 --- a/tabs/tasks/image_segmentation/evaluator.py +++ b/tabs/tasks/image_segmentation/evaluator.py @@ -1,2 +1,228 @@ +import json +import os +import tempfile + +import streamlit as st + +from perceptionmetrics.datasets.cityscapes import CityscapesImageSegmentationDataset +from tabs.tasks.utils import browse_folder + + +def browse_segmentation_predictions_outdir(): + folder = browse_folder() + if folder: + st.session_state.segmentation_predictions_outdir = folder + + def render_image_segmentation_evaluator(): - return + st.header("Evaluator") + st.markdown("Evaluate your model on the loaded dataset using PerceptionMetrics.") + + dataset_type = st.session_state.get("segmentation_dataset_type", "Cityscapes") + if dataset_type != "Cityscapes": + st.info(f"{dataset_type} image segmentation evaluation is not wired yet.") + return + + dataset = None + model = st.session_state.get("segmentation_model") + dataset_path = st.session_state.get("dataset_path", "") + split = st.session_state.get("split", "val") + + if not dataset_path or not os.path.isdir(dataset_path): + st.warning( + "No dataset path provided. Please set the dataset path in the sidebar." + ) + else: + try: + roots = {"train": None, "val": None, "test": None} + roots[split] = dataset_path + dataset_key = ( + "cityscapes_segmentation_evaluation_dataset", + os.path.abspath(dataset_path), + split, + st.session_state.get( + "segmentation_image_dir", + "leftImg8bit_trainvaltest/leftImg8bit", + ), + st.session_state.get("segmentation_label_dir", "gtFine"), + st.session_state.get( + "segmentation_image_suffix", "_leftImg8bit.png" + ), + st.session_state.get( + "segmentation_label_suffix", "_gtFine_labelIds.png" + ), + st.session_state.get("segmentation_use_train_id", False), + ) + + if dataset_key not in st.session_state: + st.session_state[dataset_key] = CityscapesImageSegmentationDataset( + train_dataset_root=roots["train"], + val_dataset_root=roots["val"], + test_dataset_root=roots["test"], + image_dir=dataset_key[3], + label_dir=dataset_key[4], + image_suffix=dataset_key[5], + label_suffix=dataset_key[6], + use_train_id=dataset_key[7], + ) + + dataset = st.session_state[dataset_key] + st.success( + f"✅ Dataset loaded: {dataset_path} ({split} split) - {len(dataset.dataset)} samples" + ) + except Exception as exc: + st.error(f"Error loading dataset: {exc}") + + if model is not None: + st.success("✅ Model loaded and ready for evaluation") + else: + st.warning( + "No model loaded. Please load a model using the " + "'Load Segmentation Model' button in the sidebar." + ) + + st.markdown("### Evaluation Configuration") + + save_predictions = st.checkbox( + "Save Predictions", + value=False, + help="Save predicted label images to an output directory.", + key="segmentation_save_predictions", + ) + + predictions_outdir = None + if save_predictions: + col1, col2 = st.columns([3, 1]) + with col1: + st.text_input( + "Predictions Output Directory", + key="segmentation_predictions_outdir", + ) + with col2: + st.markdown( + "
", + unsafe_allow_html=True, + ) + st.button( + "Browse", + on_click=browse_segmentation_predictions_outdir, + key="browse_segmentation_predictions_outdir", + ) + predictions_outdir = st.session_state.get("segmentation_predictions_outdir") + + ontology_translation = st.file_uploader( + "Ontology Translation (Optional)", + type=["json"], + key="segmentation_ontology_translation", + help="JSON file for translating between dataset and model ontologies.", + ) + st.info( + "For Cityscapes SegFormer models, use train-ID labels or provide a " + "label-ID to train-ID ontology translation." + ) + + translation_direction = st.selectbox( + "Translation Direction", + ["dataset_to_model", "model_to_dataset"], + key="segmentation_translation_direction", + help=( + "dataset_to_model maps GT labels to model IDs. " + "model_to_dataset maps predictions to dataset IDs." + ), + ) + + output_dir_missing = save_predictions and not ( + predictions_outdir and predictions_outdir.strip() + ) + if output_dir_missing: + st.warning("Please provide a predictions output directory.") + + if st.button( + "🚀 Run Evaluation", + type="primary", + disabled=dataset is None or model is None or output_dir_missing, + key="run_segmentation_evaluation", + ): + with st.spinner("Running evaluation..."): + try: + ontology_translation_path = None + if ontology_translation is not None: + with tempfile.NamedTemporaryFile( + delete=False, suffix=".json", mode="w" + ) as tmp_trans: + json.dump(json.load(ontology_translation), tmp_trans) + ontology_translation_path = tmp_trans.name + + predictions_outdir = predictions_outdir.strip() if ( + save_predictions and predictions_outdir + ) else None + if predictions_outdir is not None: + os.makedirs(predictions_outdir, exist_ok=True) + + progress_bar = st.progress(0) + status_text = st.empty() + intermediate_metrics_placeholder = st.empty() + + def progress_callback(processed, total): + progress = processed / total if total > 0 else 0 + progress_bar.progress(progress) + status_text.text( + f"Processing: {processed}/{total} images ({progress:.1%})" + ) + + def metrics_callback(metrics_df, processed, total): + with intermediate_metrics_placeholder.container(): + st.markdown( + f"#### Results (after {processed}/{total} images)" + ) + display_segmentation_evaluation_results( + metrics_df, show_download=False + ) + + results = model.eval( + dataset=dataset, + split=split, + ontology_translation=ontology_translation_path, + translation_direction=translation_direction, + predictions_outdir=predictions_outdir, + results_per_sample=save_predictions, + progress_callback=progress_callback, + metrics_callback=metrics_callback, + ) + + progress_bar.empty() + status_text.empty() + intermediate_metrics_placeholder.empty() + + st.session_state["segmentation_evaluation_results"] = results + st.success("✅ Evaluation completed successfully!") + except Exception as exc: + st.error(f"Error in model.eval(): {exc}") + + if "segmentation_evaluation_results" in st.session_state: + display_segmentation_evaluation_results( + st.session_state["segmentation_evaluation_results"] + ) + + +def display_segmentation_evaluation_results(results, show_download=True): + if results is None or results.empty: + st.warning("No evaluation results to display.") + return + + + st.markdown("#### Metrics") + display_df = results.copy() + numeric_columns = display_df.select_dtypes(include=["float64", "int64"]).columns + for col in numeric_columns: + display_df[col] = display_df[col].round(3) + st.dataframe(display_df, width="stretch") + + if show_download: + csv = results.to_csv(index=True) + st.download_button( + label="📥 Download segmentation metrics", + data=csv, + file_name="segmentation_evaluation_results.csv", + mime="text/csv", + ) diff --git a/tabs/tasks/image_segmentation/inference.py b/tabs/tasks/image_segmentation/inference.py index e3bd3aaa..ebdad937 100644 --- a/tabs/tasks/image_segmentation/inference.py +++ b/tabs/tasks/image_segmentation/inference.py @@ -1,2 +1,56 @@ +import numpy as np +import streamlit as st +from PIL import Image + +from perceptionmetrics.utils import conversion as uc + + def render_image_segmentation_inference(): - return + st.header("Model Inference") + st.markdown("Select an image and run inference using the loaded model.") + + model = st.session_state.get("segmentation_model") + if model is None: + st.warning("Load a segmentation model from the sidebar to start inference.") + return + + image_file = st.file_uploader( + "Choose an image", + type=["jpg", "jpeg", "png"], + key="segmentation_inference_image_file", + help="Upload an image to run segmentation inference.", + ) + + if image_file is None: + return + + with st.spinner("Running segmentation inference..."): + try: + image = Image.open(image_file).convert("RGB") + pred = model.predict(image) + except Exception as exc: + st.error(f"Failed to run inference: {exc}") + return + + pred = uc.label_to_rgb(pred, model.ontology) + pred = pred.resize(image.size) + prediction_overlay = _overlay_mask(image, pred, opacity=0.45) + + cols = st.columns(3) + with cols[0]: + st.image(image, caption="Image", use_container_width=True) + with cols[1]: + st.image(pred, caption="Prediction", use_container_width=True) + with cols[2]: + st.image( + prediction_overlay, + caption="Prediction Overlay", + use_container_width=True, + ) + + +def _overlay_mask(image, mask_rgb, opacity): + image_np = np.array(image) + mask_np = np.array(mask_rgb) + overlay = ((1.0 - opacity) * image_np + opacity * mask_np).astype(np.uint8) + return Image.fromarray(overlay) diff --git a/tabs/tasks/image_segmentation/sidebar.py b/tabs/tasks/image_segmentation/sidebar.py index 91c8a36c..141bea56 100644 --- a/tabs/tasks/image_segmentation/sidebar.py +++ b/tabs/tasks/image_segmentation/sidebar.py @@ -1,6 +1,8 @@ +import os + import streamlit as st -from perceptionmetrics.utils.gui import browse_folder +from tabs.tasks.utils import browse_file, browse_folder IMAGE_SEGMENTATION_DATASETS = [ @@ -17,6 +19,21 @@ def browse_dataset_path(): st.session_state.dataset_path = browse_folder() +def browse_segmentation_model_path(): + if st.session_state.get("segmentation_model_type") == "Hugging Face SegFormer": + st.session_state.segmentation_model_path = browse_folder() + else: + st.session_state.segmentation_model_path = browse_file() + + +def browse_segmentation_config_path(): + st.session_state.segmentation_config_path = browse_file() + + +def browse_segmentation_ontology_path(): + st.session_state.segmentation_ontology_path = browse_file() + + def render_image_segmentation_sidebar(_available_devices): with st.expander("Image Segmentation Dataset", expanded=True): col1, col2 = st.columns(2) @@ -40,21 +57,37 @@ def render_image_segmentation_sidebar(_available_devices): st.button("Browse", on_click=browse_dataset_path) if dataset_type == "Cityscapes": - _render_cityscapes_dataset_inputs() + render_cityscapes_dataset_inputs() else: st.info(f"{dataset_type} image segmentation inputs are not wired yet.") with st.expander("Image Segmentation Model", expanded=False): - st.info("Image segmentation model loading is a placeholder for now.") + if st.session_state.get("segmentation_dataset_type", "Cityscapes") != "Cityscapes": + st.info("Image segmentation model loading is wired for Cityscapes first.") + return + render_segmentation_model_inputs() -def _render_cityscapes_dataset_inputs(): + if st.button( + "Load Segmentation Model", + type="primary", + width="stretch", + key="sidebar_load_segmentation_model_btn", + ): + load_image_segmentation_model() + + +def render_cityscapes_dataset_inputs(): st.text_input( "Image Directory", value="leftImg8bit_trainvaltest/leftImg8bit", key="segmentation_image_dir", ) - st.text_input("Label Directory", value="gtFine", key="segmentation_label_dir") + st.text_input( + "Label Directory", + value="gtFine", + key="segmentation_label_dir", + ) st.text_input( "Image Suffix", value="_leftImg8bit.png", @@ -71,3 +104,126 @@ def _render_cityscapes_dataset_inputs(): key="segmentation_use_train_id", help="Enable when labels use _gtFine_labelTrainIds.png.", ) + + +def render_segmentation_model_inputs(): + model_type = st.selectbox( + "Model Type", + ["Torch Model File", "Hugging Face SegFormer"], + key="segmentation_model_type", + ) + + col1, col2 = st.columns([3, 1]) + with col1: + st.text_input( + "Model Path" if model_type == "Torch Model File" else "Model Name or Folder", + key="segmentation_model_path", + help=( + "Path to a TorchScript model or saved PyTorch model file." + if model_type == "Torch Model File" + else "Hugging Face model name or local folder downloaded with save_pretrained." + ), + ) + with col2: + st.markdown( + "
", + unsafe_allow_html=True, + ) + st.button( + "Browse", + on_click=browse_segmentation_model_path, + key="browse_segmentation_model_path", + ) + + col1, col2 = st.columns([3, 1]) + with col1: + st.text_input( + "Config File", + key="segmentation_config_path", + help="JSON model configuration file.", + ) + with col2: + st.markdown( + "
", + unsafe_allow_html=True, + ) + st.button( + "Browse", + on_click=browse_segmentation_config_path, + key="browse_segmentation_config_path", + ) + + col1, col2 = st.columns([3, 1]) + with col1: + st.text_input( + "Ontology File", + key="segmentation_ontology_path", + help="JSON file containing the model output ontology.", + ) + with col2: + st.markdown( + "
", + unsafe_allow_html=True, + ) + st.button( + "Browse", + on_click=browse_segmentation_ontology_path, + key="browse_segmentation_ontology_path", + ) + + +def load_image_segmentation_model(): + from perceptionmetrics.models.torch_segmentation import TorchImageSegmentationModel + + model_type = st.session_state.get("segmentation_model_type", "Torch Model File") + model_path = st.session_state.get("segmentation_model_path", "") + config_path = st.session_state.get("segmentation_config_path", "") + ontology_path = st.session_state.get("segmentation_ontology_path", "") + + if not model_path: + st.error("Please provide a model path or model name.") + return + if not config_path or not os.path.isfile(config_path): + st.error("Please provide a valid config JSON path.") + return + if not ontology_path or not os.path.isfile(ontology_path): + st.error("Please provide a valid ontology JSON path.") + return + + with st.spinner("Loading image segmentation model..."): + try: + model = load_model_for_type(model_type, model_path) + segmentation_model = TorchImageSegmentationModel( + model=model, + model_cfg=config_path, + ontology_fname=ontology_path, + ) + st.session_state.segmentation_model = segmentation_model + st.session_state.segmentation_model_loaded = True + st.success("Segmentation model loaded and saved for inference") + except Exception as exc: + st.session_state.segmentation_model = None + st.session_state.segmentation_model_loaded = False + st.error(f"Failed to load segmentation model: {exc}") + + +def load_model_for_type(model_type, model_path): + if model_type == "Torch Model File": + if not os.path.isfile(model_path): + raise ValueError( + "Torch Model File expects a .pt/.pth/.torchscript file saved with " + "torch.save or torch.jit.save. For a Hugging Face model folder, " + "select 'Hugging Face SegFormer'." + ) + return model_path + + if model_type == "Hugging Face SegFormer": + try: + from transformers import SegformerForSemanticSegmentation + except ImportError as exc: + raise ImportError( + "transformers is required for Hugging Face SegFormer models." + ) from exc + return SegformerForSemanticSegmentation.from_pretrained(model_path) + + raise ValueError(f"Unsupported segmentation model type: {model_type}") diff --git a/tabs/tasks/utils.py b/tabs/tasks/utils.py new file mode 100644 index 00000000..d14f213e --- /dev/null +++ b/tabs/tasks/utils.py @@ -0,0 +1,261 @@ +import platform +import subprocess +import sys + +import streamlit as st +from streamlit_image_select import image_select + + +def is_wsl(): + """ + Detect if running in Windows Subsystem for Linux (WSL). + Returns True if WSL is detected, False otherwise. + """ + return ( + "wsl" in platform.release().lower() or "microsoft" in platform.release().lower() + ) + + +def browse_folder(): + """ + Opens a native folder selection dialog and returns the selected folder path. + Works on Windows, macOS, and Linux (with zenity or kdialog). + Returns None if cancelled or error. + """ + try: + is_windows = sys.platform.startswith("win") + is_wsl_env = is_wsl() + if is_windows or is_wsl_env: + script = ( + "Add-Type -AssemblyName System.windows.forms;" + "$f=New-Object System.Windows.Forms.FolderBrowserDialog;" + 'if($f.ShowDialog() -eq "OK"){Write-Output $f.SelectedPath}' + ) + result = subprocess.run( + ["powershell.exe", "-NoProfile", "-Command", script], + capture_output=True, + text=True, + timeout=30, + ) + folder = result.stdout.strip() + if folder and is_wsl_env: + result = subprocess.run( + ["wslpath", "-u", folder], + capture_output=True, + text=True, + timeout=30, + ) + folder = result.stdout.strip() + return folder if folder else None + elif sys.platform == "darwin": + script = 'POSIX path of (choose folder with prompt "Select folder:")' + result = subprocess.run( + ["osascript", "-e", script], capture_output=True, text=True, timeout=30 + ) + folder = result.stdout.strip() + return folder if folder else None + else: + for cmd in [ + [ + "zenity", + "--file-selection", + "--directory", + "--title=Select folder", + ], + [ + "kdialog", + "--getexistingdirectory", + "--title", + "Select folder", + ], + ]: + try: + result = subprocess.run( + cmd, capture_output=True, text=True, timeout=30 + ) + if result.returncode == 0 or result.returncode == 1: + folder = result.stdout.strip() + return folder if folder else None + except subprocess.TimeoutExpired: + return None + except (FileNotFoundError, Exception): + continue + return None + except Exception: + return None + + +def browse_file(): + """ + Opens a native file selection dialog and returns the selected file path. + Works on Windows, macOS, and Linux (with zenity or kdialog). + Returns None if cancelled or error. + """ + try: + is_windows = sys.platform.startswith("win") + is_wsl_env = is_wsl() + if is_windows or is_wsl_env: + script = ( + "Add-Type -AssemblyName System.windows.forms;" + "$f=New-Object System.Windows.Forms.OpenFileDialog;" + 'if($f.ShowDialog() -eq "OK"){Write-Output $f.FileName}' + ) + result = subprocess.run( + ["powershell.exe", "-NoProfile", "-Command", script], + capture_output=True, + text=True, + timeout=30, + ) + file_path = result.stdout.strip() + if file_path and is_wsl_env: + result = subprocess.run( + ["wslpath", "-u", file_path], + capture_output=True, + text=True, + timeout=30, + ) + file_path = result.stdout.strip() + return file_path if file_path else None + elif sys.platform == "darwin": + script = 'POSIX path of (choose file with prompt "Select file:")' + result = subprocess.run( + ["osascript", "-e", script], capture_output=True, text=True, timeout=30 + ) + file_path = result.stdout.strip() + return file_path if file_path else None + else: + for cmd in [ + ["zenity", "--file-selection", "--title=Select file"], + ["kdialog", "--getopenfilename", "--title", "Select file"], + ]: + try: + result = subprocess.run( + cmd, capture_output=True, text=True, timeout=30 + ) + if result.returncode == 0 or result.returncode == 1: + file_path = result.stdout.strip() + return file_path if file_path else None + except subprocess.TimeoutExpired: + return None + except (FileNotFoundError, Exception): + continue + return None + except Exception: + return None + + +def render_image_grid( + item_names, + image_paths, + state_prefix, + context, + search_label="image", + images_per_page=12, +): + total_pages = (len(item_names) + images_per_page - 1) // images_per_page + page_key = f"{state_prefix}_page_{context}" + + if page_key not in st.session_state: + st.session_state[page_key] = 0 + + current_page = max(0, min(st.session_state[page_key], total_pages - 1)) + st.session_state[page_key] = current_page + + start_idx = current_page * images_per_page + page_item_names = item_names[start_idx : start_idx + images_per_page] + page_image_paths = image_paths[start_idx : start_idx + images_per_page] + + col1, col2, col3, col4 = st.columns([0.5, 9.5, 0.5, 0.5]) + with col1: + if st.button( + "⟨", + key=f"{state_prefix}_prev_page_btn", + disabled=(current_page == 0), + ): + st.session_state[page_key] = current_page - 1 + st.rerun() + with col2: + st.markdown( + f"
Page {current_page + 1} of {total_pages}
", + unsafe_allow_html=True, + ) + with col3: + if st.button( + "⟩", + key=f"{state_prefix}_next_page_btn", + disabled=(current_page >= total_pages - 1), + ): + st.session_state[page_key] = current_page + 1 + st.rerun() + with col4: + if st.button( + "🔍", + key=f"{state_prefix}_search_icon_btn", + help=f"Search for a {search_label} by name", + ): + st.session_state[f"show_{state_prefix}_search"] = True + + if st.session_state.get(f"show_{state_prefix}_search", False): + col1, col2, col3 = st.columns([4, 1, 1]) + with col1: + selected_item = st.selectbox( + f"Search {search_label}:", + options=item_names, + key=f"{state_prefix}_search_item", + ) + with col2: + st.markdown( + "
", + unsafe_allow_html=True, + ) + if st.button(f"Go to {search_label}", key=f"{state_prefix}_go_to_item"): + selected_idx = item_names.index(selected_item) + new_page = selected_idx // images_per_page + st.session_state[page_key] = new_page + st.session_state[f"{state_prefix}_select_{context}_{new_page}"] = ( + selected_idx % images_per_page + ) + st.session_state[f"show_{state_prefix}_search"] = False + st.rerun() + with col3: + st.markdown( + "
", + unsafe_allow_html=True, + ) + if st.button("Cancel", key=f"{state_prefix}_cancel_search"): + st.session_state[f"show_{state_prefix}_search"] = False + st.rerun() + + caption_len_limit = 17 + captions = [ + ( + (name[:caption_len_limit] + "..." + name[-3:]) + if len(name) > caption_len_limit + else name + ) + for name in page_item_names + ] + + select_key = f"{state_prefix}_select_{context}_{current_page}" + select_index = st.session_state.get(select_key) + if select_index is None or not isinstance(select_index, int): + select_index = 0 + + selected_image_path = ( + image_select( + label="", + images=page_image_paths, + captions=captions, + use_container_width=False, + key=select_key, + index=select_index, + ) + if page_image_paths + else None + ) + + if not selected_image_path: + return None, None + + selected_index = page_image_paths.index(selected_image_path) + return selected_image_path, page_item_names[selected_index] From de2cc39c1d426b8df730820bdc5c666884a93b51 Mon Sep 17 00:00:00 2001 From: tejass Date: Wed, 1 Jul 2026 18:53:30 +0200 Subject: [PATCH 04/10] Restore image segmentation tutorial from master --- examples/tutorial_image_segmentation.ipynb | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/examples/tutorial_image_segmentation.ipynb b/examples/tutorial_image_segmentation.ipynb index a31eadf3..6de4877b 100644 --- a/examples/tutorial_image_segmentation.ipynb +++ b/examples/tutorial_image_segmentation.ipynb @@ -1135,7 +1135,7 @@ "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", - "version": "3.11.14" + "version": "3.10.19" } }, "nbformat": 4, From d0f66306545213dabf3cd4db52039055d1a4014d Mon Sep 17 00:00:00 2001 From: tejass Date: Wed, 1 Jul 2026 19:05:24 +0200 Subject: [PATCH 05/10] Rebasing tutorials --- examples/tutorial_image_detection.ipynb | 57 +- examples/tutorial_nuimages_segmentation.ipynb | 1757 +---------------- 2 files changed, 86 insertions(+), 1728 deletions(-) diff --git a/examples/tutorial_image_detection.ipynb b/examples/tutorial_image_detection.ipynb index d4f106d5..97f8a13a 100644 --- a/examples/tutorial_image_detection.ipynb +++ b/examples/tutorial_image_detection.ipynb @@ -2,30 +2,9 @@ "cells": [ { "cell_type": "code", - "execution_count": 1, + "execution_count": null, "metadata": {}, - "outputs": [ - { - "name": "stdout", - "output_type": "stream", - "text": [ - "Jupyter environment detected. Enabling Open3D WebVisualizer.\n", - "[Open3D INFO] WebRTC GUI backend enabled.\n", - "[Open3D INFO] WebRTCWindowSystem: HTTP handshake server disabled.\n", - "Tensorflow not available\n" - ] - }, - { - "name": "stderr", - "output_type": "stream", - "text": [ - "/home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages/torchvision/datapoints/__init__.py:12: UserWarning: The torchvision.datapoints and torchvision.transforms.v2 namespaces are still Beta. While we do not expect major breaking changes, some APIs may still change according to user feedback. Please submit any feedback you may have in this issue: https://github.com/pytorch/vision/issues/6753, and you can also check out https://github.com/pytorch/vision/issues/7319 to learn more about the APIs that we suspect might involve future changes. You can silence this warning by calling torchvision.disable_beta_transforms_warning().\n", - " warnings.warn(_BETA_TRANSFORMS_WARNING)\n", - "/home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages/torchvision/transforms/v2/__init__.py:54: UserWarning: The torchvision.datapoints and torchvision.transforms.v2 namespaces are still Beta. While we do not expect major breaking changes, some APIs may still change according to user feedback. Please submit any feedback you may have in this issue: https://github.com/pytorch/vision/issues/6753, and you can also check out https://github.com/pytorch/vision/issues/7319 to learn more about the APIs that we suspect might involve future changes. You can silence this warning by calling torchvision.disable_beta_transforms_warning().\n", - " warnings.warn(_BETA_TRANSFORMS_WARNING)\n" - ] - } - ], + "outputs": [], "source": [ "# Import required libraries\n", "import torch\n", @@ -59,7 +38,7 @@ }, { "cell_type": "code", - "execution_count": 3, + "execution_count": null, "metadata": {}, "outputs": [ { @@ -67,7 +46,7 @@ "output_type": "stream", "text": [ "loading annotations into memory...\n", - "Done (t=0.40s)\n", + "Done (t=0.39s)\n", "creating index...\n", "index created!\n", "Dataset loaded with 5000 samples\n", @@ -79,8 +58,8 @@ "# Initialize COCO dataset\n", "# Update these placeholders with your local COCO paths before running\n", "\n", - "img_dir = \"/home/tejass/Downloads/TUDELFT_ROBOTICS/GSOC/gsoc2026-Tejas_Stanley/PerceptionMetrics/examples/local/data/coco/images/val2017\"\n", - "ann_file = \"/home/tejass/Downloads/TUDELFT_ROBOTICS/GSOC/gsoc2026-Tejas_Stanley/PerceptionMetrics/examples/local/data/coco/annotations/instances_val2017.json\"\n", + "img_dir = \"\"\n", + "ann_file = \"\"\n", "\n", "dataset = None\n", "\n", @@ -101,14 +80,14 @@ }, { "cell_type": "code", - "execution_count": 4, + "execution_count": null, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ - "Model, configuration, and dataset ontology saved!\n" + "Model and configuration saved!\n" ] } ], @@ -117,19 +96,11 @@ "model = torchvision.models.detection.maskrcnn_resnet50_fpn(weights=\"DEFAULT\")\n", "model.eval()\n", "\n", - "if dataset is None:\n", - " raise RuntimeError(\n", - " \"Dataset is not loaded. Set img_dir and ann_file to valid COCO paths \"\n", - " \"and run the dataset initialization cell before saving the model ontology.\"\n", - " )\n", - "\n", "# Save the model\n", "model_path = \"local/data/models/maskrcnn_model.pt\"\n", "os.makedirs(os.path.dirname(model_path), exist_ok=True)\n", "torch.save(model, model_path)\n", - "\n", "model_cfg = {\n", - " \"model_format\": \"torchvision\",\n", " \"batch_size\": 1,\n", " \"num_workers\": 0,\n", " \"confidence_threshold\": 0.8,\n", @@ -139,13 +110,12 @@ "with open(config_path, \"w\") as f:\n", " json.dump(model_cfg, f, indent=2)\n", "\n", - "# Use the dataset ontology so evaluation uses the same COCO category IDs\n", - "# as the loaded annotations.\n", + "\n", "ontology_path = \"local/data/models/coco_model_ontology.json\"\n", "with open(ontology_path, \"w\") as f:\n", " json.dump(dataset.ontology, f, indent=2)\n", "\n", - "print(\"Model, configuration, and dataset ontology saved!\")\n" + "print(\"Model and configuration saved!\")" ] }, { @@ -157,8 +127,7 @@ "name": "stdout", "output_type": "stream", "text": [ - "Open3D-ML3D not available\n", - "'resize_cfg' missing in model config. No resizing will be applied.\n", + "Model is not a TorchScript model. Loading as native PyTorch model.\n", "Detection model initialized!\n" ] } @@ -773,7 +742,7 @@ ], "metadata": { "kernelspec": { - "display_name": "pm", + "display_name": ".venv", "language": "python", "name": "python3" }, @@ -787,7 +756,7 @@ "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", - "version": "3.11.14" + "version": "3.10.4" } }, "nbformat": 4, diff --git a/examples/tutorial_nuimages_segmentation.ipynb b/examples/tutorial_nuimages_segmentation.ipynb index f2b435ec..68086277 100644 --- a/examples/tutorial_nuimages_segmentation.ipynb +++ b/examples/tutorial_nuimages_segmentation.ipynb @@ -1,112 +1,47 @@ { "cells": [ - { - "cell_type": "markdown", - "id": "eeb1b612", - "metadata": {}, - "source": [ - "# NuImages Segmentation Evaluation Tutorial\n" - ] - }, { "cell_type": "code", - "execution_count": 1, + "execution_count": null, "id": "478f413c", "metadata": {}, - "outputs": [ - { - "name": "stdout", - "output_type": "stream", - "text": [ - "Requirement already satisfied: transformers in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (4.30.2)\n", - "Requirement already satisfied: filelock in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from transformers) (3.14.0)\n", - "Requirement already satisfied: huggingface-hub<1.0,>=0.14.1 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from transformers) (0.36.2)\n", - "Requirement already satisfied: numpy>=1.17 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from transformers) (1.26.4)\n", - "Requirement already satisfied: packaging>=20.0 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from transformers) (24.2)\n", - "Requirement already satisfied: pyyaml>=5.1 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from transformers) (6.0.3)\n", - "Requirement already satisfied: regex!=2019.12.17 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from transformers) (2026.5.9)\n", - "Requirement already satisfied: requests in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from transformers) (2.28.2)\n", - "Requirement already satisfied: tokenizers!=0.11.3,<0.14,>=0.11.1 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from transformers) (0.13.3)\n", - "Requirement already satisfied: safetensors>=0.3.1 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from transformers) (0.7.0)\n", - "Requirement already satisfied: tqdm>=4.27 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from transformers) (4.65.2)\n", - "Requirement already satisfied: fsspec>=2023.5.0 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from huggingface-hub<1.0,>=0.14.1->transformers) (2025.12.0)\n", - "Requirement already satisfied: hf-xet<2.0.0,>=1.1.3 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from huggingface-hub<1.0,>=0.14.1->transformers) (1.5.0)\n", - "Requirement already satisfied: typing-extensions>=3.7.4.3 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from huggingface-hub<1.0,>=0.14.1->transformers) (4.15.0)\n", - "Requirement already satisfied: charset-normalizer<4,>=2 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from requests->transformers) (3.4.7)\n", - "Requirement already satisfied: idna<4,>=2.5 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from requests->transformers) (3.16)\n", - "Requirement already satisfied: urllib3<1.27,>=1.21.1 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from requests->transformers) (1.26.20)\n", - "Requirement already satisfied: certifi>=2017.4.17 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from requests->transformers) (2026.5.20)\n" - ] - } - ], + "outputs": [], "source": [ - "!pip install transformers" + "!pip install gdown" ] }, { - "cell_type": "markdown", - "id": "13273f08", + "cell_type": "code", + "execution_count": null, + "id": "a37f27c1", "metadata": {}, + "outputs": [], "source": [ - "## Download model from Hugging Face" + "# Download\n", + "!gdown 15Py3AKyGzWNX4M3OJKlQQsnH6IehkoEd -O local/data/image_segmentation_model.pth # model\n", + "!gdown 17eJ6aei6yBAbK1aA4mhTeMdXYycorqcT -O local/data/image_segmentation_model_cfg.json # configuration\n", + "!gdown 1spXt5_ISG1ZaHHO2DTtCSqgkohnhwuzt -O local/data/image_segmentation_model_ontology.json # ontology" ] }, { "cell_type": "code", - "execution_count": 2, + "execution_count": null, "id": "2a47f7a9", "metadata": {}, - "outputs": [ - { - "name": "stderr", - "output_type": "stream", - "text": [ - "/home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages/huggingface_hub/file_download.py:949: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.\n", - " warnings.warn(\n", - "/home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages/transformers/models/segformer/image_processing_segformer.py:99: FutureWarning: The `reduce_labels` parameter is deprecated and will be removed in a future version. Please use `do_reduce_labels` instead.\n", - " warnings.warn(\n" - ] - } - ], + "outputs": [], "source": [ - "import json\n", - "from transformers import AutoImageProcessor\n", - "\n", - "model_name = \"tejasstanley/segformer-nuimages-b0-full-bs8\"\n", - "\n", - "processor = AutoImageProcessor.from_pretrained(model_name)\n", - "cfg = processor.to_dict()\n", - "\n", - "pm_cfg = {\n", - " \"resize\": {\n", - " \"height\": cfg[\"size\"][\"height\"],\n", - " \"width\": cfg[\"size\"][\"width\"],\n", - " },\n", - " \"normalization\": {\n", - " \"mean\": list(cfg[\"image_mean\"]),\n", - " \"std\": list(cfg[\"image_std\"]),\n", - " },\n", - " \"ignored_classes\": [\"background\", \"vehicle.ego\"],\n", - "}\n", - "\n", - "model_cfg_path = \"local/data/segformer_nuimages_b0_full_bs8_cfg.json\"\n", + "from perceptionmetrics.models.torch_segmentation import TorchImageSegmentationModel\n", "\n", - "with open(model_cfg_path, \"w\", encoding=\"utf-8\") as f:\n", - " json.dump(pm_cfg, f, indent=2)" - ] - }, - { - "cell_type": "markdown", - "id": "f7f0b0e0", - "metadata": {}, - "source": [ - "## Build Dataset and Model Ontology\n", - "\n" + "model = TorchImageSegmentationModel(\n", + " model=\"local/data/image_segmentation_model.pth\", \n", + " model_cfg=\"local/data/image_segmentation_model_cfg.json\",\n", + " ontology_fname=\"local/data/image_segmentation_model_ontology.json\",\n", + ")" ] }, { "cell_type": "code", - "execution_count": 3, + "execution_count": 5, "id": "444ef811", "metadata": {}, "outputs": [ @@ -114,149 +49,60 @@ "name": "stdout", "output_type": "stream", "text": [ - "Jupyter environment detected. Enabling Open3D WebVisualizer.\n", - "[Open3D INFO] WebRTC GUI backend enabled.\n", - "[Open3D INFO] WebRTCWindowSystem: HTTP handshake server disabled.\n", "Built nuImages segmentation dataset with 50 samples and 26 classes.\n", - "Saved model ontology to local/data/nuimages_model_ontology.json\n", - "Dataset labels will be translated in memory during evaluation.\n" + "Saved compact ontology to local/data/nuimages_segmentation_ontology_compact.json with 26 classes (max idx=25)\n" ] } ], "source": [ + "## NuImages Segmentation Dataset and Model\n", + "\n", "import json\n", "from pathlib import Path\n", "\n", + "from perceptionmetrics.models.torch_segmentation import TorchImageSegmentationModel\n", "from perceptionmetrics.datasets.nuimages import NuImagesSegmentationDataset\n", "\n", "dataset = NuImagesSegmentationDataset(\n", - " dataset_dir=\"local/data/nuimages-v1.0-mini\", version=\"v1.0-mini\", split=\"val\"\n", + " dataset_dir=\"local/data/nuimages-v1.0-mini\", version=\"v1.0-mini\", split=\"train\"\n", ")\n", "\n", - "model_labels = {\n", - " 0: \"animal\",\n", - " 1: \"human.pedestrian.adult\",\n", - " 2: \"human.pedestrian.child\",\n", - " 3: \"human.pedestrian.construction_worker\",\n", - " 4: \"human.pedestrian.personal_mobility\",\n", - " 5: \"human.pedestrian.police_officer\",\n", - " 6: \"human.pedestrian.stroller\",\n", - " 7: \"human.pedestrian.wheelchair\",\n", - " 8: \"movable_object.barrier\",\n", - " 9: \"movable_object.debris\",\n", - " 10: \"movable_object.pushable_pullable\",\n", - " 11: \"movable_object.trafficcone\",\n", - " 12: \"static_object.bicycle_rack\",\n", - " 13: \"vehicle.bicycle\",\n", - " 14: \"vehicle.bus.bendy\",\n", - " 15: \"vehicle.bus.rigid\",\n", - " 16: \"vehicle.car\",\n", - " 17: \"vehicle.construction\",\n", - " 18: \"vehicle.emergency.ambulance\",\n", - " 19: \"vehicle.emergency.police\",\n", - " 20: \"vehicle.motorcycle\",\n", - " 21: \"vehicle.trailer\",\n", - " 22: \"vehicle.truck\",\n", - " 23: \"flat.driveable_surface\",\n", - "}\n", - "\n", - "model_ontology = {\n", + "# Build compact ontology (contiguous idx: 0..N-1) to avoid sparse-index issues (e.g., idx 31).\n", + "sorted_items = sorted(dataset.ontology.items(), key=lambda x: x[1][\"idx\"] )\n", + "compact_ontology = {\n", " class_name: {\n", - " \"idx\": class_idx,\n", - " \"rgb\": dataset.ontology[class_name][\"rgb\"],\n", + " \"idx\": new_idx,\n", + " \"rgb\": class_meta.get(\"rgb\", [0, 0, 0]),\n", " }\n", - " for class_idx, class_name in model_labels.items()\n", + " for new_idx, (class_name, class_meta) in enumerate(sorted_items)\n", "}\n", "\n", - "model_ontology_path = Path(\"local/data/nuimages_model_ontology.json\")\n", - "model_ontology_path.parent.mkdir(parents=True, exist_ok=True)\n", - "with open(model_ontology_path, \"w\", encoding=\"utf-8\") as f:\n", - " json.dump(model_ontology, f, indent=2)\n", + "ontology_path = Path(\"local/data/nuimages_segmentation_ontology_compact.json\")\n", + "ontology_path.parent.mkdir(parents=True, exist_ok=True)\n", + "with open(ontology_path, \"w\") as f:\n", + " json.dump(compact_ontology, f, indent=2)\n", "\n", - "print(f\"Saved model ontology to {model_ontology_path}\")\n", - "print(\"Dataset labels will be translated in memory during evaluation.\")\n" - ] - }, - { - "cell_type": "markdown", - "id": "4e68ee53", - "metadata": {}, - "source": [ - "## Initialize the PerceptionMetrics Model Wrapper\n", - "\n" - ] - }, - { - "cell_type": "code", - "execution_count": 4, - "id": "e34db9f2", - "metadata": {}, - "outputs": [ - { - "name": "stdout", - "output_type": "stream", - "text": [ - "Torch detection not available\n", - "Tensorflow not available\n" - ] - }, - { - "name": "stderr", - "output_type": "stream", - "text": [ - "/home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages/torchvision/datapoints/__init__.py:12: UserWarning: The torchvision.datapoints and torchvision.transforms.v2 namespaces are still Beta. While we do not expect major breaking changes, some APIs may still change according to user feedback. Please submit any feedback you may have in this issue: https://github.com/pytorch/vision/issues/6753, and you can also check out https://github.com/pytorch/vision/issues/7319 to learn more about the APIs that we suspect might involve future changes. You can silence this warning by calling torchvision.disable_beta_transforms_warning().\n", - " warnings.warn(_BETA_TRANSFORMS_WARNING)\n", - "/home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages/torchvision/transforms/v2/__init__.py:54: UserWarning: The torchvision.datapoints and torchvision.transforms.v2 namespaces are still Beta. While we do not expect major breaking changes, some APIs may still change according to user feedback. Please submit any feedback you may have in this issue: https://github.com/pytorch/vision/issues/6753, and you can also check out https://github.com/pytorch/vision/issues/7319 to learn more about the APIs that we suspect might involve future changes. You can silence this warning by calling torchvision.disable_beta_transforms_warning().\n", - " warnings.warn(_BETA_TRANSFORMS_WARNING)\n", - "/home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages/huggingface_hub/file_download.py:949: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.\n", - " warnings.warn(\n" - ] - } - ], - "source": [ - "from transformers import SegformerForSemanticSegmentation\n", - "from perceptionmetrics.models.torch_segmentation import TorchImageSegmentationModel\n", - "\n", - "model_name = \"tejasstanley/segformer-nuimages-b0-full-bs8\"\n", - "model_cfg_path = \"local/data/segformer_nuimages_b0_full_bs8_cfg.json\"\n", - "\n", - "\n", - "hf_model = SegformerForSemanticSegmentation.from_pretrained(model_name)\n", + "max_idx = max(v[\"idx\"] for v in compact_ontology.values())\n", + "print(\n", + " f\"Saved compact ontology to {ontology_path} with {len(compact_ontology)} classes (max idx={max_idx})\"\n", + " )\n", "\n", - "pm_model = TorchImageSegmentationModel(\n", - " model=hf_model,\n", - " model_cfg=model_cfg_path,\n", - " ontology_fname=str(model_ontology_path),\n", + "model = TorchImageSegmentationModel(\n", + " model=\"local/data/image_segmentation_model.pth\",\n", + " model_cfg=\"local/data/image_segmentation_model_cfg.json\",\n", + " ontology_fname=str(ontology_path),\n", ")" ] }, - { - "cell_type": "markdown", - "id": "16b24b8a", - "metadata": {}, - "source": [ - "## Inspect a Sample Prediction\n" - ] - }, { "cell_type": "code", - "execution_count": 5, + "execution_count": 6, "id": "47f1727b", "metadata": {}, "outputs": [ - { - "name": "stdout", - "output_type": "stream", - "text": [ - "Predicted class ids: [1, 3, 8, 11, 13, 16, 21, 22, 23]\n", - "Predicted class names: ['human.pedestrian.adult', 'human.pedestrian.construction_worker', 'movable_object.barrier', 'movable_object.trafficcone', 'vehicle.bicycle', 'vehicle.car', 'vehicle.trailer', 'vehicle.truck', 'flat.driveable_surface']\n", - "Dataset label ids in sample: [0, 4, 9, 17, 23, 24]\n", - "Dataset label names in sample: ['background', 'human.pedestrian.construction_worker', 'movable_object.barrier', 'vehicle.car', 'vehicle.truck', 'flat.driveable_surface']\n" - ] - }, { "data": { - "image/png": "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", + "image/png": "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", "text/plain": [ "
" ] @@ -272,43 +118,15 @@ "\n", "from perceptionmetrics.utils import conversion as uc\n", "\n", - "image_fname = dataset.dataset[\"image\"].iloc[1]\n", - "if dataset.dataset_dir is not None:\n", - " image_fname = Path(dataset.dataset_dir) / image_fname\n", + "image_fname = dataset.dataset[\"image\"].iloc[0]\n", "image = Image.open(image_fname)\n", "\n", - "label_fname = dataset.dataset[\"label\"].iloc[1]\n", - "if dataset.dataset_dir is not None:\n", - " label_fname = Path(dataset.dataset_dir) / label_fname\n", + "label_fname = dataset.dataset[\"label\"].iloc[0]\n", "label = Image.open(label_fname)\n", "label = uc.label_to_rgb(label, dataset.ontology)\n", "\n", - "raw_pred = pm_model.predict(image)\n", - "raw_pred_np = np.array(raw_pred)\n", - "raw_label_np = np.array(Image.open(label_fname))\n", - "\n", - "model_idx_to_name = {\n", - " class_meta[\"idx\"]: class_name for class_name, class_meta in pm_model.ontology.items()\n", - "}\n", - "dataset_idx_to_name = {\n", - " class_meta[\"idx\"]: class_name for class_name, class_meta in dataset.ontology.items()\n", - "}\n", - "\n", - "pred_ids = sorted(np.unique(raw_pred_np).tolist())\n", - "label_ids = sorted(np.unique(raw_label_np).tolist())\n", - "\n", - "print(\"Predicted class ids:\", pred_ids)\n", - "print(\n", - " \"Predicted class names:\",\n", - " [model_idx_to_name.get(class_id, f\"unknown:{class_id}\") for class_id in pred_ids],\n", - ")\n", - "print(\"Dataset label ids in sample:\", label_ids)\n", - "print(\n", - " \"Dataset label names in sample:\",\n", - " [dataset_idx_to_name.get(class_id, f\"unknown:{class_id}\") for class_id in label_ids],\n", - ")\n", - "\n", - "pred = uc.label_to_rgb(raw_pred, pm_model.ontology)\n", + "pred = model.predict(image)\n", + "pred = uc.label_to_rgb(pred, model.ontology)\n", "pred = pred.resize(label.size)\n", "\n", "plt.figure(figsize=(10, 10))\n", @@ -318,1474 +136,45 @@ "plt.show()" ] }, - { - "cell_type": "markdown", - "id": "bcb40b82", - "metadata": {}, - "source": [ - "## Translate Dataset Labels and Run Evaluation\n", - "\n", - "The evaluation call maps ground-truth labels from the nuImages ontology into the model ontology with `translation_direction=\"dataset_to_model\"`.\n" - ] - }, { "cell_type": "code", - "execution_count": 6, + "execution_count": null, "id": "a1f6ca6b", "metadata": {}, - "outputs": [ - { - "name": "stderr", - "output_type": "stream", - "text": [ - "100%|██████████| 50/50 [00:01<00:00, 35.81it/s]\n", - "/home/tejass/Downloads/TUDELFT_ROBOTICS/GSOC/gsoc2026-Tejas_Stanley/PerceptionMetrics/perceptionmetrics/utils/segmentation_metrics.py:147: RuntimeWarning: invalid value encountered in divide\n", - " return np.where(denominator > 0, tp / denominator, np.nan)\n", - "/home/tejass/Downloads/TUDELFT_ROBOTICS/GSOC/gsoc2026-Tejas_Stanley/PerceptionMetrics/perceptionmetrics/utils/segmentation_metrics.py:164: RuntimeWarning: invalid value encountered in divide\n", - " return np.where(denominator > 0, tp / denominator, np.nan)\n", - "/home/tejass/Downloads/TUDELFT_ROBOTICS/GSOC/gsoc2026-Tejas_Stanley/PerceptionMetrics/perceptionmetrics/utils/segmentation_metrics.py:203: RuntimeWarning: invalid value encountered in divide\n", - " denominator > 0, 2 * (precision * recall) / denominator, np.nan\n", - "/home/tejass/Downloads/TUDELFT_ROBOTICS/GSOC/gsoc2026-Tejas_Stanley/PerceptionMetrics/perceptionmetrics/utils/segmentation_metrics.py:222: RuntimeWarning: invalid value encountered in divide\n", - " return np.where(union > 0, tp / union, np.nan)\n", - "/home/tejass/Downloads/TUDELFT_ROBOTICS/GSOC/gsoc2026-Tejas_Stanley/PerceptionMetrics/perceptionmetrics/utils/segmentation_metrics.py:240: RuntimeWarning: invalid value encountered in divide\n", - " return np.where(denominator > 0, 2 * tp / denominator, np.nan)\n" - ] - }, - { - "data": { - "text/html": [ - "
\n", - "\n", - "\n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - "
animalhuman.pedestrian.adulthuman.pedestrian.childhuman.pedestrian.construction_workerhuman.pedestrian.personal_mobilityhuman.pedestrian.police_officerhuman.pedestrian.strollerhuman.pedestrian.wheelchairmovable_object.barriermovable_object.debris...vehicle.carvehicle.constructionvehicle.emergency.ambulancevehicle.emergency.policevehicle.motorcyclevehicle.trailervehicle.truckflat.driveable_surfacemacromicro
tp0.04.578900e+040.02.280000e+020.000000e+000.00.00.07.052100e+040.000000e+00...3.026940e+051.243870e+050.00.02.299100e+046.436000e+037.031500e+042.559201e+06NaNNaN
fp0.06.403000e+030.02.690000e+020.000000e+000.00.00.03.562000e+031.000000e+00...1.643800e+043.742000e+030.00.03.728000e+035.490000e+022.202200e+048.568000e+03NaNNaN
fn0.05.353000e+030.01.494000e+031.254000e+030.00.00.03.064000e+031.146000e+03...1.231400e+043.680000e+030.00.01.579000e+031.489800e+049.345000e+039.131000e+03NaNNaN
tn3326172.03.268627e+063326172.03.324181e+063.324918e+063326172.03326172.03326172.03.249025e+063.325025e+06...2.994726e+063.194363e+063326172.03326172.03.297874e+063.304289e+063.224490e+067.492720e+05NaNNaN
precisionNaN8.773184e-01NaN4.587525e-01NaNNaNNaNNaN9.519188e-010.000000e+00...9.484915e-019.707951e-01NaNNaN8.604738e-019.214030e-017.615041e-019.966633e-010.7892790.978781
recallNaN8.953306e-01NaN1.324042e-010.000000e+00NaNNaNNaN9.583611e-010.000000e+00...9.609089e-019.712650e-01NaNNaN9.357346e-013.016781e-018.826889e-019.964448e-010.6760580.978781
accuracy1.09.964656e-011.09.994700e-019.996230e-011.01.01.09.980079e-019.996552e-01...9.913558e-019.977686e-011.01.09.984045e-019.953559e-019.905696e-019.946789e-010.9982320.998232
f1_scoreNaN8.862330e-01NaN2.054980e-01NaNNaNNaNNaN9.551291e-01NaN...9.546599e-019.710300e-01NaNNaN8.965275e-014.545358e-018.176305e-019.965540e-010.7891150.978781
iouNaN7.957077e-01NaN1.145153e-010.000000e+00NaNNaNNaN9.141120e-010.000000e+00...9.132528e-019.436913e-01NaNNaN8.124602e-012.941096e-016.915187e-019.931317e-010.6177490.958443
dice_scoreNaN8.862330e-01NaN2.054980e-010.000000e+00NaNNaNNaN9.551291e-010.000000e+00...9.546599e-019.710300e-01NaNNaN8.965275e-014.545358e-018.176305e-019.965540e-010.6904760.978781
animal0.00.000000e+000.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...0.000000e+000.000000e+000.00.00.000000e+000.000000e+000.000000e+000.000000e+00NaNNaN
human.pedestrian.adult0.04.578900e+040.09.000000e+014.470000e+020.00.00.05.200000e+019.140000e+02...1.235000e+032.140000e+020.00.06.370000e+020.000000e+006.580000e+021.288000e+03NaNNaN
human.pedestrian.child0.00.000000e+000.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...0.000000e+000.000000e+000.00.00.000000e+000.000000e+000.000000e+000.000000e+00NaNNaN
human.pedestrian.construction_worker0.02.390000e+020.02.280000e+020.000000e+000.00.00.02.000000e+000.000000e+00...0.000000e+002.000000e+010.00.00.000000e+000.000000e+008.000000e+000.000000e+00NaNNaN
human.pedestrian.personal_mobility0.00.000000e+000.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...0.000000e+000.000000e+000.00.00.000000e+000.000000e+000.000000e+000.000000e+00NaNNaN
human.pedestrian.police_officer0.00.000000e+000.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...0.000000e+000.000000e+000.00.00.000000e+000.000000e+000.000000e+000.000000e+00NaNNaN
human.pedestrian.stroller0.00.000000e+000.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...0.000000e+000.000000e+000.00.00.000000e+000.000000e+000.000000e+000.000000e+00NaNNaN
human.pedestrian.wheelchair0.00.000000e+000.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...0.000000e+000.000000e+000.00.00.000000e+000.000000e+000.000000e+000.000000e+00NaNNaN
movable_object.barrier0.03.130000e+020.01.500000e+010.000000e+000.00.00.07.052100e+046.000000e+00...3.000000e+007.490000e+020.00.00.000000e+000.000000e+003.480000e+021.260000e+03NaNNaN
movable_object.debris0.00.000000e+000.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...0.000000e+000.000000e+000.00.00.000000e+000.000000e+000.000000e+000.000000e+00NaNNaN
movable_object.pushable_pullable0.01.100000e+020.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...0.000000e+000.000000e+000.00.00.000000e+000.000000e+000.000000e+000.000000e+00NaNNaN
movable_object.trafficcone0.09.600000e+010.00.000000e+000.000000e+000.00.00.02.320000e+020.000000e+00...2.000000e+019.700000e+010.00.00.000000e+004.600000e+011.360000e+024.700000e+02NaNNaN
static_object.bicycle_rack0.00.000000e+000.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...0.000000e+000.000000e+000.00.04.000000e+010.000000e+000.000000e+000.000000e+00NaNNaN
vehicle.bicycle0.02.070000e+020.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...0.000000e+000.000000e+000.00.00.000000e+000.000000e+000.000000e+000.000000e+00NaNNaN
vehicle.bus.bendy0.00.000000e+000.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...0.000000e+000.000000e+000.00.00.000000e+000.000000e+000.000000e+000.000000e+00NaNNaN
vehicle.bus.rigid0.00.000000e+000.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...1.400000e+028.200000e+010.00.00.000000e+000.000000e+001.290000e+021.900000e+02NaNNaN
vehicle.car0.01.746000e+030.05.300000e+010.000000e+000.00.00.02.890000e+020.000000e+00...3.026940e+055.070000e+020.00.07.370000e+021.821000e+036.235000e+034.619000e+03NaNNaN
vehicle.construction0.03.890000e+020.06.820000e+020.000000e+000.00.00.08.730000e+020.000000e+00...8.930000e+021.243870e+050.00.00.000000e+000.000000e+008.140000e+029.100000e+01NaNNaN
vehicle.emergency.ambulance0.00.000000e+000.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...0.000000e+000.000000e+000.00.00.000000e+000.000000e+000.000000e+000.000000e+00NaNNaN
vehicle.emergency.police0.00.000000e+000.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...0.000000e+000.000000e+000.00.00.000000e+000.000000e+000.000000e+000.000000e+00NaNNaN
vehicle.motorcycle0.06.670000e+020.02.220000e+027.190000e+020.00.00.01.760000e+020.000000e+00...1.369000e+031.490000e+020.00.02.299100e+040.000000e+000.000000e+004.260000e+02NaNNaN
vehicle.trailer0.00.000000e+000.00.000000e+000.000000e+000.00.00.06.000000e+010.000000e+00...0.000000e+002.900000e+010.00.00.000000e+006.436000e+034.100000e+020.000000e+00NaNNaN
vehicle.truck0.05.660000e+020.03.350000e+026.600000e+010.00.00.04.200000e+010.000000e+00...5.173000e+031.542000e+030.00.04.000000e+001.282800e+047.031500e+047.870000e+02NaNNaN
flat.driveable_surface0.01.020000e+030.09.700000e+012.200000e+010.00.00.01.338000e+032.260000e+02...3.481000e+032.910000e+020.00.01.610000e+022.030000e+026.070000e+022.559201e+06NaNNaN
\n", - "

34 rows × 26 columns

\n", - "
" - ], - "text/plain": [ - " animal human.pedestrian.adult \\\n", - "tp 0.0 4.578900e+04 \n", - "fp 0.0 6.403000e+03 \n", - "fn 0.0 5.353000e+03 \n", - "tn 3326172.0 3.268627e+06 \n", - "precision NaN 8.773184e-01 \n", - "recall NaN 8.953306e-01 \n", - "accuracy 1.0 9.964656e-01 \n", - "f1_score NaN 8.862330e-01 \n", - "iou NaN 7.957077e-01 \n", - "dice_score NaN 8.862330e-01 \n", - "animal 0.0 0.000000e+00 \n", - "human.pedestrian.adult 0.0 4.578900e+04 \n", - "human.pedestrian.child 0.0 0.000000e+00 \n", - "human.pedestrian.construction_worker 0.0 2.390000e+02 \n", - "human.pedestrian.personal_mobility 0.0 0.000000e+00 \n", - "human.pedestrian.police_officer 0.0 0.000000e+00 \n", - "human.pedestrian.stroller 0.0 0.000000e+00 \n", - "human.pedestrian.wheelchair 0.0 0.000000e+00 \n", - "movable_object.barrier 0.0 3.130000e+02 \n", - "movable_object.debris 0.0 0.000000e+00 \n", - "movable_object.pushable_pullable 0.0 1.100000e+02 \n", - "movable_object.trafficcone 0.0 9.600000e+01 \n", - "static_object.bicycle_rack 0.0 0.000000e+00 \n", - "vehicle.bicycle 0.0 2.070000e+02 \n", - "vehicle.bus.bendy 0.0 0.000000e+00 \n", - "vehicle.bus.rigid 0.0 0.000000e+00 \n", - "vehicle.car 0.0 1.746000e+03 \n", - "vehicle.construction 0.0 3.890000e+02 \n", - "vehicle.emergency.ambulance 0.0 0.000000e+00 \n", - "vehicle.emergency.police 0.0 0.000000e+00 \n", - "vehicle.motorcycle 0.0 6.670000e+02 \n", - "vehicle.trailer 0.0 0.000000e+00 \n", - "vehicle.truck 0.0 5.660000e+02 \n", - "flat.driveable_surface 0.0 1.020000e+03 \n", - "\n", - " human.pedestrian.child \\\n", - "tp 0.0 \n", - "fp 0.0 \n", - "fn 0.0 \n", - "tn 3326172.0 \n", - "precision NaN \n", - "recall NaN \n", - "accuracy 1.0 \n", - "f1_score NaN \n", - "iou NaN \n", - "dice_score NaN \n", - "animal 0.0 \n", - "human.pedestrian.adult 0.0 \n", - "human.pedestrian.child 0.0 \n", - "human.pedestrian.construction_worker 0.0 \n", - "human.pedestrian.personal_mobility 0.0 \n", - "human.pedestrian.police_officer 0.0 \n", - "human.pedestrian.stroller 0.0 \n", - "human.pedestrian.wheelchair 0.0 \n", - "movable_object.barrier 0.0 \n", - "movable_object.debris 0.0 \n", - "movable_object.pushable_pullable 0.0 \n", - "movable_object.trafficcone 0.0 \n", - "static_object.bicycle_rack 0.0 \n", - "vehicle.bicycle 0.0 \n", - "vehicle.bus.bendy 0.0 \n", - "vehicle.bus.rigid 0.0 \n", - "vehicle.car 0.0 \n", - "vehicle.construction 0.0 \n", - "vehicle.emergency.ambulance 0.0 \n", - "vehicle.emergency.police 0.0 \n", - "vehicle.motorcycle 0.0 \n", - "vehicle.trailer 0.0 \n", - "vehicle.truck 0.0 \n", - "flat.driveable_surface 0.0 \n", - "\n", - " human.pedestrian.construction_worker \\\n", - "tp 2.280000e+02 \n", - "fp 2.690000e+02 \n", - "fn 1.494000e+03 \n", - "tn 3.324181e+06 \n", - "precision 4.587525e-01 \n", - "recall 1.324042e-01 \n", - "accuracy 9.994700e-01 \n", - "f1_score 2.054980e-01 \n", - "iou 1.145153e-01 \n", - "dice_score 2.054980e-01 \n", - "animal 0.000000e+00 \n", - "human.pedestrian.adult 9.000000e+01 \n", - "human.pedestrian.child 0.000000e+00 \n", - "human.pedestrian.construction_worker 2.280000e+02 \n", - "human.pedestrian.personal_mobility 0.000000e+00 \n", - "human.pedestrian.police_officer 0.000000e+00 \n", - "human.pedestrian.stroller 0.000000e+00 \n", - "human.pedestrian.wheelchair 0.000000e+00 \n", - "movable_object.barrier 1.500000e+01 \n", - "movable_object.debris 0.000000e+00 \n", - "movable_object.pushable_pullable 0.000000e+00 \n", - "movable_object.trafficcone 0.000000e+00 \n", - "static_object.bicycle_rack 0.000000e+00 \n", - "vehicle.bicycle 0.000000e+00 \n", - "vehicle.bus.bendy 0.000000e+00 \n", - "vehicle.bus.rigid 0.000000e+00 \n", - "vehicle.car 5.300000e+01 \n", - "vehicle.construction 6.820000e+02 \n", - "vehicle.emergency.ambulance 0.000000e+00 \n", - "vehicle.emergency.police 0.000000e+00 \n", - "vehicle.motorcycle 2.220000e+02 \n", - "vehicle.trailer 0.000000e+00 \n", - "vehicle.truck 3.350000e+02 \n", - "flat.driveable_surface 9.700000e+01 \n", - "\n", - " human.pedestrian.personal_mobility \\\n", - "tp 0.000000e+00 \n", - "fp 0.000000e+00 \n", - "fn 1.254000e+03 \n", - "tn 3.324918e+06 \n", - "precision NaN \n", - "recall 0.000000e+00 \n", - "accuracy 9.996230e-01 \n", - "f1_score NaN \n", - "iou 0.000000e+00 \n", - "dice_score 0.000000e+00 \n", - "animal 0.000000e+00 \n", - "human.pedestrian.adult 4.470000e+02 \n", - "human.pedestrian.child 0.000000e+00 \n", - "human.pedestrian.construction_worker 0.000000e+00 \n", - "human.pedestrian.personal_mobility 0.000000e+00 \n", - "human.pedestrian.police_officer 0.000000e+00 \n", - "human.pedestrian.stroller 0.000000e+00 \n", - "human.pedestrian.wheelchair 0.000000e+00 \n", - "movable_object.barrier 0.000000e+00 \n", - "movable_object.debris 0.000000e+00 \n", - "movable_object.pushable_pullable 0.000000e+00 \n", - "movable_object.trafficcone 0.000000e+00 \n", - "static_object.bicycle_rack 0.000000e+00 \n", - "vehicle.bicycle 0.000000e+00 \n", - "vehicle.bus.bendy 0.000000e+00 \n", - "vehicle.bus.rigid 0.000000e+00 \n", - "vehicle.car 0.000000e+00 \n", - "vehicle.construction 0.000000e+00 \n", - "vehicle.emergency.ambulance 0.000000e+00 \n", - "vehicle.emergency.police 0.000000e+00 \n", - "vehicle.motorcycle 7.190000e+02 \n", - "vehicle.trailer 0.000000e+00 \n", - "vehicle.truck 6.600000e+01 \n", - "flat.driveable_surface 2.200000e+01 \n", - "\n", - " human.pedestrian.police_officer \\\n", - "tp 0.0 \n", - "fp 0.0 \n", - "fn 0.0 \n", - "tn 3326172.0 \n", - "precision NaN \n", - "recall NaN \n", - "accuracy 1.0 \n", - "f1_score NaN \n", - "iou NaN \n", - "dice_score NaN \n", - "animal 0.0 \n", - "human.pedestrian.adult 0.0 \n", - "human.pedestrian.child 0.0 \n", - "human.pedestrian.construction_worker 0.0 \n", - "human.pedestrian.personal_mobility 0.0 \n", - "human.pedestrian.police_officer 0.0 \n", - "human.pedestrian.stroller 0.0 \n", - "human.pedestrian.wheelchair 0.0 \n", - "movable_object.barrier 0.0 \n", - "movable_object.debris 0.0 \n", - "movable_object.pushable_pullable 0.0 \n", - "movable_object.trafficcone 0.0 \n", - "static_object.bicycle_rack 0.0 \n", - "vehicle.bicycle 0.0 \n", - "vehicle.bus.bendy 0.0 \n", - "vehicle.bus.rigid 0.0 \n", - "vehicle.car 0.0 \n", - "vehicle.construction 0.0 \n", - "vehicle.emergency.ambulance 0.0 \n", - "vehicle.emergency.police 0.0 \n", - "vehicle.motorcycle 0.0 \n", - "vehicle.trailer 0.0 \n", - "vehicle.truck 0.0 \n", - "flat.driveable_surface 0.0 \n", - "\n", - " human.pedestrian.stroller \\\n", - "tp 0.0 \n", - "fp 0.0 \n", - "fn 0.0 \n", - "tn 3326172.0 \n", - "precision NaN \n", - "recall NaN \n", - "accuracy 1.0 \n", - "f1_score NaN \n", - "iou NaN \n", - "dice_score NaN \n", - "animal 0.0 \n", - "human.pedestrian.adult 0.0 \n", - "human.pedestrian.child 0.0 \n", - "human.pedestrian.construction_worker 0.0 \n", - "human.pedestrian.personal_mobility 0.0 \n", - "human.pedestrian.police_officer 0.0 \n", - "human.pedestrian.stroller 0.0 \n", - "human.pedestrian.wheelchair 0.0 \n", - "movable_object.barrier 0.0 \n", - "movable_object.debris 0.0 \n", - "movable_object.pushable_pullable 0.0 \n", - "movable_object.trafficcone 0.0 \n", - "static_object.bicycle_rack 0.0 \n", - "vehicle.bicycle 0.0 \n", - "vehicle.bus.bendy 0.0 \n", - "vehicle.bus.rigid 0.0 \n", - "vehicle.car 0.0 \n", - "vehicle.construction 0.0 \n", - "vehicle.emergency.ambulance 0.0 \n", - "vehicle.emergency.police 0.0 \n", - "vehicle.motorcycle 0.0 \n", - "vehicle.trailer 0.0 \n", - "vehicle.truck 0.0 \n", - "flat.driveable_surface 0.0 \n", - "\n", - " human.pedestrian.wheelchair \\\n", - "tp 0.0 \n", - "fp 0.0 \n", - "fn 0.0 \n", - "tn 3326172.0 \n", - "precision NaN \n", - "recall NaN \n", - "accuracy 1.0 \n", - "f1_score NaN \n", - "iou NaN \n", - "dice_score NaN \n", - "animal 0.0 \n", - "human.pedestrian.adult 0.0 \n", - "human.pedestrian.child 0.0 \n", - "human.pedestrian.construction_worker 0.0 \n", - "human.pedestrian.personal_mobility 0.0 \n", - "human.pedestrian.police_officer 0.0 \n", - "human.pedestrian.stroller 0.0 \n", - "human.pedestrian.wheelchair 0.0 \n", - "movable_object.barrier 0.0 \n", - "movable_object.debris 0.0 \n", - "movable_object.pushable_pullable 0.0 \n", - "movable_object.trafficcone 0.0 \n", - "static_object.bicycle_rack 0.0 \n", - "vehicle.bicycle 0.0 \n", - "vehicle.bus.bendy 0.0 \n", - "vehicle.bus.rigid 0.0 \n", - "vehicle.car 0.0 \n", - "vehicle.construction 0.0 \n", - "vehicle.emergency.ambulance 0.0 \n", - "vehicle.emergency.police 0.0 \n", - "vehicle.motorcycle 0.0 \n", - "vehicle.trailer 0.0 \n", - "vehicle.truck 0.0 \n", - "flat.driveable_surface 0.0 \n", - "\n", - " movable_object.barrier \\\n", - "tp 7.052100e+04 \n", - "fp 3.562000e+03 \n", - "fn 3.064000e+03 \n", - "tn 3.249025e+06 \n", - "precision 9.519188e-01 \n", - "recall 9.583611e-01 \n", - "accuracy 9.980079e-01 \n", - "f1_score 9.551291e-01 \n", - "iou 9.141120e-01 \n", - "dice_score 9.551291e-01 \n", - "animal 0.000000e+00 \n", - "human.pedestrian.adult 5.200000e+01 \n", - "human.pedestrian.child 0.000000e+00 \n", - "human.pedestrian.construction_worker 2.000000e+00 \n", - "human.pedestrian.personal_mobility 0.000000e+00 \n", - "human.pedestrian.police_officer 0.000000e+00 \n", - "human.pedestrian.stroller 0.000000e+00 \n", - "human.pedestrian.wheelchair 0.000000e+00 \n", - "movable_object.barrier 7.052100e+04 \n", - "movable_object.debris 0.000000e+00 \n", - "movable_object.pushable_pullable 0.000000e+00 \n", - "movable_object.trafficcone 2.320000e+02 \n", - "static_object.bicycle_rack 0.000000e+00 \n", - "vehicle.bicycle 0.000000e+00 \n", - "vehicle.bus.bendy 0.000000e+00 \n", - "vehicle.bus.rigid 0.000000e+00 \n", - "vehicle.car 2.890000e+02 \n", - "vehicle.construction 8.730000e+02 \n", - "vehicle.emergency.ambulance 0.000000e+00 \n", - "vehicle.emergency.police 0.000000e+00 \n", - "vehicle.motorcycle 1.760000e+02 \n", - "vehicle.trailer 6.000000e+01 \n", - "vehicle.truck 4.200000e+01 \n", - "flat.driveable_surface 1.338000e+03 \n", - "\n", - " movable_object.debris ... \\\n", - "tp 0.000000e+00 ... \n", - "fp 1.000000e+00 ... \n", - "fn 1.146000e+03 ... \n", - "tn 3.325025e+06 ... \n", - "precision 0.000000e+00 ... \n", - "recall 0.000000e+00 ... \n", - "accuracy 9.996552e-01 ... \n", - "f1_score NaN ... \n", - "iou 0.000000e+00 ... \n", - "dice_score 0.000000e+00 ... \n", - "animal 0.000000e+00 ... \n", - "human.pedestrian.adult 9.140000e+02 ... \n", - "human.pedestrian.child 0.000000e+00 ... \n", - "human.pedestrian.construction_worker 0.000000e+00 ... \n", - "human.pedestrian.personal_mobility 0.000000e+00 ... \n", - "human.pedestrian.police_officer 0.000000e+00 ... \n", - "human.pedestrian.stroller 0.000000e+00 ... \n", - "human.pedestrian.wheelchair 0.000000e+00 ... \n", - "movable_object.barrier 6.000000e+00 ... \n", - "movable_object.debris 0.000000e+00 ... \n", - "movable_object.pushable_pullable 0.000000e+00 ... \n", - "movable_object.trafficcone 0.000000e+00 ... \n", - "static_object.bicycle_rack 0.000000e+00 ... \n", - "vehicle.bicycle 0.000000e+00 ... \n", - "vehicle.bus.bendy 0.000000e+00 ... \n", - "vehicle.bus.rigid 0.000000e+00 ... \n", - "vehicle.car 0.000000e+00 ... \n", - "vehicle.construction 0.000000e+00 ... \n", - "vehicle.emergency.ambulance 0.000000e+00 ... \n", - "vehicle.emergency.police 0.000000e+00 ... \n", - "vehicle.motorcycle 0.000000e+00 ... \n", - "vehicle.trailer 0.000000e+00 ... \n", - "vehicle.truck 0.000000e+00 ... \n", - "flat.driveable_surface 2.260000e+02 ... \n", - "\n", - " vehicle.car vehicle.construction \\\n", - "tp 3.026940e+05 1.243870e+05 \n", - "fp 1.643800e+04 3.742000e+03 \n", - "fn 1.231400e+04 3.680000e+03 \n", - "tn 2.994726e+06 3.194363e+06 \n", - "precision 9.484915e-01 9.707951e-01 \n", - "recall 9.609089e-01 9.712650e-01 \n", - "accuracy 9.913558e-01 9.977686e-01 \n", - "f1_score 9.546599e-01 9.710300e-01 \n", - "iou 9.132528e-01 9.436913e-01 \n", - "dice_score 9.546599e-01 9.710300e-01 \n", - "animal 0.000000e+00 0.000000e+00 \n", - "human.pedestrian.adult 1.235000e+03 2.140000e+02 \n", - "human.pedestrian.child 0.000000e+00 0.000000e+00 \n", - "human.pedestrian.construction_worker 0.000000e+00 2.000000e+01 \n", - "human.pedestrian.personal_mobility 0.000000e+00 0.000000e+00 \n", - "human.pedestrian.police_officer 0.000000e+00 0.000000e+00 \n", - "human.pedestrian.stroller 0.000000e+00 0.000000e+00 \n", - "human.pedestrian.wheelchair 0.000000e+00 0.000000e+00 \n", - "movable_object.barrier 3.000000e+00 7.490000e+02 \n", - "movable_object.debris 0.000000e+00 0.000000e+00 \n", - "movable_object.pushable_pullable 0.000000e+00 0.000000e+00 \n", - "movable_object.trafficcone 2.000000e+01 9.700000e+01 \n", - "static_object.bicycle_rack 0.000000e+00 0.000000e+00 \n", - "vehicle.bicycle 0.000000e+00 0.000000e+00 \n", - "vehicle.bus.bendy 0.000000e+00 0.000000e+00 \n", - "vehicle.bus.rigid 1.400000e+02 8.200000e+01 \n", - "vehicle.car 3.026940e+05 5.070000e+02 \n", - "vehicle.construction 8.930000e+02 1.243870e+05 \n", - "vehicle.emergency.ambulance 0.000000e+00 0.000000e+00 \n", - "vehicle.emergency.police 0.000000e+00 0.000000e+00 \n", - "vehicle.motorcycle 1.369000e+03 1.490000e+02 \n", - "vehicle.trailer 0.000000e+00 2.900000e+01 \n", - "vehicle.truck 5.173000e+03 1.542000e+03 \n", - "flat.driveable_surface 3.481000e+03 2.910000e+02 \n", - "\n", - " vehicle.emergency.ambulance \\\n", - "tp 0.0 \n", - "fp 0.0 \n", - "fn 0.0 \n", - "tn 3326172.0 \n", - "precision NaN \n", - "recall NaN \n", - "accuracy 1.0 \n", - "f1_score NaN \n", - "iou NaN \n", - "dice_score NaN \n", - "animal 0.0 \n", - "human.pedestrian.adult 0.0 \n", - "human.pedestrian.child 0.0 \n", - "human.pedestrian.construction_worker 0.0 \n", - "human.pedestrian.personal_mobility 0.0 \n", - "human.pedestrian.police_officer 0.0 \n", - "human.pedestrian.stroller 0.0 \n", - "human.pedestrian.wheelchair 0.0 \n", - "movable_object.barrier 0.0 \n", - "movable_object.debris 0.0 \n", - "movable_object.pushable_pullable 0.0 \n", - "movable_object.trafficcone 0.0 \n", - "static_object.bicycle_rack 0.0 \n", - "vehicle.bicycle 0.0 \n", - "vehicle.bus.bendy 0.0 \n", - "vehicle.bus.rigid 0.0 \n", - "vehicle.car 0.0 \n", - "vehicle.construction 0.0 \n", - "vehicle.emergency.ambulance 0.0 \n", - "vehicle.emergency.police 0.0 \n", - "vehicle.motorcycle 0.0 \n", - "vehicle.trailer 0.0 \n", - "vehicle.truck 0.0 \n", - "flat.driveable_surface 0.0 \n", - "\n", - " vehicle.emergency.police \\\n", - "tp 0.0 \n", - "fp 0.0 \n", - "fn 0.0 \n", - "tn 3326172.0 \n", - "precision NaN \n", - "recall NaN \n", - "accuracy 1.0 \n", - "f1_score NaN \n", - "iou NaN \n", - "dice_score NaN \n", - "animal 0.0 \n", - "human.pedestrian.adult 0.0 \n", - "human.pedestrian.child 0.0 \n", - "human.pedestrian.construction_worker 0.0 \n", - "human.pedestrian.personal_mobility 0.0 \n", - "human.pedestrian.police_officer 0.0 \n", - "human.pedestrian.stroller 0.0 \n", - "human.pedestrian.wheelchair 0.0 \n", - "movable_object.barrier 0.0 \n", - "movable_object.debris 0.0 \n", - "movable_object.pushable_pullable 0.0 \n", - "movable_object.trafficcone 0.0 \n", - "static_object.bicycle_rack 0.0 \n", - "vehicle.bicycle 0.0 \n", - "vehicle.bus.bendy 0.0 \n", - "vehicle.bus.rigid 0.0 \n", - "vehicle.car 0.0 \n", - "vehicle.construction 0.0 \n", - "vehicle.emergency.ambulance 0.0 \n", - "vehicle.emergency.police 0.0 \n", - "vehicle.motorcycle 0.0 \n", - "vehicle.trailer 0.0 \n", - "vehicle.truck 0.0 \n", - "flat.driveable_surface 0.0 \n", - "\n", - " vehicle.motorcycle vehicle.trailer \\\n", - "tp 2.299100e+04 6.436000e+03 \n", - "fp 3.728000e+03 5.490000e+02 \n", - "fn 1.579000e+03 1.489800e+04 \n", - "tn 3.297874e+06 3.304289e+06 \n", - "precision 8.604738e-01 9.214030e-01 \n", - "recall 9.357346e-01 3.016781e-01 \n", - "accuracy 9.984045e-01 9.953559e-01 \n", - "f1_score 8.965275e-01 4.545358e-01 \n", - "iou 8.124602e-01 2.941096e-01 \n", - "dice_score 8.965275e-01 4.545358e-01 \n", - "animal 0.000000e+00 0.000000e+00 \n", - "human.pedestrian.adult 6.370000e+02 0.000000e+00 \n", - "human.pedestrian.child 0.000000e+00 0.000000e+00 \n", - "human.pedestrian.construction_worker 0.000000e+00 0.000000e+00 \n", - "human.pedestrian.personal_mobility 0.000000e+00 0.000000e+00 \n", - "human.pedestrian.police_officer 0.000000e+00 0.000000e+00 \n", - "human.pedestrian.stroller 0.000000e+00 0.000000e+00 \n", - "human.pedestrian.wheelchair 0.000000e+00 0.000000e+00 \n", - "movable_object.barrier 0.000000e+00 0.000000e+00 \n", - "movable_object.debris 0.000000e+00 0.000000e+00 \n", - "movable_object.pushable_pullable 0.000000e+00 0.000000e+00 \n", - "movable_object.trafficcone 0.000000e+00 4.600000e+01 \n", - "static_object.bicycle_rack 4.000000e+01 0.000000e+00 \n", - "vehicle.bicycle 0.000000e+00 0.000000e+00 \n", - "vehicle.bus.bendy 0.000000e+00 0.000000e+00 \n", - "vehicle.bus.rigid 0.000000e+00 0.000000e+00 \n", - "vehicle.car 7.370000e+02 1.821000e+03 \n", - "vehicle.construction 0.000000e+00 0.000000e+00 \n", - "vehicle.emergency.ambulance 0.000000e+00 0.000000e+00 \n", - "vehicle.emergency.police 0.000000e+00 0.000000e+00 \n", - "vehicle.motorcycle 2.299100e+04 0.000000e+00 \n", - "vehicle.trailer 0.000000e+00 6.436000e+03 \n", - "vehicle.truck 4.000000e+00 1.282800e+04 \n", - "flat.driveable_surface 1.610000e+02 2.030000e+02 \n", - "\n", - " vehicle.truck flat.driveable_surface \\\n", - "tp 7.031500e+04 2.559201e+06 \n", - "fp 2.202200e+04 8.568000e+03 \n", - "fn 9.345000e+03 9.131000e+03 \n", - "tn 3.224490e+06 7.492720e+05 \n", - "precision 7.615041e-01 9.966633e-01 \n", - "recall 8.826889e-01 9.964448e-01 \n", - "accuracy 9.905696e-01 9.946789e-01 \n", - "f1_score 8.176305e-01 9.965540e-01 \n", - "iou 6.915187e-01 9.931317e-01 \n", - "dice_score 8.176305e-01 9.965540e-01 \n", - "animal 0.000000e+00 0.000000e+00 \n", - "human.pedestrian.adult 6.580000e+02 1.288000e+03 \n", - "human.pedestrian.child 0.000000e+00 0.000000e+00 \n", - "human.pedestrian.construction_worker 8.000000e+00 0.000000e+00 \n", - "human.pedestrian.personal_mobility 0.000000e+00 0.000000e+00 \n", - "human.pedestrian.police_officer 0.000000e+00 0.000000e+00 \n", - "human.pedestrian.stroller 0.000000e+00 0.000000e+00 \n", - "human.pedestrian.wheelchair 0.000000e+00 0.000000e+00 \n", - "movable_object.barrier 3.480000e+02 1.260000e+03 \n", - "movable_object.debris 0.000000e+00 0.000000e+00 \n", - "movable_object.pushable_pullable 0.000000e+00 0.000000e+00 \n", - "movable_object.trafficcone 1.360000e+02 4.700000e+02 \n", - "static_object.bicycle_rack 0.000000e+00 0.000000e+00 \n", - "vehicle.bicycle 0.000000e+00 0.000000e+00 \n", - "vehicle.bus.bendy 0.000000e+00 0.000000e+00 \n", - "vehicle.bus.rigid 1.290000e+02 1.900000e+02 \n", - "vehicle.car 6.235000e+03 4.619000e+03 \n", - "vehicle.construction 8.140000e+02 9.100000e+01 \n", - "vehicle.emergency.ambulance 0.000000e+00 0.000000e+00 \n", - "vehicle.emergency.police 0.000000e+00 0.000000e+00 \n", - "vehicle.motorcycle 0.000000e+00 4.260000e+02 \n", - "vehicle.trailer 4.100000e+02 0.000000e+00 \n", - "vehicle.truck 7.031500e+04 7.870000e+02 \n", - "flat.driveable_surface 6.070000e+02 2.559201e+06 \n", - "\n", - " macro micro \n", - "tp NaN NaN \n", - "fp NaN NaN \n", - "fn NaN NaN \n", - "tn NaN NaN \n", - "precision 0.789279 0.978781 \n", - "recall 0.676058 0.978781 \n", - "accuracy 0.998232 0.998232 \n", - "f1_score 0.789115 0.978781 \n", - "iou 0.617749 0.958443 \n", - "dice_score 0.690476 0.978781 \n", - "animal NaN NaN \n", - "human.pedestrian.adult NaN NaN \n", - "human.pedestrian.child NaN NaN \n", - "human.pedestrian.construction_worker NaN NaN \n", - "human.pedestrian.personal_mobility NaN NaN \n", - "human.pedestrian.police_officer NaN NaN \n", - "human.pedestrian.stroller NaN NaN \n", - "human.pedestrian.wheelchair NaN NaN \n", - "movable_object.barrier NaN NaN \n", - "movable_object.debris NaN NaN \n", - "movable_object.pushable_pullable NaN NaN \n", - "movable_object.trafficcone NaN NaN \n", - "static_object.bicycle_rack NaN NaN \n", - "vehicle.bicycle NaN NaN \n", - "vehicle.bus.bendy NaN NaN \n", - "vehicle.bus.rigid NaN NaN \n", - "vehicle.car NaN NaN \n", - "vehicle.construction NaN NaN \n", - "vehicle.emergency.ambulance NaN NaN \n", - "vehicle.emergency.police NaN NaN \n", - "vehicle.motorcycle NaN NaN \n", - "vehicle.trailer NaN NaN \n", - "vehicle.truck NaN NaN \n", - "flat.driveable_surface NaN NaN \n", - "\n", - "[34 rows x 26 columns]" - ] - }, - "metadata": {}, - "output_type": "display_data" - } - ], + "outputs": [], "source": [ - "# Translate original nuImages dataset labels into the model ontology.\n", - "# Classes ignored by the model config are masked before this translation is applied.\n", - "fallback_class = next(iter(pm_model.ontology.keys()))\n", - "dataset_to_model_translation = {\n", - " class_name: (class_name if class_name in pm_model.ontology else fallback_class)\n", + "# Inspect overlap first\n", + "model_classes = set(model.ontology.keys())\n", + "dataset_classes = set(dataset.ontology.keys())\n", + "print(\"Model classes:\", len(model_classes))\n", + "print(\"Dataset classes:\", len(dataset_classes))\n", + "print(\n", + " \"Overlap:\",\n", + " len(model_classes & dataset_classes),\n", + " sorted(model_classes & dataset_classes)[:20],\n", + ")\n", + "\n", + "# Build a safe translation map\n", + "# old_ontology = dataset.ontology, new_ontology = model.ontology\n", + "# dataset ontology -> model ontology\n", + "fallback = list(model.ontology.keys())[0]\n", + "ontology_translation = {\n", + " class_name: (class_name if class_name in model.ontology else fallback)\n", " for class_name in dataset.ontology.keys()\n", "}\n", "\n", + "# TorchImageSegmentationModel.eval expects a JSON filename, not a Python dict.\n", "translation_path = Path(\"local/data/nuimages_to_model_ontology_translation.json\")\n", "translation_path.parent.mkdir(parents=True, exist_ok=True)\n", "with open(translation_path, \"w\", encoding=\"utf-8\") as f:\n", - " json.dump(dataset_to_model_translation, f, indent=2)\n", + " json.dump(ontology_translation, f, indent=2)\n", "\n", - "results = pm_model.eval(\n", + "results = model.eval(\n", " dataset,\n", - " split=\"val\",\n", + " split=\"train\",\n", " ontology_translation=str(translation_path),\n", - " translation_direction=\"dataset_to_model\",\n", ")\n", - "display(results)\n" + "display(results)" ] } ], From 5c1b02150c75ddc088562684b120e629c38ec58c Mon Sep 17 00:00:00 2001 From: tejass Date: Wed, 1 Jul 2026 19:17:52 +0200 Subject: [PATCH 06/10] Restore NuImages segmentation tutorial from master --- examples/tutorial_nuimages_segmentation.ipynb | 1757 ++++++++++++++++- 1 file changed, 1684 insertions(+), 73 deletions(-) diff --git a/examples/tutorial_nuimages_segmentation.ipynb b/examples/tutorial_nuimages_segmentation.ipynb index 68086277..f2b435ec 100644 --- a/examples/tutorial_nuimages_segmentation.ipynb +++ b/examples/tutorial_nuimages_segmentation.ipynb @@ -1,47 +1,112 @@ { "cells": [ + { + "cell_type": "markdown", + "id": "eeb1b612", + "metadata": {}, + "source": [ + "# NuImages Segmentation Evaluation Tutorial\n" + ] + }, { "cell_type": "code", - "execution_count": null, + "execution_count": 1, "id": "478f413c", "metadata": {}, - "outputs": [], + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Requirement already satisfied: transformers in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (4.30.2)\n", + "Requirement already satisfied: filelock in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from transformers) (3.14.0)\n", + "Requirement already satisfied: huggingface-hub<1.0,>=0.14.1 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from transformers) (0.36.2)\n", + "Requirement already satisfied: numpy>=1.17 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from transformers) (1.26.4)\n", + "Requirement already satisfied: packaging>=20.0 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from transformers) (24.2)\n", + "Requirement already satisfied: pyyaml>=5.1 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from transformers) (6.0.3)\n", + "Requirement already satisfied: regex!=2019.12.17 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from transformers) (2026.5.9)\n", + "Requirement already satisfied: requests in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from transformers) (2.28.2)\n", + "Requirement already satisfied: tokenizers!=0.11.3,<0.14,>=0.11.1 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from transformers) (0.13.3)\n", + "Requirement already satisfied: safetensors>=0.3.1 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from transformers) (0.7.0)\n", + "Requirement already satisfied: tqdm>=4.27 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from transformers) (4.65.2)\n", + "Requirement already satisfied: fsspec>=2023.5.0 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from huggingface-hub<1.0,>=0.14.1->transformers) (2025.12.0)\n", + "Requirement already satisfied: hf-xet<2.0.0,>=1.1.3 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from huggingface-hub<1.0,>=0.14.1->transformers) (1.5.0)\n", + "Requirement already satisfied: typing-extensions>=3.7.4.3 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from huggingface-hub<1.0,>=0.14.1->transformers) (4.15.0)\n", + "Requirement already satisfied: charset-normalizer<4,>=2 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from requests->transformers) (3.4.7)\n", + "Requirement already satisfied: idna<4,>=2.5 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from requests->transformers) (3.16)\n", + "Requirement already satisfied: urllib3<1.27,>=1.21.1 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from requests->transformers) (1.26.20)\n", + "Requirement already satisfied: certifi>=2017.4.17 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from requests->transformers) (2026.5.20)\n" + ] + } + ], "source": [ - "!pip install gdown" + "!pip install transformers" ] }, { - "cell_type": "code", - "execution_count": null, - "id": "a37f27c1", + "cell_type": "markdown", + "id": "13273f08", "metadata": {}, - "outputs": [], "source": [ - "# Download\n", - "!gdown 15Py3AKyGzWNX4M3OJKlQQsnH6IehkoEd -O local/data/image_segmentation_model.pth # model\n", - "!gdown 17eJ6aei6yBAbK1aA4mhTeMdXYycorqcT -O local/data/image_segmentation_model_cfg.json # configuration\n", - "!gdown 1spXt5_ISG1ZaHHO2DTtCSqgkohnhwuzt -O local/data/image_segmentation_model_ontology.json # ontology" + "## Download model from Hugging Face" ] }, { "cell_type": "code", - "execution_count": null, + "execution_count": 2, "id": "2a47f7a9", "metadata": {}, - "outputs": [], + "outputs": [ + { + "name": "stderr", + "output_type": "stream", + "text": [ + "/home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages/huggingface_hub/file_download.py:949: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.\n", + " warnings.warn(\n", + "/home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages/transformers/models/segformer/image_processing_segformer.py:99: FutureWarning: The `reduce_labels` parameter is deprecated and will be removed in a future version. Please use `do_reduce_labels` instead.\n", + " warnings.warn(\n" + ] + } + ], "source": [ - "from perceptionmetrics.models.torch_segmentation import TorchImageSegmentationModel\n", + "import json\n", + "from transformers import AutoImageProcessor\n", "\n", - "model = TorchImageSegmentationModel(\n", - " model=\"local/data/image_segmentation_model.pth\", \n", - " model_cfg=\"local/data/image_segmentation_model_cfg.json\",\n", - " ontology_fname=\"local/data/image_segmentation_model_ontology.json\",\n", - ")" + "model_name = \"tejasstanley/segformer-nuimages-b0-full-bs8\"\n", + "\n", + "processor = AutoImageProcessor.from_pretrained(model_name)\n", + "cfg = processor.to_dict()\n", + "\n", + "pm_cfg = {\n", + " \"resize\": {\n", + " \"height\": cfg[\"size\"][\"height\"],\n", + " \"width\": cfg[\"size\"][\"width\"],\n", + " },\n", + " \"normalization\": {\n", + " \"mean\": list(cfg[\"image_mean\"]),\n", + " \"std\": list(cfg[\"image_std\"]),\n", + " },\n", + " \"ignored_classes\": [\"background\", \"vehicle.ego\"],\n", + "}\n", + "\n", + "model_cfg_path = \"local/data/segformer_nuimages_b0_full_bs8_cfg.json\"\n", + "\n", + "with open(model_cfg_path, \"w\", encoding=\"utf-8\") as f:\n", + " json.dump(pm_cfg, f, indent=2)" + ] + }, + { + "cell_type": "markdown", + "id": "f7f0b0e0", + "metadata": {}, + "source": [ + "## Build Dataset and Model Ontology\n", + "\n" ] }, { "cell_type": "code", - "execution_count": 5, + "execution_count": 3, "id": "444ef811", "metadata": {}, "outputs": [ @@ -49,60 +114,149 @@ "name": "stdout", "output_type": "stream", "text": [ + "Jupyter environment detected. Enabling Open3D WebVisualizer.\n", + "[Open3D INFO] WebRTC GUI backend enabled.\n", + "[Open3D INFO] WebRTCWindowSystem: HTTP handshake server disabled.\n", "Built nuImages segmentation dataset with 50 samples and 26 classes.\n", - "Saved compact ontology to local/data/nuimages_segmentation_ontology_compact.json with 26 classes (max idx=25)\n" + "Saved model ontology to local/data/nuimages_model_ontology.json\n", + "Dataset labels will be translated in memory during evaluation.\n" ] } ], "source": [ - "## NuImages Segmentation Dataset and Model\n", - "\n", "import json\n", "from pathlib import Path\n", "\n", - "from perceptionmetrics.models.torch_segmentation import TorchImageSegmentationModel\n", "from perceptionmetrics.datasets.nuimages import NuImagesSegmentationDataset\n", "\n", "dataset = NuImagesSegmentationDataset(\n", - " dataset_dir=\"local/data/nuimages-v1.0-mini\", version=\"v1.0-mini\", split=\"train\"\n", + " dataset_dir=\"local/data/nuimages-v1.0-mini\", version=\"v1.0-mini\", split=\"val\"\n", ")\n", "\n", - "# Build compact ontology (contiguous idx: 0..N-1) to avoid sparse-index issues (e.g., idx 31).\n", - "sorted_items = sorted(dataset.ontology.items(), key=lambda x: x[1][\"idx\"] )\n", - "compact_ontology = {\n", + "model_labels = {\n", + " 0: \"animal\",\n", + " 1: \"human.pedestrian.adult\",\n", + " 2: \"human.pedestrian.child\",\n", + " 3: \"human.pedestrian.construction_worker\",\n", + " 4: \"human.pedestrian.personal_mobility\",\n", + " 5: \"human.pedestrian.police_officer\",\n", + " 6: \"human.pedestrian.stroller\",\n", + " 7: \"human.pedestrian.wheelchair\",\n", + " 8: \"movable_object.barrier\",\n", + " 9: \"movable_object.debris\",\n", + " 10: \"movable_object.pushable_pullable\",\n", + " 11: \"movable_object.trafficcone\",\n", + " 12: \"static_object.bicycle_rack\",\n", + " 13: \"vehicle.bicycle\",\n", + " 14: \"vehicle.bus.bendy\",\n", + " 15: \"vehicle.bus.rigid\",\n", + " 16: \"vehicle.car\",\n", + " 17: \"vehicle.construction\",\n", + " 18: \"vehicle.emergency.ambulance\",\n", + " 19: \"vehicle.emergency.police\",\n", + " 20: \"vehicle.motorcycle\",\n", + " 21: \"vehicle.trailer\",\n", + " 22: \"vehicle.truck\",\n", + " 23: \"flat.driveable_surface\",\n", + "}\n", + "\n", + "model_ontology = {\n", " class_name: {\n", - " \"idx\": new_idx,\n", - " \"rgb\": class_meta.get(\"rgb\", [0, 0, 0]),\n", + " \"idx\": class_idx,\n", + " \"rgb\": dataset.ontology[class_name][\"rgb\"],\n", " }\n", - " for new_idx, (class_name, class_meta) in enumerate(sorted_items)\n", + " for class_idx, class_name in model_labels.items()\n", "}\n", "\n", - "ontology_path = Path(\"local/data/nuimages_segmentation_ontology_compact.json\")\n", - "ontology_path.parent.mkdir(parents=True, exist_ok=True)\n", - "with open(ontology_path, \"w\") as f:\n", - " json.dump(compact_ontology, f, indent=2)\n", + "model_ontology_path = Path(\"local/data/nuimages_model_ontology.json\")\n", + "model_ontology_path.parent.mkdir(parents=True, exist_ok=True)\n", + "with open(model_ontology_path, \"w\", encoding=\"utf-8\") as f:\n", + " json.dump(model_ontology, f, indent=2)\n", "\n", - "max_idx = max(v[\"idx\"] for v in compact_ontology.values())\n", - "print(\n", - " f\"Saved compact ontology to {ontology_path} with {len(compact_ontology)} classes (max idx={max_idx})\"\n", - " )\n", + "print(f\"Saved model ontology to {model_ontology_path}\")\n", + "print(\"Dataset labels will be translated in memory during evaluation.\")\n" + ] + }, + { + "cell_type": "markdown", + "id": "4e68ee53", + "metadata": {}, + "source": [ + "## Initialize the PerceptionMetrics Model Wrapper\n", + "\n" + ] + }, + { + "cell_type": "code", + "execution_count": 4, + "id": "e34db9f2", + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Torch detection not available\n", + "Tensorflow not available\n" + ] + }, + { + "name": "stderr", + "output_type": "stream", + "text": [ + "/home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages/torchvision/datapoints/__init__.py:12: UserWarning: The torchvision.datapoints and torchvision.transforms.v2 namespaces are still Beta. While we do not expect major breaking changes, some APIs may still change according to user feedback. Please submit any feedback you may have in this issue: https://github.com/pytorch/vision/issues/6753, and you can also check out https://github.com/pytorch/vision/issues/7319 to learn more about the APIs that we suspect might involve future changes. You can silence this warning by calling torchvision.disable_beta_transforms_warning().\n", + " warnings.warn(_BETA_TRANSFORMS_WARNING)\n", + "/home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages/torchvision/transforms/v2/__init__.py:54: UserWarning: The torchvision.datapoints and torchvision.transforms.v2 namespaces are still Beta. While we do not expect major breaking changes, some APIs may still change according to user feedback. Please submit any feedback you may have in this issue: https://github.com/pytorch/vision/issues/6753, and you can also check out https://github.com/pytorch/vision/issues/7319 to learn more about the APIs that we suspect might involve future changes. You can silence this warning by calling torchvision.disable_beta_transforms_warning().\n", + " warnings.warn(_BETA_TRANSFORMS_WARNING)\n", + "/home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages/huggingface_hub/file_download.py:949: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.\n", + " warnings.warn(\n" + ] + } + ], + "source": [ + "from transformers import SegformerForSemanticSegmentation\n", + "from perceptionmetrics.models.torch_segmentation import TorchImageSegmentationModel\n", "\n", - "model = TorchImageSegmentationModel(\n", - " model=\"local/data/image_segmentation_model.pth\",\n", - " model_cfg=\"local/data/image_segmentation_model_cfg.json\",\n", - " ontology_fname=str(ontology_path),\n", + "model_name = \"tejasstanley/segformer-nuimages-b0-full-bs8\"\n", + "model_cfg_path = \"local/data/segformer_nuimages_b0_full_bs8_cfg.json\"\n", + "\n", + "\n", + "hf_model = SegformerForSemanticSegmentation.from_pretrained(model_name)\n", + "\n", + "pm_model = TorchImageSegmentationModel(\n", + " model=hf_model,\n", + " model_cfg=model_cfg_path,\n", + " ontology_fname=str(model_ontology_path),\n", ")" ] }, + { + "cell_type": "markdown", + "id": "16b24b8a", + "metadata": {}, + "source": [ + "## Inspect a Sample Prediction\n" + ] + }, { "cell_type": "code", - "execution_count": 6, + "execution_count": 5, "id": "47f1727b", "metadata": {}, "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Predicted class ids: [1, 3, 8, 11, 13, 16, 21, 22, 23]\n", + "Predicted class names: ['human.pedestrian.adult', 'human.pedestrian.construction_worker', 'movable_object.barrier', 'movable_object.trafficcone', 'vehicle.bicycle', 'vehicle.car', 'vehicle.trailer', 'vehicle.truck', 'flat.driveable_surface']\n", + "Dataset label ids in sample: [0, 4, 9, 17, 23, 24]\n", + "Dataset label names in sample: ['background', 'human.pedestrian.construction_worker', 'movable_object.barrier', 'vehicle.car', 'vehicle.truck', 'flat.driveable_surface']\n" + ] + }, { "data": { - "image/png": "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", + "image/png": "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", "text/plain": [ "
" ] @@ -118,15 +272,43 @@ "\n", "from perceptionmetrics.utils import conversion as uc\n", "\n", - "image_fname = dataset.dataset[\"image\"].iloc[0]\n", + "image_fname = dataset.dataset[\"image\"].iloc[1]\n", + "if dataset.dataset_dir is not None:\n", + " image_fname = Path(dataset.dataset_dir) / image_fname\n", "image = Image.open(image_fname)\n", "\n", - "label_fname = dataset.dataset[\"label\"].iloc[0]\n", + "label_fname = dataset.dataset[\"label\"].iloc[1]\n", + "if dataset.dataset_dir is not None:\n", + " label_fname = Path(dataset.dataset_dir) / label_fname\n", "label = Image.open(label_fname)\n", "label = uc.label_to_rgb(label, dataset.ontology)\n", "\n", - "pred = model.predict(image)\n", - "pred = uc.label_to_rgb(pred, model.ontology)\n", + "raw_pred = pm_model.predict(image)\n", + "raw_pred_np = np.array(raw_pred)\n", + "raw_label_np = np.array(Image.open(label_fname))\n", + "\n", + "model_idx_to_name = {\n", + " class_meta[\"idx\"]: class_name for class_name, class_meta in pm_model.ontology.items()\n", + "}\n", + "dataset_idx_to_name = {\n", + " class_meta[\"idx\"]: class_name for class_name, class_meta in dataset.ontology.items()\n", + "}\n", + "\n", + "pred_ids = sorted(np.unique(raw_pred_np).tolist())\n", + "label_ids = sorted(np.unique(raw_label_np).tolist())\n", + "\n", + "print(\"Predicted class ids:\", pred_ids)\n", + "print(\n", + " \"Predicted class names:\",\n", + " [model_idx_to_name.get(class_id, f\"unknown:{class_id}\") for class_id in pred_ids],\n", + ")\n", + "print(\"Dataset label ids in sample:\", label_ids)\n", + "print(\n", + " \"Dataset label names in sample:\",\n", + " [dataset_idx_to_name.get(class_id, f\"unknown:{class_id}\") for class_id in label_ids],\n", + ")\n", + "\n", + "pred = uc.label_to_rgb(raw_pred, pm_model.ontology)\n", "pred = pred.resize(label.size)\n", "\n", "plt.figure(figsize=(10, 10))\n", @@ -136,45 +318,1474 @@ "plt.show()" ] }, + { + "cell_type": "markdown", + "id": "bcb40b82", + "metadata": {}, + "source": [ + "## Translate Dataset Labels and Run Evaluation\n", + "\n", + "The evaluation call maps ground-truth labels from the nuImages ontology into the model ontology with `translation_direction=\"dataset_to_model\"`.\n" + ] + }, { "cell_type": "code", - "execution_count": null, + "execution_count": 6, "id": "a1f6ca6b", "metadata": {}, - "outputs": [], + "outputs": [ + { + "name": "stderr", + "output_type": "stream", + "text": [ + "100%|██████████| 50/50 [00:01<00:00, 35.81it/s]\n", + "/home/tejass/Downloads/TUDELFT_ROBOTICS/GSOC/gsoc2026-Tejas_Stanley/PerceptionMetrics/perceptionmetrics/utils/segmentation_metrics.py:147: RuntimeWarning: invalid value encountered in divide\n", + " return np.where(denominator > 0, tp / denominator, np.nan)\n", + "/home/tejass/Downloads/TUDELFT_ROBOTICS/GSOC/gsoc2026-Tejas_Stanley/PerceptionMetrics/perceptionmetrics/utils/segmentation_metrics.py:164: RuntimeWarning: invalid value encountered in divide\n", + " return np.where(denominator > 0, tp / denominator, np.nan)\n", + "/home/tejass/Downloads/TUDELFT_ROBOTICS/GSOC/gsoc2026-Tejas_Stanley/PerceptionMetrics/perceptionmetrics/utils/segmentation_metrics.py:203: RuntimeWarning: invalid value encountered in divide\n", + " denominator > 0, 2 * (precision * recall) / denominator, np.nan\n", + "/home/tejass/Downloads/TUDELFT_ROBOTICS/GSOC/gsoc2026-Tejas_Stanley/PerceptionMetrics/perceptionmetrics/utils/segmentation_metrics.py:222: RuntimeWarning: invalid value encountered in divide\n", + " return np.where(union > 0, tp / union, np.nan)\n", + "/home/tejass/Downloads/TUDELFT_ROBOTICS/GSOC/gsoc2026-Tejas_Stanley/PerceptionMetrics/perceptionmetrics/utils/segmentation_metrics.py:240: RuntimeWarning: invalid value encountered in divide\n", + " return np.where(denominator > 0, 2 * tp / denominator, np.nan)\n" + ] + }, + { + "data": { + "text/html": [ + "
\n", + "\n", + "\n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + "
animalhuman.pedestrian.adulthuman.pedestrian.childhuman.pedestrian.construction_workerhuman.pedestrian.personal_mobilityhuman.pedestrian.police_officerhuman.pedestrian.strollerhuman.pedestrian.wheelchairmovable_object.barriermovable_object.debris...vehicle.carvehicle.constructionvehicle.emergency.ambulancevehicle.emergency.policevehicle.motorcyclevehicle.trailervehicle.truckflat.driveable_surfacemacromicro
tp0.04.578900e+040.02.280000e+020.000000e+000.00.00.07.052100e+040.000000e+00...3.026940e+051.243870e+050.00.02.299100e+046.436000e+037.031500e+042.559201e+06NaNNaN
fp0.06.403000e+030.02.690000e+020.000000e+000.00.00.03.562000e+031.000000e+00...1.643800e+043.742000e+030.00.03.728000e+035.490000e+022.202200e+048.568000e+03NaNNaN
fn0.05.353000e+030.01.494000e+031.254000e+030.00.00.03.064000e+031.146000e+03...1.231400e+043.680000e+030.00.01.579000e+031.489800e+049.345000e+039.131000e+03NaNNaN
tn3326172.03.268627e+063326172.03.324181e+063.324918e+063326172.03326172.03326172.03.249025e+063.325025e+06...2.994726e+063.194363e+063326172.03326172.03.297874e+063.304289e+063.224490e+067.492720e+05NaNNaN
precisionNaN8.773184e-01NaN4.587525e-01NaNNaNNaNNaN9.519188e-010.000000e+00...9.484915e-019.707951e-01NaNNaN8.604738e-019.214030e-017.615041e-019.966633e-010.7892790.978781
recallNaN8.953306e-01NaN1.324042e-010.000000e+00NaNNaNNaN9.583611e-010.000000e+00...9.609089e-019.712650e-01NaNNaN9.357346e-013.016781e-018.826889e-019.964448e-010.6760580.978781
accuracy1.09.964656e-011.09.994700e-019.996230e-011.01.01.09.980079e-019.996552e-01...9.913558e-019.977686e-011.01.09.984045e-019.953559e-019.905696e-019.946789e-010.9982320.998232
f1_scoreNaN8.862330e-01NaN2.054980e-01NaNNaNNaNNaN9.551291e-01NaN...9.546599e-019.710300e-01NaNNaN8.965275e-014.545358e-018.176305e-019.965540e-010.7891150.978781
iouNaN7.957077e-01NaN1.145153e-010.000000e+00NaNNaNNaN9.141120e-010.000000e+00...9.132528e-019.436913e-01NaNNaN8.124602e-012.941096e-016.915187e-019.931317e-010.6177490.958443
dice_scoreNaN8.862330e-01NaN2.054980e-010.000000e+00NaNNaNNaN9.551291e-010.000000e+00...9.546599e-019.710300e-01NaNNaN8.965275e-014.545358e-018.176305e-019.965540e-010.6904760.978781
animal0.00.000000e+000.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...0.000000e+000.000000e+000.00.00.000000e+000.000000e+000.000000e+000.000000e+00NaNNaN
human.pedestrian.adult0.04.578900e+040.09.000000e+014.470000e+020.00.00.05.200000e+019.140000e+02...1.235000e+032.140000e+020.00.06.370000e+020.000000e+006.580000e+021.288000e+03NaNNaN
human.pedestrian.child0.00.000000e+000.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...0.000000e+000.000000e+000.00.00.000000e+000.000000e+000.000000e+000.000000e+00NaNNaN
human.pedestrian.construction_worker0.02.390000e+020.02.280000e+020.000000e+000.00.00.02.000000e+000.000000e+00...0.000000e+002.000000e+010.00.00.000000e+000.000000e+008.000000e+000.000000e+00NaNNaN
human.pedestrian.personal_mobility0.00.000000e+000.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...0.000000e+000.000000e+000.00.00.000000e+000.000000e+000.000000e+000.000000e+00NaNNaN
human.pedestrian.police_officer0.00.000000e+000.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...0.000000e+000.000000e+000.00.00.000000e+000.000000e+000.000000e+000.000000e+00NaNNaN
human.pedestrian.stroller0.00.000000e+000.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...0.000000e+000.000000e+000.00.00.000000e+000.000000e+000.000000e+000.000000e+00NaNNaN
human.pedestrian.wheelchair0.00.000000e+000.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...0.000000e+000.000000e+000.00.00.000000e+000.000000e+000.000000e+000.000000e+00NaNNaN
movable_object.barrier0.03.130000e+020.01.500000e+010.000000e+000.00.00.07.052100e+046.000000e+00...3.000000e+007.490000e+020.00.00.000000e+000.000000e+003.480000e+021.260000e+03NaNNaN
movable_object.debris0.00.000000e+000.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...0.000000e+000.000000e+000.00.00.000000e+000.000000e+000.000000e+000.000000e+00NaNNaN
movable_object.pushable_pullable0.01.100000e+020.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...0.000000e+000.000000e+000.00.00.000000e+000.000000e+000.000000e+000.000000e+00NaNNaN
movable_object.trafficcone0.09.600000e+010.00.000000e+000.000000e+000.00.00.02.320000e+020.000000e+00...2.000000e+019.700000e+010.00.00.000000e+004.600000e+011.360000e+024.700000e+02NaNNaN
static_object.bicycle_rack0.00.000000e+000.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...0.000000e+000.000000e+000.00.04.000000e+010.000000e+000.000000e+000.000000e+00NaNNaN
vehicle.bicycle0.02.070000e+020.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...0.000000e+000.000000e+000.00.00.000000e+000.000000e+000.000000e+000.000000e+00NaNNaN
vehicle.bus.bendy0.00.000000e+000.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...0.000000e+000.000000e+000.00.00.000000e+000.000000e+000.000000e+000.000000e+00NaNNaN
vehicle.bus.rigid0.00.000000e+000.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...1.400000e+028.200000e+010.00.00.000000e+000.000000e+001.290000e+021.900000e+02NaNNaN
vehicle.car0.01.746000e+030.05.300000e+010.000000e+000.00.00.02.890000e+020.000000e+00...3.026940e+055.070000e+020.00.07.370000e+021.821000e+036.235000e+034.619000e+03NaNNaN
vehicle.construction0.03.890000e+020.06.820000e+020.000000e+000.00.00.08.730000e+020.000000e+00...8.930000e+021.243870e+050.00.00.000000e+000.000000e+008.140000e+029.100000e+01NaNNaN
vehicle.emergency.ambulance0.00.000000e+000.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...0.000000e+000.000000e+000.00.00.000000e+000.000000e+000.000000e+000.000000e+00NaNNaN
vehicle.emergency.police0.00.000000e+000.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...0.000000e+000.000000e+000.00.00.000000e+000.000000e+000.000000e+000.000000e+00NaNNaN
vehicle.motorcycle0.06.670000e+020.02.220000e+027.190000e+020.00.00.01.760000e+020.000000e+00...1.369000e+031.490000e+020.00.02.299100e+040.000000e+000.000000e+004.260000e+02NaNNaN
vehicle.trailer0.00.000000e+000.00.000000e+000.000000e+000.00.00.06.000000e+010.000000e+00...0.000000e+002.900000e+010.00.00.000000e+006.436000e+034.100000e+020.000000e+00NaNNaN
vehicle.truck0.05.660000e+020.03.350000e+026.600000e+010.00.00.04.200000e+010.000000e+00...5.173000e+031.542000e+030.00.04.000000e+001.282800e+047.031500e+047.870000e+02NaNNaN
flat.driveable_surface0.01.020000e+030.09.700000e+012.200000e+010.00.00.01.338000e+032.260000e+02...3.481000e+032.910000e+020.00.01.610000e+022.030000e+026.070000e+022.559201e+06NaNNaN
\n", + "

34 rows × 26 columns

\n", + "
" + ], + "text/plain": [ + " animal human.pedestrian.adult \\\n", + "tp 0.0 4.578900e+04 \n", + "fp 0.0 6.403000e+03 \n", + "fn 0.0 5.353000e+03 \n", + "tn 3326172.0 3.268627e+06 \n", + "precision NaN 8.773184e-01 \n", + "recall NaN 8.953306e-01 \n", + "accuracy 1.0 9.964656e-01 \n", + "f1_score NaN 8.862330e-01 \n", + "iou NaN 7.957077e-01 \n", + "dice_score NaN 8.862330e-01 \n", + "animal 0.0 0.000000e+00 \n", + "human.pedestrian.adult 0.0 4.578900e+04 \n", + "human.pedestrian.child 0.0 0.000000e+00 \n", + "human.pedestrian.construction_worker 0.0 2.390000e+02 \n", + "human.pedestrian.personal_mobility 0.0 0.000000e+00 \n", + "human.pedestrian.police_officer 0.0 0.000000e+00 \n", + "human.pedestrian.stroller 0.0 0.000000e+00 \n", + "human.pedestrian.wheelchair 0.0 0.000000e+00 \n", + "movable_object.barrier 0.0 3.130000e+02 \n", + "movable_object.debris 0.0 0.000000e+00 \n", + "movable_object.pushable_pullable 0.0 1.100000e+02 \n", + "movable_object.trafficcone 0.0 9.600000e+01 \n", + "static_object.bicycle_rack 0.0 0.000000e+00 \n", + "vehicle.bicycle 0.0 2.070000e+02 \n", + "vehicle.bus.bendy 0.0 0.000000e+00 \n", + "vehicle.bus.rigid 0.0 0.000000e+00 \n", + "vehicle.car 0.0 1.746000e+03 \n", + "vehicle.construction 0.0 3.890000e+02 \n", + "vehicle.emergency.ambulance 0.0 0.000000e+00 \n", + "vehicle.emergency.police 0.0 0.000000e+00 \n", + "vehicle.motorcycle 0.0 6.670000e+02 \n", + "vehicle.trailer 0.0 0.000000e+00 \n", + "vehicle.truck 0.0 5.660000e+02 \n", + "flat.driveable_surface 0.0 1.020000e+03 \n", + "\n", + " human.pedestrian.child \\\n", + "tp 0.0 \n", + "fp 0.0 \n", + "fn 0.0 \n", + "tn 3326172.0 \n", + "precision NaN \n", + "recall NaN \n", + "accuracy 1.0 \n", + "f1_score NaN \n", + "iou NaN \n", + "dice_score NaN \n", + "animal 0.0 \n", + "human.pedestrian.adult 0.0 \n", + "human.pedestrian.child 0.0 \n", + "human.pedestrian.construction_worker 0.0 \n", + "human.pedestrian.personal_mobility 0.0 \n", + "human.pedestrian.police_officer 0.0 \n", + "human.pedestrian.stroller 0.0 \n", + "human.pedestrian.wheelchair 0.0 \n", + "movable_object.barrier 0.0 \n", + "movable_object.debris 0.0 \n", + "movable_object.pushable_pullable 0.0 \n", + "movable_object.trafficcone 0.0 \n", + "static_object.bicycle_rack 0.0 \n", + "vehicle.bicycle 0.0 \n", + "vehicle.bus.bendy 0.0 \n", + "vehicle.bus.rigid 0.0 \n", + "vehicle.car 0.0 \n", + "vehicle.construction 0.0 \n", + "vehicle.emergency.ambulance 0.0 \n", + "vehicle.emergency.police 0.0 \n", + "vehicle.motorcycle 0.0 \n", + "vehicle.trailer 0.0 \n", + "vehicle.truck 0.0 \n", + "flat.driveable_surface 0.0 \n", + "\n", + " human.pedestrian.construction_worker \\\n", + "tp 2.280000e+02 \n", + "fp 2.690000e+02 \n", + "fn 1.494000e+03 \n", + "tn 3.324181e+06 \n", + "precision 4.587525e-01 \n", + "recall 1.324042e-01 \n", + "accuracy 9.994700e-01 \n", + "f1_score 2.054980e-01 \n", + "iou 1.145153e-01 \n", + "dice_score 2.054980e-01 \n", + "animal 0.000000e+00 \n", + "human.pedestrian.adult 9.000000e+01 \n", + "human.pedestrian.child 0.000000e+00 \n", + "human.pedestrian.construction_worker 2.280000e+02 \n", + "human.pedestrian.personal_mobility 0.000000e+00 \n", + "human.pedestrian.police_officer 0.000000e+00 \n", + "human.pedestrian.stroller 0.000000e+00 \n", + "human.pedestrian.wheelchair 0.000000e+00 \n", + "movable_object.barrier 1.500000e+01 \n", + "movable_object.debris 0.000000e+00 \n", + "movable_object.pushable_pullable 0.000000e+00 \n", + "movable_object.trafficcone 0.000000e+00 \n", + "static_object.bicycle_rack 0.000000e+00 \n", + "vehicle.bicycle 0.000000e+00 \n", + "vehicle.bus.bendy 0.000000e+00 \n", + "vehicle.bus.rigid 0.000000e+00 \n", + "vehicle.car 5.300000e+01 \n", + "vehicle.construction 6.820000e+02 \n", + "vehicle.emergency.ambulance 0.000000e+00 \n", + "vehicle.emergency.police 0.000000e+00 \n", + "vehicle.motorcycle 2.220000e+02 \n", + "vehicle.trailer 0.000000e+00 \n", + "vehicle.truck 3.350000e+02 \n", + "flat.driveable_surface 9.700000e+01 \n", + "\n", + " human.pedestrian.personal_mobility \\\n", + "tp 0.000000e+00 \n", + "fp 0.000000e+00 \n", + "fn 1.254000e+03 \n", + "tn 3.324918e+06 \n", + "precision NaN \n", + "recall 0.000000e+00 \n", + "accuracy 9.996230e-01 \n", + "f1_score NaN \n", + "iou 0.000000e+00 \n", + "dice_score 0.000000e+00 \n", + "animal 0.000000e+00 \n", + "human.pedestrian.adult 4.470000e+02 \n", + "human.pedestrian.child 0.000000e+00 \n", + "human.pedestrian.construction_worker 0.000000e+00 \n", + "human.pedestrian.personal_mobility 0.000000e+00 \n", + "human.pedestrian.police_officer 0.000000e+00 \n", + "human.pedestrian.stroller 0.000000e+00 \n", + "human.pedestrian.wheelchair 0.000000e+00 \n", + "movable_object.barrier 0.000000e+00 \n", + "movable_object.debris 0.000000e+00 \n", + "movable_object.pushable_pullable 0.000000e+00 \n", + "movable_object.trafficcone 0.000000e+00 \n", + "static_object.bicycle_rack 0.000000e+00 \n", + "vehicle.bicycle 0.000000e+00 \n", + "vehicle.bus.bendy 0.000000e+00 \n", + "vehicle.bus.rigid 0.000000e+00 \n", + "vehicle.car 0.000000e+00 \n", + "vehicle.construction 0.000000e+00 \n", + "vehicle.emergency.ambulance 0.000000e+00 \n", + "vehicle.emergency.police 0.000000e+00 \n", + "vehicle.motorcycle 7.190000e+02 \n", + "vehicle.trailer 0.000000e+00 \n", + "vehicle.truck 6.600000e+01 \n", + "flat.driveable_surface 2.200000e+01 \n", + "\n", + " human.pedestrian.police_officer \\\n", + "tp 0.0 \n", + "fp 0.0 \n", + "fn 0.0 \n", + "tn 3326172.0 \n", + "precision NaN \n", + "recall NaN \n", + "accuracy 1.0 \n", + "f1_score NaN \n", + "iou NaN \n", + "dice_score NaN \n", + "animal 0.0 \n", + "human.pedestrian.adult 0.0 \n", + "human.pedestrian.child 0.0 \n", + "human.pedestrian.construction_worker 0.0 \n", + "human.pedestrian.personal_mobility 0.0 \n", + "human.pedestrian.police_officer 0.0 \n", + "human.pedestrian.stroller 0.0 \n", + "human.pedestrian.wheelchair 0.0 \n", + "movable_object.barrier 0.0 \n", + "movable_object.debris 0.0 \n", + "movable_object.pushable_pullable 0.0 \n", + "movable_object.trafficcone 0.0 \n", + "static_object.bicycle_rack 0.0 \n", + "vehicle.bicycle 0.0 \n", + "vehicle.bus.bendy 0.0 \n", + "vehicle.bus.rigid 0.0 \n", + "vehicle.car 0.0 \n", + "vehicle.construction 0.0 \n", + "vehicle.emergency.ambulance 0.0 \n", + "vehicle.emergency.police 0.0 \n", + "vehicle.motorcycle 0.0 \n", + "vehicle.trailer 0.0 \n", + "vehicle.truck 0.0 \n", + "flat.driveable_surface 0.0 \n", + "\n", + " human.pedestrian.stroller \\\n", + "tp 0.0 \n", + "fp 0.0 \n", + "fn 0.0 \n", + "tn 3326172.0 \n", + "precision NaN \n", + "recall NaN \n", + "accuracy 1.0 \n", + "f1_score NaN \n", + "iou NaN \n", + "dice_score NaN \n", + "animal 0.0 \n", + "human.pedestrian.adult 0.0 \n", + "human.pedestrian.child 0.0 \n", + "human.pedestrian.construction_worker 0.0 \n", + "human.pedestrian.personal_mobility 0.0 \n", + "human.pedestrian.police_officer 0.0 \n", + "human.pedestrian.stroller 0.0 \n", + "human.pedestrian.wheelchair 0.0 \n", + "movable_object.barrier 0.0 \n", + "movable_object.debris 0.0 \n", + "movable_object.pushable_pullable 0.0 \n", + "movable_object.trafficcone 0.0 \n", + "static_object.bicycle_rack 0.0 \n", + "vehicle.bicycle 0.0 \n", + "vehicle.bus.bendy 0.0 \n", + "vehicle.bus.rigid 0.0 \n", + "vehicle.car 0.0 \n", + "vehicle.construction 0.0 \n", + "vehicle.emergency.ambulance 0.0 \n", + "vehicle.emergency.police 0.0 \n", + "vehicle.motorcycle 0.0 \n", + "vehicle.trailer 0.0 \n", + "vehicle.truck 0.0 \n", + "flat.driveable_surface 0.0 \n", + "\n", + " human.pedestrian.wheelchair \\\n", + "tp 0.0 \n", + "fp 0.0 \n", + "fn 0.0 \n", + "tn 3326172.0 \n", + "precision NaN \n", + "recall NaN \n", + "accuracy 1.0 \n", + "f1_score NaN \n", + "iou NaN \n", + "dice_score NaN \n", + "animal 0.0 \n", + "human.pedestrian.adult 0.0 \n", + "human.pedestrian.child 0.0 \n", + "human.pedestrian.construction_worker 0.0 \n", + "human.pedestrian.personal_mobility 0.0 \n", + "human.pedestrian.police_officer 0.0 \n", + "human.pedestrian.stroller 0.0 \n", + "human.pedestrian.wheelchair 0.0 \n", + "movable_object.barrier 0.0 \n", + "movable_object.debris 0.0 \n", + "movable_object.pushable_pullable 0.0 \n", + "movable_object.trafficcone 0.0 \n", + "static_object.bicycle_rack 0.0 \n", + "vehicle.bicycle 0.0 \n", + "vehicle.bus.bendy 0.0 \n", + "vehicle.bus.rigid 0.0 \n", + "vehicle.car 0.0 \n", + "vehicle.construction 0.0 \n", + "vehicle.emergency.ambulance 0.0 \n", + "vehicle.emergency.police 0.0 \n", + "vehicle.motorcycle 0.0 \n", + "vehicle.trailer 0.0 \n", + "vehicle.truck 0.0 \n", + "flat.driveable_surface 0.0 \n", + "\n", + " movable_object.barrier \\\n", + "tp 7.052100e+04 \n", + "fp 3.562000e+03 \n", + "fn 3.064000e+03 \n", + "tn 3.249025e+06 \n", + "precision 9.519188e-01 \n", + "recall 9.583611e-01 \n", + "accuracy 9.980079e-01 \n", + "f1_score 9.551291e-01 \n", + "iou 9.141120e-01 \n", + "dice_score 9.551291e-01 \n", + "animal 0.000000e+00 \n", + "human.pedestrian.adult 5.200000e+01 \n", + "human.pedestrian.child 0.000000e+00 \n", + "human.pedestrian.construction_worker 2.000000e+00 \n", + "human.pedestrian.personal_mobility 0.000000e+00 \n", + "human.pedestrian.police_officer 0.000000e+00 \n", + "human.pedestrian.stroller 0.000000e+00 \n", + "human.pedestrian.wheelchair 0.000000e+00 \n", + "movable_object.barrier 7.052100e+04 \n", + "movable_object.debris 0.000000e+00 \n", + "movable_object.pushable_pullable 0.000000e+00 \n", + "movable_object.trafficcone 2.320000e+02 \n", + "static_object.bicycle_rack 0.000000e+00 \n", + "vehicle.bicycle 0.000000e+00 \n", + "vehicle.bus.bendy 0.000000e+00 \n", + "vehicle.bus.rigid 0.000000e+00 \n", + "vehicle.car 2.890000e+02 \n", + "vehicle.construction 8.730000e+02 \n", + "vehicle.emergency.ambulance 0.000000e+00 \n", + "vehicle.emergency.police 0.000000e+00 \n", + "vehicle.motorcycle 1.760000e+02 \n", + "vehicle.trailer 6.000000e+01 \n", + "vehicle.truck 4.200000e+01 \n", + "flat.driveable_surface 1.338000e+03 \n", + "\n", + " movable_object.debris ... \\\n", + "tp 0.000000e+00 ... \n", + "fp 1.000000e+00 ... \n", + "fn 1.146000e+03 ... \n", + "tn 3.325025e+06 ... \n", + "precision 0.000000e+00 ... \n", + "recall 0.000000e+00 ... \n", + "accuracy 9.996552e-01 ... \n", + "f1_score NaN ... \n", + "iou 0.000000e+00 ... \n", + "dice_score 0.000000e+00 ... \n", + "animal 0.000000e+00 ... \n", + "human.pedestrian.adult 9.140000e+02 ... \n", + "human.pedestrian.child 0.000000e+00 ... \n", + "human.pedestrian.construction_worker 0.000000e+00 ... \n", + "human.pedestrian.personal_mobility 0.000000e+00 ... \n", + "human.pedestrian.police_officer 0.000000e+00 ... \n", + "human.pedestrian.stroller 0.000000e+00 ... \n", + "human.pedestrian.wheelchair 0.000000e+00 ... \n", + "movable_object.barrier 6.000000e+00 ... \n", + "movable_object.debris 0.000000e+00 ... \n", + "movable_object.pushable_pullable 0.000000e+00 ... \n", + "movable_object.trafficcone 0.000000e+00 ... \n", + "static_object.bicycle_rack 0.000000e+00 ... \n", + "vehicle.bicycle 0.000000e+00 ... \n", + "vehicle.bus.bendy 0.000000e+00 ... \n", + "vehicle.bus.rigid 0.000000e+00 ... \n", + "vehicle.car 0.000000e+00 ... \n", + "vehicle.construction 0.000000e+00 ... \n", + "vehicle.emergency.ambulance 0.000000e+00 ... \n", + "vehicle.emergency.police 0.000000e+00 ... \n", + "vehicle.motorcycle 0.000000e+00 ... \n", + "vehicle.trailer 0.000000e+00 ... \n", + "vehicle.truck 0.000000e+00 ... \n", + "flat.driveable_surface 2.260000e+02 ... \n", + "\n", + " vehicle.car vehicle.construction \\\n", + "tp 3.026940e+05 1.243870e+05 \n", + "fp 1.643800e+04 3.742000e+03 \n", + "fn 1.231400e+04 3.680000e+03 \n", + "tn 2.994726e+06 3.194363e+06 \n", + "precision 9.484915e-01 9.707951e-01 \n", + "recall 9.609089e-01 9.712650e-01 \n", + "accuracy 9.913558e-01 9.977686e-01 \n", + "f1_score 9.546599e-01 9.710300e-01 \n", + "iou 9.132528e-01 9.436913e-01 \n", + "dice_score 9.546599e-01 9.710300e-01 \n", + "animal 0.000000e+00 0.000000e+00 \n", + "human.pedestrian.adult 1.235000e+03 2.140000e+02 \n", + "human.pedestrian.child 0.000000e+00 0.000000e+00 \n", + "human.pedestrian.construction_worker 0.000000e+00 2.000000e+01 \n", + "human.pedestrian.personal_mobility 0.000000e+00 0.000000e+00 \n", + "human.pedestrian.police_officer 0.000000e+00 0.000000e+00 \n", + "human.pedestrian.stroller 0.000000e+00 0.000000e+00 \n", + "human.pedestrian.wheelchair 0.000000e+00 0.000000e+00 \n", + "movable_object.barrier 3.000000e+00 7.490000e+02 \n", + "movable_object.debris 0.000000e+00 0.000000e+00 \n", + "movable_object.pushable_pullable 0.000000e+00 0.000000e+00 \n", + "movable_object.trafficcone 2.000000e+01 9.700000e+01 \n", + "static_object.bicycle_rack 0.000000e+00 0.000000e+00 \n", + "vehicle.bicycle 0.000000e+00 0.000000e+00 \n", + "vehicle.bus.bendy 0.000000e+00 0.000000e+00 \n", + "vehicle.bus.rigid 1.400000e+02 8.200000e+01 \n", + "vehicle.car 3.026940e+05 5.070000e+02 \n", + "vehicle.construction 8.930000e+02 1.243870e+05 \n", + "vehicle.emergency.ambulance 0.000000e+00 0.000000e+00 \n", + "vehicle.emergency.police 0.000000e+00 0.000000e+00 \n", + "vehicle.motorcycle 1.369000e+03 1.490000e+02 \n", + "vehicle.trailer 0.000000e+00 2.900000e+01 \n", + "vehicle.truck 5.173000e+03 1.542000e+03 \n", + "flat.driveable_surface 3.481000e+03 2.910000e+02 \n", + "\n", + " vehicle.emergency.ambulance \\\n", + "tp 0.0 \n", + "fp 0.0 \n", + "fn 0.0 \n", + "tn 3326172.0 \n", + "precision NaN \n", + "recall NaN \n", + "accuracy 1.0 \n", + "f1_score NaN \n", + "iou NaN \n", + "dice_score NaN \n", + "animal 0.0 \n", + "human.pedestrian.adult 0.0 \n", + "human.pedestrian.child 0.0 \n", + "human.pedestrian.construction_worker 0.0 \n", + "human.pedestrian.personal_mobility 0.0 \n", + "human.pedestrian.police_officer 0.0 \n", + "human.pedestrian.stroller 0.0 \n", + "human.pedestrian.wheelchair 0.0 \n", + "movable_object.barrier 0.0 \n", + "movable_object.debris 0.0 \n", + "movable_object.pushable_pullable 0.0 \n", + "movable_object.trafficcone 0.0 \n", + "static_object.bicycle_rack 0.0 \n", + "vehicle.bicycle 0.0 \n", + "vehicle.bus.bendy 0.0 \n", + "vehicle.bus.rigid 0.0 \n", + "vehicle.car 0.0 \n", + "vehicle.construction 0.0 \n", + "vehicle.emergency.ambulance 0.0 \n", + "vehicle.emergency.police 0.0 \n", + "vehicle.motorcycle 0.0 \n", + "vehicle.trailer 0.0 \n", + "vehicle.truck 0.0 \n", + "flat.driveable_surface 0.0 \n", + "\n", + " vehicle.emergency.police \\\n", + "tp 0.0 \n", + "fp 0.0 \n", + "fn 0.0 \n", + "tn 3326172.0 \n", + "precision NaN \n", + "recall NaN \n", + "accuracy 1.0 \n", + "f1_score NaN \n", + "iou NaN \n", + "dice_score NaN \n", + "animal 0.0 \n", + "human.pedestrian.adult 0.0 \n", + "human.pedestrian.child 0.0 \n", + "human.pedestrian.construction_worker 0.0 \n", + "human.pedestrian.personal_mobility 0.0 \n", + "human.pedestrian.police_officer 0.0 \n", + "human.pedestrian.stroller 0.0 \n", + "human.pedestrian.wheelchair 0.0 \n", + "movable_object.barrier 0.0 \n", + "movable_object.debris 0.0 \n", + "movable_object.pushable_pullable 0.0 \n", + "movable_object.trafficcone 0.0 \n", + "static_object.bicycle_rack 0.0 \n", + "vehicle.bicycle 0.0 \n", + "vehicle.bus.bendy 0.0 \n", + "vehicle.bus.rigid 0.0 \n", + "vehicle.car 0.0 \n", + "vehicle.construction 0.0 \n", + "vehicle.emergency.ambulance 0.0 \n", + "vehicle.emergency.police 0.0 \n", + "vehicle.motorcycle 0.0 \n", + "vehicle.trailer 0.0 \n", + "vehicle.truck 0.0 \n", + "flat.driveable_surface 0.0 \n", + "\n", + " vehicle.motorcycle vehicle.trailer \\\n", + "tp 2.299100e+04 6.436000e+03 \n", + "fp 3.728000e+03 5.490000e+02 \n", + "fn 1.579000e+03 1.489800e+04 \n", + "tn 3.297874e+06 3.304289e+06 \n", + "precision 8.604738e-01 9.214030e-01 \n", + "recall 9.357346e-01 3.016781e-01 \n", + "accuracy 9.984045e-01 9.953559e-01 \n", + "f1_score 8.965275e-01 4.545358e-01 \n", + "iou 8.124602e-01 2.941096e-01 \n", + "dice_score 8.965275e-01 4.545358e-01 \n", + "animal 0.000000e+00 0.000000e+00 \n", + "human.pedestrian.adult 6.370000e+02 0.000000e+00 \n", + "human.pedestrian.child 0.000000e+00 0.000000e+00 \n", + "human.pedestrian.construction_worker 0.000000e+00 0.000000e+00 \n", + "human.pedestrian.personal_mobility 0.000000e+00 0.000000e+00 \n", + "human.pedestrian.police_officer 0.000000e+00 0.000000e+00 \n", + "human.pedestrian.stroller 0.000000e+00 0.000000e+00 \n", + "human.pedestrian.wheelchair 0.000000e+00 0.000000e+00 \n", + "movable_object.barrier 0.000000e+00 0.000000e+00 \n", + "movable_object.debris 0.000000e+00 0.000000e+00 \n", + "movable_object.pushable_pullable 0.000000e+00 0.000000e+00 \n", + "movable_object.trafficcone 0.000000e+00 4.600000e+01 \n", + "static_object.bicycle_rack 4.000000e+01 0.000000e+00 \n", + "vehicle.bicycle 0.000000e+00 0.000000e+00 \n", + "vehicle.bus.bendy 0.000000e+00 0.000000e+00 \n", + "vehicle.bus.rigid 0.000000e+00 0.000000e+00 \n", + "vehicle.car 7.370000e+02 1.821000e+03 \n", + "vehicle.construction 0.000000e+00 0.000000e+00 \n", + "vehicle.emergency.ambulance 0.000000e+00 0.000000e+00 \n", + "vehicle.emergency.police 0.000000e+00 0.000000e+00 \n", + "vehicle.motorcycle 2.299100e+04 0.000000e+00 \n", + "vehicle.trailer 0.000000e+00 6.436000e+03 \n", + "vehicle.truck 4.000000e+00 1.282800e+04 \n", + "flat.driveable_surface 1.610000e+02 2.030000e+02 \n", + "\n", + " vehicle.truck flat.driveable_surface \\\n", + "tp 7.031500e+04 2.559201e+06 \n", + "fp 2.202200e+04 8.568000e+03 \n", + "fn 9.345000e+03 9.131000e+03 \n", + "tn 3.224490e+06 7.492720e+05 \n", + "precision 7.615041e-01 9.966633e-01 \n", + "recall 8.826889e-01 9.964448e-01 \n", + "accuracy 9.905696e-01 9.946789e-01 \n", + "f1_score 8.176305e-01 9.965540e-01 \n", + "iou 6.915187e-01 9.931317e-01 \n", + "dice_score 8.176305e-01 9.965540e-01 \n", + "animal 0.000000e+00 0.000000e+00 \n", + "human.pedestrian.adult 6.580000e+02 1.288000e+03 \n", + "human.pedestrian.child 0.000000e+00 0.000000e+00 \n", + "human.pedestrian.construction_worker 8.000000e+00 0.000000e+00 \n", + "human.pedestrian.personal_mobility 0.000000e+00 0.000000e+00 \n", + "human.pedestrian.police_officer 0.000000e+00 0.000000e+00 \n", + "human.pedestrian.stroller 0.000000e+00 0.000000e+00 \n", + "human.pedestrian.wheelchair 0.000000e+00 0.000000e+00 \n", + "movable_object.barrier 3.480000e+02 1.260000e+03 \n", + "movable_object.debris 0.000000e+00 0.000000e+00 \n", + "movable_object.pushable_pullable 0.000000e+00 0.000000e+00 \n", + "movable_object.trafficcone 1.360000e+02 4.700000e+02 \n", + "static_object.bicycle_rack 0.000000e+00 0.000000e+00 \n", + "vehicle.bicycle 0.000000e+00 0.000000e+00 \n", + "vehicle.bus.bendy 0.000000e+00 0.000000e+00 \n", + "vehicle.bus.rigid 1.290000e+02 1.900000e+02 \n", + "vehicle.car 6.235000e+03 4.619000e+03 \n", + "vehicle.construction 8.140000e+02 9.100000e+01 \n", + "vehicle.emergency.ambulance 0.000000e+00 0.000000e+00 \n", + "vehicle.emergency.police 0.000000e+00 0.000000e+00 \n", + "vehicle.motorcycle 0.000000e+00 4.260000e+02 \n", + "vehicle.trailer 4.100000e+02 0.000000e+00 \n", + "vehicle.truck 7.031500e+04 7.870000e+02 \n", + "flat.driveable_surface 6.070000e+02 2.559201e+06 \n", + "\n", + " macro micro \n", + "tp NaN NaN \n", + "fp NaN NaN \n", + "fn NaN NaN \n", + "tn NaN NaN \n", + "precision 0.789279 0.978781 \n", + "recall 0.676058 0.978781 \n", + "accuracy 0.998232 0.998232 \n", + "f1_score 0.789115 0.978781 \n", + "iou 0.617749 0.958443 \n", + "dice_score 0.690476 0.978781 \n", + "animal NaN NaN \n", + "human.pedestrian.adult NaN NaN \n", + "human.pedestrian.child NaN NaN \n", + "human.pedestrian.construction_worker NaN NaN \n", + "human.pedestrian.personal_mobility NaN NaN \n", + "human.pedestrian.police_officer NaN NaN \n", + "human.pedestrian.stroller NaN NaN \n", + "human.pedestrian.wheelchair NaN NaN \n", + "movable_object.barrier NaN NaN \n", + "movable_object.debris NaN NaN \n", + "movable_object.pushable_pullable NaN NaN \n", + "movable_object.trafficcone NaN NaN \n", + "static_object.bicycle_rack NaN NaN \n", + "vehicle.bicycle NaN NaN \n", + "vehicle.bus.bendy NaN NaN \n", + "vehicle.bus.rigid NaN NaN \n", + "vehicle.car NaN NaN \n", + "vehicle.construction NaN NaN \n", + "vehicle.emergency.ambulance NaN NaN \n", + "vehicle.emergency.police NaN NaN \n", + "vehicle.motorcycle NaN NaN \n", + "vehicle.trailer NaN NaN \n", + "vehicle.truck NaN NaN \n", + "flat.driveable_surface NaN NaN \n", + "\n", + "[34 rows x 26 columns]" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], "source": [ - "# Inspect overlap first\n", - "model_classes = set(model.ontology.keys())\n", - "dataset_classes = set(dataset.ontology.keys())\n", - "print(\"Model classes:\", len(model_classes))\n", - "print(\"Dataset classes:\", len(dataset_classes))\n", - "print(\n", - " \"Overlap:\",\n", - " len(model_classes & dataset_classes),\n", - " sorted(model_classes & dataset_classes)[:20],\n", - ")\n", - "\n", - "# Build a safe translation map\n", - "# old_ontology = dataset.ontology, new_ontology = model.ontology\n", - "# dataset ontology -> model ontology\n", - "fallback = list(model.ontology.keys())[0]\n", - "ontology_translation = {\n", - " class_name: (class_name if class_name in model.ontology else fallback)\n", + "# Translate original nuImages dataset labels into the model ontology.\n", + "# Classes ignored by the model config are masked before this translation is applied.\n", + "fallback_class = next(iter(pm_model.ontology.keys()))\n", + "dataset_to_model_translation = {\n", + " class_name: (class_name if class_name in pm_model.ontology else fallback_class)\n", " for class_name in dataset.ontology.keys()\n", "}\n", "\n", - "# TorchImageSegmentationModel.eval expects a JSON filename, not a Python dict.\n", "translation_path = Path(\"local/data/nuimages_to_model_ontology_translation.json\")\n", "translation_path.parent.mkdir(parents=True, exist_ok=True)\n", "with open(translation_path, \"w\", encoding=\"utf-8\") as f:\n", - " json.dump(ontology_translation, f, indent=2)\n", + " json.dump(dataset_to_model_translation, f, indent=2)\n", "\n", - "results = model.eval(\n", + "results = pm_model.eval(\n", " dataset,\n", - " split=\"train\",\n", + " split=\"val\",\n", " ontology_translation=str(translation_path),\n", + " translation_direction=\"dataset_to_model\",\n", ")\n", - "display(results)" + "display(results)\n" ] } ], From 075c59c5f679ff2eee6135ab4181f804e59bea3e Mon Sep 17 00:00:00 2001 From: tejass Date: Wed, 1 Jul 2026 19:25:31 +0200 Subject: [PATCH 07/10] Rebasing tutorials --- examples/tutorial_nuimages_segmentation.ipynb | 1757 +---------------- 1 file changed, 73 insertions(+), 1684 deletions(-) diff --git a/examples/tutorial_nuimages_segmentation.ipynb b/examples/tutorial_nuimages_segmentation.ipynb index f2b435ec..68086277 100644 --- a/examples/tutorial_nuimages_segmentation.ipynb +++ b/examples/tutorial_nuimages_segmentation.ipynb @@ -1,112 +1,47 @@ { "cells": [ - { - "cell_type": "markdown", - "id": "eeb1b612", - "metadata": {}, - "source": [ - "# NuImages Segmentation Evaluation Tutorial\n" - ] - }, { "cell_type": "code", - "execution_count": 1, + "execution_count": null, "id": "478f413c", "metadata": {}, - "outputs": [ - { - "name": "stdout", - "output_type": "stream", - "text": [ - "Requirement already satisfied: transformers in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (4.30.2)\n", - "Requirement already satisfied: filelock in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from transformers) (3.14.0)\n", - "Requirement already satisfied: huggingface-hub<1.0,>=0.14.1 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from transformers) (0.36.2)\n", - "Requirement already satisfied: numpy>=1.17 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from transformers) (1.26.4)\n", - "Requirement already satisfied: packaging>=20.0 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from transformers) (24.2)\n", - "Requirement already satisfied: pyyaml>=5.1 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from transformers) (6.0.3)\n", - "Requirement already satisfied: regex!=2019.12.17 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from transformers) (2026.5.9)\n", - "Requirement already satisfied: requests in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from transformers) (2.28.2)\n", - "Requirement already satisfied: tokenizers!=0.11.3,<0.14,>=0.11.1 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from transformers) (0.13.3)\n", - "Requirement already satisfied: safetensors>=0.3.1 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from transformers) (0.7.0)\n", - "Requirement already satisfied: tqdm>=4.27 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from transformers) (4.65.2)\n", - "Requirement already satisfied: fsspec>=2023.5.0 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from huggingface-hub<1.0,>=0.14.1->transformers) (2025.12.0)\n", - "Requirement already satisfied: hf-xet<2.0.0,>=1.1.3 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from huggingface-hub<1.0,>=0.14.1->transformers) (1.5.0)\n", - "Requirement already satisfied: typing-extensions>=3.7.4.3 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from huggingface-hub<1.0,>=0.14.1->transformers) (4.15.0)\n", - "Requirement already satisfied: charset-normalizer<4,>=2 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from requests->transformers) (3.4.7)\n", - "Requirement already satisfied: idna<4,>=2.5 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from requests->transformers) (3.16)\n", - "Requirement already satisfied: urllib3<1.27,>=1.21.1 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from requests->transformers) (1.26.20)\n", - "Requirement already satisfied: certifi>=2017.4.17 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from requests->transformers) (2026.5.20)\n" - ] - } - ], + "outputs": [], "source": [ - "!pip install transformers" + "!pip install gdown" ] }, { - "cell_type": "markdown", - "id": "13273f08", + "cell_type": "code", + "execution_count": null, + "id": "a37f27c1", "metadata": {}, + "outputs": [], "source": [ - "## Download model from Hugging Face" + "# Download\n", + "!gdown 15Py3AKyGzWNX4M3OJKlQQsnH6IehkoEd -O local/data/image_segmentation_model.pth # model\n", + "!gdown 17eJ6aei6yBAbK1aA4mhTeMdXYycorqcT -O local/data/image_segmentation_model_cfg.json # configuration\n", + "!gdown 1spXt5_ISG1ZaHHO2DTtCSqgkohnhwuzt -O local/data/image_segmentation_model_ontology.json # ontology" ] }, { "cell_type": "code", - "execution_count": 2, + "execution_count": null, "id": "2a47f7a9", "metadata": {}, - "outputs": [ - { - "name": "stderr", - "output_type": "stream", - "text": [ - "/home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages/huggingface_hub/file_download.py:949: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.\n", - " warnings.warn(\n", - "/home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages/transformers/models/segformer/image_processing_segformer.py:99: FutureWarning: The `reduce_labels` parameter is deprecated and will be removed in a future version. Please use `do_reduce_labels` instead.\n", - " warnings.warn(\n" - ] - } - ], + "outputs": [], "source": [ - "import json\n", - "from transformers import AutoImageProcessor\n", - "\n", - "model_name = \"tejasstanley/segformer-nuimages-b0-full-bs8\"\n", - "\n", - "processor = AutoImageProcessor.from_pretrained(model_name)\n", - "cfg = processor.to_dict()\n", - "\n", - "pm_cfg = {\n", - " \"resize\": {\n", - " \"height\": cfg[\"size\"][\"height\"],\n", - " \"width\": cfg[\"size\"][\"width\"],\n", - " },\n", - " \"normalization\": {\n", - " \"mean\": list(cfg[\"image_mean\"]),\n", - " \"std\": list(cfg[\"image_std\"]),\n", - " },\n", - " \"ignored_classes\": [\"background\", \"vehicle.ego\"],\n", - "}\n", - "\n", - "model_cfg_path = \"local/data/segformer_nuimages_b0_full_bs8_cfg.json\"\n", + "from perceptionmetrics.models.torch_segmentation import TorchImageSegmentationModel\n", "\n", - "with open(model_cfg_path, \"w\", encoding=\"utf-8\") as f:\n", - " json.dump(pm_cfg, f, indent=2)" - ] - }, - { - "cell_type": "markdown", - "id": "f7f0b0e0", - "metadata": {}, - "source": [ - "## Build Dataset and Model Ontology\n", - "\n" + "model = TorchImageSegmentationModel(\n", + " model=\"local/data/image_segmentation_model.pth\", \n", + " model_cfg=\"local/data/image_segmentation_model_cfg.json\",\n", + " ontology_fname=\"local/data/image_segmentation_model_ontology.json\",\n", + ")" ] }, { "cell_type": "code", - "execution_count": 3, + "execution_count": 5, "id": "444ef811", "metadata": {}, "outputs": [ @@ -114,149 +49,60 @@ "name": "stdout", "output_type": "stream", "text": [ - "Jupyter environment detected. Enabling Open3D WebVisualizer.\n", - "[Open3D INFO] WebRTC GUI backend enabled.\n", - "[Open3D INFO] WebRTCWindowSystem: HTTP handshake server disabled.\n", "Built nuImages segmentation dataset with 50 samples and 26 classes.\n", - "Saved model ontology to local/data/nuimages_model_ontology.json\n", - "Dataset labels will be translated in memory during evaluation.\n" + "Saved compact ontology to local/data/nuimages_segmentation_ontology_compact.json with 26 classes (max idx=25)\n" ] } ], "source": [ + "## NuImages Segmentation Dataset and Model\n", + "\n", "import json\n", "from pathlib import Path\n", "\n", + "from perceptionmetrics.models.torch_segmentation import TorchImageSegmentationModel\n", "from perceptionmetrics.datasets.nuimages import NuImagesSegmentationDataset\n", "\n", "dataset = NuImagesSegmentationDataset(\n", - " dataset_dir=\"local/data/nuimages-v1.0-mini\", version=\"v1.0-mini\", split=\"val\"\n", + " dataset_dir=\"local/data/nuimages-v1.0-mini\", version=\"v1.0-mini\", split=\"train\"\n", ")\n", "\n", - "model_labels = {\n", - " 0: \"animal\",\n", - " 1: \"human.pedestrian.adult\",\n", - " 2: \"human.pedestrian.child\",\n", - " 3: \"human.pedestrian.construction_worker\",\n", - " 4: \"human.pedestrian.personal_mobility\",\n", - " 5: \"human.pedestrian.police_officer\",\n", - " 6: \"human.pedestrian.stroller\",\n", - " 7: \"human.pedestrian.wheelchair\",\n", - " 8: \"movable_object.barrier\",\n", - " 9: \"movable_object.debris\",\n", - " 10: \"movable_object.pushable_pullable\",\n", - " 11: \"movable_object.trafficcone\",\n", - " 12: \"static_object.bicycle_rack\",\n", - " 13: \"vehicle.bicycle\",\n", - " 14: \"vehicle.bus.bendy\",\n", - " 15: \"vehicle.bus.rigid\",\n", - " 16: \"vehicle.car\",\n", - " 17: \"vehicle.construction\",\n", - " 18: \"vehicle.emergency.ambulance\",\n", - " 19: \"vehicle.emergency.police\",\n", - " 20: \"vehicle.motorcycle\",\n", - " 21: \"vehicle.trailer\",\n", - " 22: \"vehicle.truck\",\n", - " 23: \"flat.driveable_surface\",\n", - "}\n", - "\n", - "model_ontology = {\n", + "# Build compact ontology (contiguous idx: 0..N-1) to avoid sparse-index issues (e.g., idx 31).\n", + "sorted_items = sorted(dataset.ontology.items(), key=lambda x: x[1][\"idx\"] )\n", + "compact_ontology = {\n", " class_name: {\n", - " \"idx\": class_idx,\n", - " \"rgb\": dataset.ontology[class_name][\"rgb\"],\n", + " \"idx\": new_idx,\n", + " \"rgb\": class_meta.get(\"rgb\", [0, 0, 0]),\n", " }\n", - " for class_idx, class_name in model_labels.items()\n", + " for new_idx, (class_name, class_meta) in enumerate(sorted_items)\n", "}\n", "\n", - "model_ontology_path = Path(\"local/data/nuimages_model_ontology.json\")\n", - "model_ontology_path.parent.mkdir(parents=True, exist_ok=True)\n", - "with open(model_ontology_path, \"w\", encoding=\"utf-8\") as f:\n", - " json.dump(model_ontology, f, indent=2)\n", + "ontology_path = Path(\"local/data/nuimages_segmentation_ontology_compact.json\")\n", + "ontology_path.parent.mkdir(parents=True, exist_ok=True)\n", + "with open(ontology_path, \"w\") as f:\n", + " json.dump(compact_ontology, f, indent=2)\n", "\n", - "print(f\"Saved model ontology to {model_ontology_path}\")\n", - "print(\"Dataset labels will be translated in memory during evaluation.\")\n" - ] - }, - { - "cell_type": "markdown", - "id": "4e68ee53", - "metadata": {}, - "source": [ - "## Initialize the PerceptionMetrics Model Wrapper\n", - "\n" - ] - }, - { - "cell_type": "code", - "execution_count": 4, - "id": "e34db9f2", - "metadata": {}, - "outputs": [ - { - "name": "stdout", - "output_type": "stream", - "text": [ - "Torch detection not available\n", - "Tensorflow not available\n" - ] - }, - { - "name": "stderr", - "output_type": "stream", - "text": [ - "/home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages/torchvision/datapoints/__init__.py:12: UserWarning: The torchvision.datapoints and torchvision.transforms.v2 namespaces are still Beta. While we do not expect major breaking changes, some APIs may still change according to user feedback. Please submit any feedback you may have in this issue: https://github.com/pytorch/vision/issues/6753, and you can also check out https://github.com/pytorch/vision/issues/7319 to learn more about the APIs that we suspect might involve future changes. You can silence this warning by calling torchvision.disable_beta_transforms_warning().\n", - " warnings.warn(_BETA_TRANSFORMS_WARNING)\n", - "/home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages/torchvision/transforms/v2/__init__.py:54: UserWarning: The torchvision.datapoints and torchvision.transforms.v2 namespaces are still Beta. While we do not expect major breaking changes, some APIs may still change according to user feedback. Please submit any feedback you may have in this issue: https://github.com/pytorch/vision/issues/6753, and you can also check out https://github.com/pytorch/vision/issues/7319 to learn more about the APIs that we suspect might involve future changes. You can silence this warning by calling torchvision.disable_beta_transforms_warning().\n", - " warnings.warn(_BETA_TRANSFORMS_WARNING)\n", - "/home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages/huggingface_hub/file_download.py:949: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.\n", - " warnings.warn(\n" - ] - } - ], - "source": [ - "from transformers import SegformerForSemanticSegmentation\n", - "from perceptionmetrics.models.torch_segmentation import TorchImageSegmentationModel\n", - "\n", - "model_name = \"tejasstanley/segformer-nuimages-b0-full-bs8\"\n", - "model_cfg_path = \"local/data/segformer_nuimages_b0_full_bs8_cfg.json\"\n", - "\n", - "\n", - "hf_model = SegformerForSemanticSegmentation.from_pretrained(model_name)\n", + "max_idx = max(v[\"idx\"] for v in compact_ontology.values())\n", + "print(\n", + " f\"Saved compact ontology to {ontology_path} with {len(compact_ontology)} classes (max idx={max_idx})\"\n", + " )\n", "\n", - "pm_model = TorchImageSegmentationModel(\n", - " model=hf_model,\n", - " model_cfg=model_cfg_path,\n", - " ontology_fname=str(model_ontology_path),\n", + "model = TorchImageSegmentationModel(\n", + " model=\"local/data/image_segmentation_model.pth\",\n", + " model_cfg=\"local/data/image_segmentation_model_cfg.json\",\n", + " ontology_fname=str(ontology_path),\n", ")" ] }, - { - "cell_type": "markdown", - "id": "16b24b8a", - "metadata": {}, - "source": [ - "## Inspect a Sample Prediction\n" - ] - }, { "cell_type": "code", - "execution_count": 5, + "execution_count": 6, "id": "47f1727b", "metadata": {}, "outputs": [ - { - "name": "stdout", - "output_type": "stream", - "text": [ - "Predicted class ids: [1, 3, 8, 11, 13, 16, 21, 22, 23]\n", - "Predicted class names: ['human.pedestrian.adult', 'human.pedestrian.construction_worker', 'movable_object.barrier', 'movable_object.trafficcone', 'vehicle.bicycle', 'vehicle.car', 'vehicle.trailer', 'vehicle.truck', 'flat.driveable_surface']\n", - "Dataset label ids in sample: [0, 4, 9, 17, 23, 24]\n", - "Dataset label names in sample: ['background', 'human.pedestrian.construction_worker', 'movable_object.barrier', 'vehicle.car', 'vehicle.truck', 'flat.driveable_surface']\n" - ] - }, { "data": { - "image/png": "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", + "image/png": "iVBORw0KGgoAAAANSUhEUgAAAxoAAACqCAYAAAAqYjm6AAAAOnRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjEwLjgsIGh0dHBzOi8vbWF0cGxvdGxpYi5vcmcvwVt1zgAAAAlwSFlzAAAPYQAAD2EBqD+naQABAABJREFUeJzs/Xm0bWdZ549+3mY2q9nt2fv0Jyd9QwiERlBEwChEGhUpG0SvisMqqhRQ6+fwZ1G3yipl6C1La4Blxy2r8HcVBhalpaWCtBEQAWmCpCEhzUly2t03q5vN29w/3jnX3jvnJNlJTpITmN+RnbP3WnPN9c7+ed7n+/0+wnvvadCgQYMGDRo0aNCgQYPzCPlUD6BBgwYNGjRo0KBBgwZff2gSjQYNGjRo0KBBgwYNGpx3NIlGgwYNGjRo0KBBgwYNzjuaRKNBgwYNGjRo0KBBgwbnHU2i0aBBgwYNGjRo0KBBg/OOJtFo0KBBgwYNGjRo0KDBeUeTaDRo0KBBgwYNGjRo0OC8o0k0GjRo0KBBgwYNGjRocN7RJBoNGjRo0KBBgwYNGjQ472gSjQYNGjRosAMXX3wxP/ETPzH+++/+7u8QQvB3f/d35+07hBD8h//wH87b+ho0aNCgwYWHJtG4wPBHf/RHCCH4whe+8FQPpUGDBk8R6vtA/ZOmKVdeeSVvfvObWVhYeKqHt2t84AMfaJKJBg2+QXHs2DHe/OY3c+WVV9Jut2m32zzjGc/gZ37mZ/jKV77Cy172sh33uYf6ae4hT2/op3oADRo0aNDg3PiVX/kVLrnkErIs4+///u/5/d//fT7wgQ9w66230m63n7RxvOQlL2E0GhHH8aP63Ac+8AF+93d/95yBwmg0QuvmEdSgwdcj/vqv/5of+qEfQmvNj/zIj/DsZz8bKSV33HEHf/7nf87v//7v8+53v5uf+qmfGn/m85//PL/927/N2972Nq655prx68961rOeik1ocJ7Q3OUbNGjQ4ALFK1/5Sp7//OcD8FM/9VPs2bOH//Jf/gt/+Zd/yQ//8A+ftfxgMKDT6Zz3cUgpSdP0vK7zfK+vQYMGFwbuueceXv/613P06FE+9rGPceDAgR3v/6f/9J/4vd/7PW644QaOHDkyfj1NU377t3+bl7/85bzsZS97kkfd4IlCQ526wPETP/ETdLtdHnjgAV7zmtfQ7XY5dOgQv/u7vwvALbfcwg033ECn0+Ho0aO8973v3fH51dVVfuEXfoHrrruObrfL5OQkr3zlK/mnf/qns77r/vvv53u+53vodDrs3buXn//5n+dDH/rQObnZn/vc5/iu7/oupqamaLfbvPSlL+XTn/70E7YfGjRoADfccAMQKAn1veGee+7hVa96FRMTE/zIj/wIAM453vGOd3DttdeSpin79u3jTW96E2trazvW573n7W9/O4cPH6bdbvPt3/7t3HbbbWd970NpND73uc/xqle9ipmZGTqdDs961rN45zvfCYR7V32f2k6DqHEuSsTNN9/MK1/5SiYnJ+l2u3zHd3wHn/3sZ3csU9PKPv3pT/Ov//W/Zn5+nk6nw/d93/extLT06HdqgwYNzit+4zd+g8FgwLvf/e6zkgwArTVvfetbdyQZDb5+0VQ0ngaw1vLKV76Sl7zkJfzGb/wG73nPe3jzm99Mp9Ph3/7bf8uP/MiP8LrXvY4/+IM/4Md+7Mf4lm/5Fi655BIA7r33Xv7iL/6CH/iBH+CSSy5hYWGBd73rXbz0pS/l9ttv5+DBg0CYCb3hhhs4ffo0P/uzP8v+/ft573vfy0033XTWeD7+8Y/zyle+kuc973n88i//MlJK3v3ud3PDDTfwqU99ihe84AVP6v5p0OAbBffccw8Ae/bsAcAYw4033siLX/xifvM3f3NMp3rTm97EH/3RH/HGN76Rt771rRw7dozf+Z3f4eabb+bTn/40URQB8O///b/n7W9/O6961at41atexZe+9CVe8YpXUBTFI47lIx/5CK95zWs4cODA+J7x1a9+lb/+67/mZ3/2Z3nTm97EqVOn+MhHPsIf//EfP+L6brvtNr7t276NyclJfvEXf5EoinjXu97Fy172Mj7xiU/wwhe+cMfyb3nLW5iZmeGXf/mXue+++3jHO97Bm9/8Zv70T//0Ue3TBg0anF/89V//NZdffvlZ12yDb1D4BhcU3v3ud3vAf/7zn/fee//jP/7jHvC/9mu/Nl5mbW3Nt1otL4Tw73vf+8av33HHHR7wv/zLvzx+Lcsyb63d8R3Hjh3zSZL4X/mVXxm/9lu/9Vse8H/xF38xfm00Gvmrr77aA/6mm27y3nvvnPNXXHGFv/HGG71zbrzscDj0l1xyiX/5y19+XvZDgwbfyKjvAx/96Ef90tKSP378uH/f+97n9+zZ41utlj9x4sT43vBLv/RLOz77qU99ygP+Pe95z47X//Zv/3bH64uLiz6OY//qV796x7X8tre9zQP+x3/8x8ev3XTTTTvuA8YYf8kll/ijR4/6tbW1Hd+zfV0/8zM/4x/qMfPge9VrX/taH8exv+eee8avnTp1yk9MTPiXvOQlZ+2b7/zO79zxXT//8z/vlVJ+fX39nN/XoEGDJx4bGxse8K997WvPem9tbc0vLS2Nf4bD4Y733//+9++4zzT4+kBDnXqaYLtganp6mquuuopOp8MP/uAPjl+/6qqrmJ6e5t577x2/liQJUobDbK1lZWWFbrfLVVddxZe+9KXxcn/7t3/LoUOH+J7v+Z7xa2ma8s//+T/fMY4vf/nL3HXXXbzhDW9gZWWF5eVllpeXGQwGfMd3fAef/OQncc6d9+1v0OAbEd/5nd/J/Pw8R44c4fWvfz3dbpf//b//N4cOHRov86/+1b/a8Zn3v//9TE1N8fKXv3x8fS4vL/O85z2Pbrc7rlJ+9KMfpSgK3vKWt+ygNP3cz/3cI47r5ptv5tixY/zcz/0c09PTO97bvq7dwlrLhz/8YV772tdy6aWXjl8/cOAAb3jDG/j7v/97Njc3d3zmX/yLf7Hju77t274Nay3333//o/7+Bg0anB/U12m32z3rvZe97GXMz8+Pf2pqZYOvbzTUqacB0jRlfn5+x2tTU1McPnz4rIf61NTUDh62c453vvOd/N7v/R7Hjh3DWjt+r6ZfQNBnXHbZZWet7/LLL9/x91133QXAj//4jz/keDc2NpiZmdnl1jVo0OCh8Lu/+7tceeWVaK3Zt28fV1111XjiAALX+fDhwzs+c9ddd7GxscHevXvPuc7FxUWAcUB+xRVX7Hh/fn7+Ea/fmsL1zGc+89Ft0ENgaWmJ4XDIVVddddZ711xzDc45jh8/zrXXXjt+/aKLLtqxXD3mB+tQGjRo8ORhYmICgH6/f9Z773rXu+j1eiwsLPCjP/qjT/bQGjxFaBKNpwGUUo/qde/9+Pdf+7Vf49/9u3/HT/7kT/Krv/qrzM7OIqXk537u5x5T5aH+zH/+z/+Z66+//pzLnGsmo0GDBo8eL3jBC8auU+fC9oplDecce/fu5T3vec85P/PgSYunK3Zz/2vQoMGTi6mpKQ4cOMCtt9561nu1ZuO+++57kkfV4KlEk2h8neN//a//xbd/+7fz3//7f9/x+vr6OnNzc+O/jx49yu233473fkdV4+67797xucsuuwyAyclJvvM7v/MJHHmDBg0eCy677DI++tGP8q3f+q20Wq2HXO7o0aNAqIBspystLS09YlWgvg/ceuutD3sf2C2Nan5+nna7zZ133nnWe3fccQdSysahpkGDpwle/epX84d/+If84z/+Y2MO06Cxt/16h1LqrBm+97///Zw8eXLHazfeeCMnT57k//yf/zN+Lcsy/tt/+287lnve857HZZddxm/+5m+eszTa2Es2aPDU4gd/8Aex1vKrv/qrZ71njGF9fR0I+o8oiviv//W/7rhHvOMd73jE73juc5/LJZdcwjve8Y7x+mpsX1fd0+PByzwYSile8YpX8Jd/+Zc7ZjsXFhZ473vfy4tf/GImJycfcVwNGjR46vGLv/iLtNttfvInf5KFhYWz3m+qjt9YaCoaX+d4zWtew6/8yq/wxje+kRe96EXccsstvOc979kxgwnBDvN3fud3+OEf/mF+9md/lgMHDvCe97xn3FSrnpmUUvKHf/iHvPKVr+Taa6/ljW98I4cOHeLkyZPcdNNNTE5O8ld/9VdP+nY2aNAg4KUvfSlvetOb+PVf/3W+/OUv84pXvIIoirjrrrt4//vfzzvf+U6+//u/n/n5eX7hF36BX//1X+c1r3kNr3rVq7j55pv54Ac/uKPaeS5IKfn93/99vvu7v5vrr7+eN77xjRw4cIA77riD2267jQ996ENAmJgAeOtb38qNN96IUorXv/7151zn29/+dj7ykY/w4he/mJ/+6Z9Ga8273vUu8jznN37jN87vTmrQoMEThiuuuIL3vve9/PAP/zBXXXXVuDO4955jx47x3ve+FynlWfqyBl+faBKNr3O87W1vYzAY8N73vpc//dM/5bnPfS5/8zd/wy/90i/tWK7b7fLxj3+ct7zlLbzzne+k2+3yYz/2Y7zoRS/in/2zf7aji+/LXvYyPvOZz/Crv/qr/M7v/A79fp/9+/fzwhe+kDe96U1P9iY2aNDgQfiDP/gDnve85/Gud72Lt73tbWitufjii/nRH/1RvvVbv3W83Nvf/nbSNOUP/uAPuOmmm3jhC1/Ihz/8YV796lc/4nfceOON3HTTTfzH//gf+a3f+i2cc1x22WU7nOpe97rX8Za3vIX3ve99/Mmf/Ane+4dMNK699lo+9alP8W/+zb/h13/913HO8cIXvpA/+ZM/afz4GzR4muF7v/d7ueWWW/it3/otPvzhD/M//sf/QAjB0aNHefWrX82//Jf/kmc/+9lP9TAbPAkQvqlhNXgYvOMd7+Dnf/7nOXHixA5LzQYNGjRo0KBBgwYNHg5NotFgjNFotEM8mmUZz3nOc7DW8rWvfe0pHFmDBg0aNGjQoEGDpxsa6lSDMV73utdx0UUXcf3117OxscGf/MmfcMcddzykTWaDBg0aNGjQoEGDBg+FJtFoMMaNN97IH/7hH/Ke97wHay3PeMYzeN/73scP/dAPPdVDa9CgQYMGDRo0aPA0Q0OdatCgQYMGDRo0aNCgwXlH00ejQYMGDRo0aNCgQYMG5x1NotGgQYMGDRo0aNCgQYPzjibRaNCgQYMGDRo0aNCgwXnHrsXgv/v//SNmZ+cRQtJtddBaoiNJUWY4K0jihG4nxTmHUgpjSjzgvEcAWmuKvAAvSJKI9f4m3oMrS+742ldZXd1ASM9wMCCKY4QUCMBay2avh1aaKIqqDtWessjpD4ZYD955nPNIEVrbW2txziOEJNIarTVKSXQUkcQxadrGGEOaxiglCd8EHk+e56RxgpIKaw15kSOFQkiBlAIdR2R5xmAwpCwNUgiyLEMIGcYsBN47jPEIPAgwxuGwSATOGPAglKAsLc4JnCuQSmG9xzuPF6AReO+JIk273ebQgUO89CU30Ol0EULgnENKidZ6/HvdvVsphXMOrRQT3TZHLjqIEOA8eBd+HB7vwnfUrzvv8d6DD/uRavwi7BwgfE6IsMdE/bL31V81wsK+Wp9zlizPOH3mNLfd9k+cPn2aoiw5cOAQMzNzWOcQIhxrqSTGWkbDEVk+REqNtZ6lpQW6rTaFMZRlSZqmeO+wtgQEzoeB2qLE2BKpIgQW5yzOe4ajIXmeo7UGL2m1Uw4eOMj+w0cQKkIqSa83Yun0Scp8QJomaBWjZJ2Lh30lhK22q9o+AWXp2BgMQUcIqQGPcw7vDba05NmI1aVF1peXyMsBQsSgYuI4JkoilIxw+YB4coI0nURpjVKqOo/BOwcyHBelFM57/p/f/vXHcLk/9bj+429/qodwFrz3KCH4wcs77Gs/On+Mfun48AND7u+He8EThXDuCb51f8o37U14rMI6AfzTSs7HToyQ44ubh1yf8J6ODTNSFhgq8ELwxG3pQ6M+Tq862ubyqegx74MnAwL40PEht60W5+28cMD1CyPe/onTJMaF7a/vxd5zbDrmLa84TK7lro/Pl2/4f5+XsT2ZEE/gddagQYNHh93IvHf9VC3yHIFERRovPMaWFKXBe0EUxSHAxiMEyCrgxnukECgpUVKitQIUVO9553F4SlMCHmfL8BnCg208/B0bUgXDCLwH4QXeW0QIK8GHIM85j1ISL1yV8Di8d/hqrXWgXEXK45uX8AIbfkFEmlQr8rygt7lBFMW0RIfhMKMsTRVo1g/rajwC8AKBq1cd9omXeFE9pAUgJB4TAshqNNJ7rHNIIRFKgJR4KcnygqIoMdaGYFwqQFTb6aqg1o8TjJDshOQrL3K8rxIRz/j7ZL2diK3Eosotdpw4fus4OO+2XvZbx6X+/4Nv/847jLWcPn2CL33x86yurTHo91FKkcYp3e4EpbNIITDGIKSkKEqyLMNaWyVEjo3NTQpjQgIyGhDHEd4biqKgLA1p2qqOu8E5g1J6PC7nIc8L8lFGFEXEcYLxEi8UXoBUmrwo6PV6bKxvkA2HTHU7xFFUnyHj/bH9+Vb/7jwID9J5LOEa8NZRFhmj4SZkGVIqumnCqNXCxRGxTkhbHeIkQUYKZw3ZRomONEoqZJXceBEuBKkUeI/wPiS87kIOsZ5eqI/tpZOayfjRFXi999yzUfJA3/BEht51knH5lOa6PfHjCrA9cKijmW8plkcWEAjvmTawp/DMF57nbIQ7AwJiB8/ZcMTOczoV/OdLY86kggdPLTzR8D7cua+Yirikq8f33m+kkFMAt82n/Mdv288bbl3jGcsZstov30j7oUGDBk8v7DrRUEqjlcZZR+ELrCmJIoVSUVigSh7Cr1u3ve03QCEEWiu8d2Gm1hNmpb1D4HDWjWsLIZbauR6lFN65aibdVzO9Go8I1QIvEYRZ5xClVcmMdOAFvgrInbdVQhDW4YSoQ0Sg4pPVs9Z4dBTRmZjAGMNwMMSZqjqBD7OC26b9va+qBdXd34sqAJfhge4AgUAKiZQyLCer2oCoHuAiJAlCV8kPAmtKrDUY59BChX3gPdaG4N97j5R1BSFsp0BQFIbSWHQkkVSBsQTrPFIKrN15nOvAa3uy4X2obHhfb2OdfVQPOVEH5DuPufOekyeOc9NNNyGVRClFkmrarRadzhSRjjHe4T1Y6xDeU5bluCoFgmyU0e/38N6SZYNw7GSEtY7hMENHEdb5cB6ZsqoqVfsAMKVjNBqAEEil0DoGB1IqSuNYOHOGXn/IYJThgCSKq+qQHG+7ENu2e7xPquNePehd9f1lkZON+lAWKCnQaYyQGmUTJqdhQiuUTkICEWpeOA9SaiQKpAQhcFXWGvJYMa6sWO8ojKHB48NWJQ72tRUvOdiipeWjCtqGxvP5xfwJGmGF6pQ70lW87GCLRD3+kHJPqviWAj5Veg7nju87Yzg69OzLw71KPcQM1XQJ/+mOgj/fr/nAXkUp6/vTE4xqPBOx5NWLQ5778VPc+r1HGM4mW5M7Fxg8sK+luEsKSucf1Sz89r0vHvS7lYIvHGhz21zKyx7o8/1fXWd+aBDALfMtrBRnraNBgwYNnkrsnicgJA6PMRaLDdUCL3ZUHepA1FV0HCEk+DDjjggBlFIS50BphTEmzMjbKvyWCqUcUoqwfLWOdpKilCJSCounLB1aRSRpgvMeqTTSgvD17L7EKx8qA+MxCMrSgMjHVCrrHAKBExUdqKrK+JAtENICqKsVcZyA96wOh5RFESo3SRzoRL6iSlXL1+UB4QQCgRcgpAwVCR9m4upEI1C/QvAopcSLENhLH6oQXhjKig7mrMNLX+V1VQVEyWr/b6vMVNtivSfLC9JWhJAgfdgWJUJFQ3ioN/Os5CL8FrZqa1fsSEY8HuFFRafaqu7gPf3+Jp/61Cfw3tDtTqIFZLGkNJaJialQdZKSLC8RQmJt2Af1eVHT5jwOqiqajjQISTbKcM4iZTL+TJ5ltFqtULnyDuscw9EA7z1aRyitEFKiqkRmo9cjKwGpiFptrDUoWdMORLXRVZmHENA4J6oEWWCdoywLBoMe/V4P6x3SS1Itke02UkhUFCGjBOslLm5RlGV1PQiUkOF8lQqkQCsVjmVdHawSaCEEzlpMkWOLEWpbZanB7rH9/BZCkGrJVCx59lxMN3psQfP4in+Cgl2PZ29b8e2HWnSixy+pq8f7ujOGNxzPST2kdqsmGaqtD7ExHvbmjjceL3EC/mrfk9OGyQtQTvD9H57kO28t0e01nr9+N7e/5gjLV05Szd5ccLhmJubm5YL13D7ywhUccM1KxgtPDPjza6bpxeqsTRPAKJJ88LJJPnWkw6FeSct4bp1PcQIi5znYK1EVDffoRsEzljPsuY7rDY9nCxs0aNDgkbHrJ4W1nmGWIYUgjmIQ1f3dh2B2B91GjOe6Q7hW0/2rSWClFK04YWBt0HJ4i7EW7wylMRRlGSoDzhJpzWg0ovQ+VEGswTvQWpEVI2wVpEuvxvqBMCMdOC3OWkAhhMcKDyWURYHVEoREIUF6nA3VD2ttpccIN3gp6sqLrwJE0HGE9yEIzUYZ1hi0jsbJRF0dEVVFBCHHiUFIMEQ1Wx0SElFRvkJVIgSz9Xu+TthkSPQEHrzFV8Eo3lWkpZ3VJK1UqMZIFbQgVJUaEbbJuWrbJHgV3giz9NsCsvq55LeSii09xvakKlDOhJSBjyzAWMMXv/h5llcWuPLKK5DA/Nw8m5ubTExMY71kZXWVKI6x1qJ0oH2Fc80ihGAwGGCMoaaJpUkKSmGNpyhytFZIESpV2Wg4Lqp5L7HGM8iGWFcitUZEGqk1SgmssfSHPaJ2h7jdGieA1srwcAbC2W23FeoEzocqjTOOUTZic3OT0WiI8BbvLUnaItItpFYoHaHTFKGiQH3LM6yzocKlJFKo8eXihUCqUOGqk2wpFN5ZXJ5T5CMkjolWQmtyGiWj3V623/Coz2UlBLGWHOootBRcOqmZSxVTSaj0XYi873rse1uK2USd11lq4WG6rG7K9eTEI32m+reQcKz95O0vJ+A1n5/kDR+Zp1VIyn0Fk/HXOPLFFTYOtylb+kLMM3Zgt5UXASy1NV880CZ/mOpVfQe+cjXndXdskNityQfl4LL1HF1RLJ0Q3LyvxV9fMUXxoHX+9GPYlgYNGjR4NNh1olGaEqQKQZKqpYEiZBHOBaoH1QPbU2kF/FgQjAiz9YHyYyuhbEGe5ZSV/gAPQtUPDQ+m0j1IiQYsHi+C+DnQjWRYlxMYb6nD9jDzHLQOzjmct4AP1C/nGfb7tDsdSgqcUkgnMaVhlI9w1jElJ4miKmgVAus8WTZCSkkUR4FDLwP9JlEKP3RkoxEQxNsqiqp4vEp4tsELgVMV4UoIXFWFMSZUiUDivEHVMQAy0NVGw/Hsdq1RQVbVCbaCBSklEhFm8KUkTVNiHQXBN6DC4am0NNV8fZU/CCfGwc0OOlRVKfGu+l5Rp5DjNCMkZ1XFCqC3OeSeu+9h7/w8vV6PQwcOcnDfPtI4wiI5s7jC+voa3ckucdKiLEI1w5qQZBhjGAyHYZsgmAHEMdZL8nyEkhFSaRCC0WiEc4441niC+HyUj7DGEkUxCIlWCd5LjLFsbK5jRdA7eFlVYaRAoZDY6tx24+2v11kUBb1+n2GvV1EHI7qdFghJUeYkrRat1gQySiqBucA6SyoShFQY62rlzni/aSFARuhq30mlccZg8yHCl7SjmHRmcmyEML6eGjwi6grfXCr5pr0JhzqalhZPqGj7scITJm0g3CNqTMSSF+xNz2uSEeUOXT5oMmFXCLqIj80pvtaRj/7jjwEeT1oKrj6R0sklToBeOYLdez/7b1nj2Iv2sn5UUz1iLigoCVdNR3x+0W2jm+48vueqVqy0NCstfc73azjgGcsZb/38Eod65TnPDy8Eynk+ePkE7/ymeYy80PZQgwYNvhGw60RDRqriuCuSSGNMmCP34yD37FtdPUPtrK3E3UGcXOY5wkOapgxHQ6IowlkoTYGotRUevAqz7EIppK9FzoKy9FWOEwKJWrtRMbcQSqMraa5UIISqAkaBswWlk0A36DaqQFIJQRrFjOyIXr9HFMcVxSq4CHnnKcvgDpXnOda6SrQbZvqD9sRSFDkYg5S6CnLZ2p46oq94/XWFQ4iKPiUsSspAB3NuzNcXErrtFqoW2VMnCbLSG+x0nBKAkpIkjZiYaNGKo0oD7qvKRk2RqL4CgvC6qgbtoEZt+5FSVpWquloVjlU9G6+lCuPynrW1ZYSwlGVJp9Om020xGPZZXlrAAqdPLWAcJJ0unSQly9eRSIwNx7koiiphDGOJogitFMVwCN6gE40UGmtLynKEVCEJ9njyMqMoR8RRitR6LNj3zrDe61GUOSqJx5UaCMmMUgLhZbW9Eo/FWc9gNGI46JENBjjhiOOEVqeLUiokBzJCjvp0JmaI004Q2XsBIjiwSaGCBkUKpKsCCBkqW94Fcbfw4KxBjfq0tKA12UHFGsWWsL9JMB4Fqrju2pmIbzvYIj0P2oYHw9QFvscJB1y8UfD629bIdODgf2l/m1Ek2dtStPT5GbsnnJb7jmXsOZk/usi8mtz4uz2K9xyKKOSTIcQO1J+rj6e86vNTOLHtdUA4z2WfXODLP3Qx9jzQys4nwqSO4Ll7E64elFz6d2dobZZ8eV+bE5MRpRTcsjel/xDUqEdad7t0vOJYj8O9Mkz5nSN5rm7xjLSkUAJ5ASZjDRo0+PrH7km2FQ+2puADoWKBxyLHaUZNsQml4iD8NtZSmhJvLWm7RayDrazLhpU2IQTTCI9EVZ+3IRCzDklwXwqJi6sC81o4roJ1rCDoFWQUuO8+2JAa79A6BO8h4XEYBGur6wgRtCJREo+TAGMtpgguWIiQ5AggkgKldOX8FETZQsjgrGQcgX0kiaIYHcU468jzLDgJVW5Q20XO9cy/r7QpQplq+wyKEDC7qnIQ64QkbSOkGjsQSaHCw0VQiexDJSjSGqkgjTWdTos0jbdE5XWVouL/CwmKQE0IlKGKPuWrmXMk3rot21u206oqmpoQKClQqv4XjIO15WXarTbD0RClNcsLS4y6HU4uLXPf/cexDubm52l1OoFOJzVlJeY2RRESFyVxwiCVRkUxziuKoiSO4yACt1AUWQj4dYRQMWWWk2UZUkqUFsSxDpUED4NBhjHgpEY5gXcWbyyiSujq89o7V1kY9xgM+jhriLWmO1EfgxgVp0RRiowUeEvhSmQUkh0x5ozXCXCVZAgZEjspAm3NGnyZ47IhrTRhstMhTZJx4hic0iqlkKy4iv7RCUu/UeHxTMWSa2fj85pk1MRBgM+eyRjZx59peAHXLY64ZiXj5ETEvoFBV9fZSub4zJmMI91A9epG4nEd/2Tk2PtALWB/dOsZKPjSpKSnnxy3Jw8kheSGL0+gzyFzEB66CyPEBejCVu8fIQUXHR9w0ekRJydjrIRDvZJ7p2Ny/diTI+08E7l9yEpXXSErlGA9UQ+xVIMGDRo88dh1oiEESAdeemxFC5J18O88wgV6SejxEMTKWgny3JKNMqQMvTaSKEYKifdbtqyCoCnARljvsNZTc32UlDjpAo9dKkQVrHlbf2cIIr2XKIJI1zkHFT1J+aoqQKgahATI41Xor+Cdx2bVOET1/rYZeghBi3GOsiwRJlhZOu+QVQLhfHC1qrUXHk8cR0gpMMaSlyHwrTUMD65g11WZQPACqRXCOYQL6xxlGYtLS2iliOMYX/UqUSqMTylFFGkEDqUkSZLQSmKmJrthpr/SZdS1iLHcuaLiSAk6CrPsNQXBO4F14AiBs5RUiaZDSDk+xlKC1iJUjqgocqWhP+hRFDlzc3Nsrm9wZvEMq6vrWOuY6HZCYD01TZq2yEajcX+PcG6IqpoUkSQtkCEhy0ajqqoW+l4Mhr3gZhUllftY0G5IKVEqQimNEArhYTQYUOQ5Oo4QhcXJkHjWdsDgsdZj85zhxiq2zBBK0EpbaK2D4YDU6CRCRS2EjhAIrDM4C1poApVPjPU4NXVnzJ4TwfbYFiXWZCQCJlttZGsmHEf94ICgJgOyRbvY7QX7DYw6GZ5JFAfa5zfIEsDIOG5ZKbivZ3YYMDxWSA8fvGySj1w6gRViTHERHtZzy5eWLF9eLkiV4FsPpFw7GwM7k55HQh14zh3PmTuRb23Mrj8v6FrPj540aA8fnVeYbdqOhzovH8ueqcd6MPf80FciXv2liWBjLqiupWrCRMDG4Tb+CahWnQ+UzvOR40Pum0mRrzqClWIHfamuOOwG2/evADYSxQcvm+Sq1ZypzOIftCJJqHTfdLjDn109FSraDRo0aPAUYPfUqXqW3W3RaMRYSBgCdGuDWFpVLjmDbIgpLa20VQWxwSXJi2BHmuc5o+EIZw2mNBRFTlmGakFNg5JSUJZlCOC9wFlDUQZaTVmYisIUAuSx4NlTuUiFsUgZ7Gulk6HBX5UohJ4HFiHcVsDgQ8DshcMrO6bGRFqOaTzWBSqYsBbhHKpq1ucIvTtarQSJDLalgBbgbAgwPRCJUJVwxgSqUKU4F96DlFVS4AE3TsjKShOQphFUn4nieBzktNKUWEmsF0xNTRLHkjiJtqIAsf3XbQ+7cbIhxmLFUDEKiaUta/tV0BqkklvuS46KmjZeGwCDwZC7776bufk9DAZDFhaWscYxOzNFlo/Y2Owx1+owMzOPtw5jSkpTIqWsEgWN844oCpUvpMDaEi0BHSGVosgzjAmN+7SK8R5KmyOVQKCRIkIKjcAzzAYMhsPwnggibyrXKGsNzlqssdhiiM0GeG/pdifQcYTWCSpKUFGEEBKhQzIRpEmVQUHlcOVdcFyraj8VXSskCrbIcMNNIuGZShOSielQnYJKo2N2HI8tml2tqRGhaVqj0dgVEiX4lv3Jo04Ctgfv3nt6ZTC7sA6O9UoKC8c2S1Zyd16SDKrvc1JQ7NA9sUML5YGhcXzqdMZG4Xjh3gS1S859fbbMnSi46nM95GOowggBzgv25Y43PVCSOvjctGQ5FpgqYm5ZmC231r0SCTK1tUW7GW396SsHnn9zT87Rz+5DD3XYF9WbdmoBVImXgvtetDfQpi6gS6IeStQ3rPQNJSD0Y9e01OvbMzIIHzQcUsAXD7R56ysOkZqdG+89HJnQvOxQiz9ezMkL95i/u0GDBg0eL3adaHhRd/neoswIqGboQxJS2qqJXVFgckOSxOg0QoSmARhjMKXFOUOeZ1jrQoO2UVnZmYZu3lptqyRYi1IxiKrSIQiUGu9RkRpTm6y1CGerCknd8CzcWp33yDowFIFyJREorakzmtBYWo6jceccpbWUAqQo0ToiTmKssyGhchZrSqRUoQIhBIpQTSlLQ6SjQCurGuqJUOIB5xiORmMbydAPRKBEoIB576CqwgB4GwJY6xxxFDHRbYfXgHa7TRzFOGtot9KgnbGOpJOMNRi1zW2Ncwlh6+B2+1tCgpaMNRey0oqM11N3VWQnP9h7z+KZBcAilWA4GvHiF30zG+vrfOAjH6Yoc7SOmZmdJ4o0ZZEFepEAIUOgrpSizPKx1atAILwgjmMwDmsco9FoXNVRMnRfV0KRpi1KY0OzQ2A4HDEYDGt35RCtVCYFpiwY9dbBGGIlSZMI0Qk2t2m7S9zpEMcJQqid9DF8JeoO+8VVDmLWWaS3IATOespiiMkytIDYGfZOT4Qq1Fhov82SWNTn6ZZ+B+93xE/bJnMb7ALJYxC/CsJx+OpayV0bJYtDS1FNMJhtFJ1w+T75ByMzjjvXS543n/BoajWTKyXX/MMmycju5L8+CoRKnSC1nn/+QMmPnYC7OpKsivOnjefi4ZYQ4N6W5GtdwRengnh8oBlXQcbr3PZ7Xcm4ZOT52WMF+/Jt5/+2somPMwI/8dzruRCQbhZc/FcnSC+ehInz4xI3mVtUlWhA2Oa19MGP8DDpN2cm6d5/gITjoIY8uvpXgwYNGpw/7D7RqO7ytUmPH1cNthrP5VmBdZZOq41KdeDHmzIEzj4E5t4ZhFChQZmUOBcefKF3RGXvSZglrt2HhAii89rBSVY9F4KVak2HIlBsXBitrGhPstJCKCFxlDhn8MbgpcK6qnHfuI+EqYKHUAWR6HGwLoAyL6rmcBY/ponZSi9SV06CC1KhChCy0mRUuogqqLQicHeNMdUYJKCwVVXD1Q9RIcaOWRJBp91ibmqa+ombpClREmb/AzVLoOvqxXZnk10ERA+1jJRbD6hHWk/trLJwZoH5fXMsr27wTc95DiqK+OhNH6MwFus8ZZYhtKIsi1BNsBZd/d2dmGbY3wgJrBShb0hFMhBKI53B2KLSZehg9SpCwzwBaB1hfagu5VlBvzegbse4jTmNkmCzAUJ6kqRFFCfESaD1lWVO2p1Ex9tmxMez1yHiUSIcc6EkxlYGw95j8pwsG+CLEil9SLaVBvRZIvsd+61+XYgdf4+TkiphNbYcVz++3vBgesjjQeHgMws5r7yotavzf/vZ8dW1kk+cGlFUM//be9M81RBCcPmkZrf68Dp4764ZWr1a6PDYt0NUSYoAWtbzrM2d4glflU6F91zTt1zTF3z3gmUlFtzdFtw6IfnCtGI5FgyV2HHM5wrP9ywYblw0TBp/Fh2ohl66CDt7nJWrIrKp+IKqZkCYlJu7t89NSnKm+/h7jdS74dh0ctZr4+rbtp3witsu5w2feTb7N7q0D1/C/+c1n2SlO4Rt95MGDRo0eLKwe42GDzaqimBzKivSvnOW0hRgPHEaEUet0BPBmvE9rSgKjM2r7tCVTWdFj7LeIlVIOEIXbl1Rd8RYv6FkmFov8iJ8VqogQRcC78ODTgiJ9ME+VApZNYqQY0tdhA/iZy8oIVC8qiSkbpYRgjkf3IfYquJ4a8EHao8UKkzkV3ZNzm+tZzxDTRCxe1mNxQdhd11NEFKCCo3uvAtuQ1puVYjCyhh3h460Jokj0kQHO1UpQiKmJEIFW9atAzX+33nB7oOr8KAry9DBfG19gysuu4zVjR4f/dhHObW4QF4YtJa0W52qK7nF+9p9KfRZT9tdBr1VlFbYIgjv5VggInHCo7QgEQlCRECdjIoguq6qCyYvGI4GKC1xtqoOeIdAoWWEd4ZYaVqtFmlniihtEeuo0vGEppG1I9nWzpA1jwlFRSdxDlPk5FmGLEbEcUqqIJrosqWK2ZlMjPfYtoSjdvDa7vIFjBOx+qdOqL7e4IFuYblmOedrexI2krPdeHaLIKT3DI3b9TyuADLruX214AtLOYUN1/GFsLN9zSMFrpmJeMG+FCXFrrdNF57Lv9gf2+eeD4TbzDmqo+Nftihf0sNc7pjP4ZvXHT92wnB/W3B3W/K5GYUDDmWe1y4Y9mfhmjlnkrGtqlG2FcefP0c2FV1wDfukhy/Pp/zd0IVq53la7znX48O+iqzk+ccO8x23X8ozTu5lZtAC4Jkn9vEvP/4C/uvLP8NmK7+QdlODBg2+QfAop1vCTLGHKpkoGY36SBmRJEloWkdwVnLWkmUZRVEiqDokV/apAErr0DjPOJTSuNIisaGvQBBXYG1lxypCD4w4ibG2pidtq2SMOcAV5z10ykNX1Zaa2x4SoyqQGwvDtx6I1lvwEifkWLAthQi9NGxwJ5JSBH6/d5Vw3QafLO+DVoNKiwIo9Hh760qQALyyoEQQsDuB80WgIvmQQHhvEUIB4XtSEZG2WiH41LKqwjz0I+OpmLSqg6HhcMj9J06x78AhpFR89OMfYnV9A6U10+02Ko7Zs2cPsQ7WsGVZopRilGVMTs2AL6pKV0g+lRJVUhISMik8SqpAZ3IC4wxCWHDBJUsIhS1z+pu9kMypqszlgnuZQpObEVpLdJLQmZ4lTacQKiSWviixLnDyg6taTbkK9K1AbQuVi2LUpyxy8JZWoknSdCzQr62dvR/vHGCLphbeq4Iq78fLAxhXJRXGVklHOPe0DgYD8jFQgp4OSI3nqsp5aeNJdspx3vPp0xl3b5SMTHXsCJMBT0V09mAdTktL9rcVLz6QkqjdJRn1GuYfyImzrSrpk4n69ry9707s4Mq+48q+55VLZryk8JUfntiqQZ5znVbR717MmWfGlR7vid+OR4N+6fhA37LWiYKu5AkaX3XH44oze3j9Z6/j+ccOEdmKdlyfth5e/LWjfOSZd/PZy46jHqpM1KBBgwZPEB6F65SognYP3jEcDkKTuagVqhCesR6hNBZrDK1WGxXpMQ1GVu5KrrKmdd5R2mD/6aQPDCshxjNXtb2nrOxTwxuK2tMdAnWrFuA6b3HeImzohWCFQVoR2q+5MDtu2aI5befIh98B4Xe4QgWBeHC4kqp6/ElQSKQTSAlKKqr2H+EB6QmBMRbvqpFKMe4sLiJFpCTClQitKb0bU2as90SVda0QoJEYHxrIlaaoKi6BSuadR1WT/edLmPp4YI3jS1/4MsO84OChGT72sQ+h0hatbkknUnS7Xby37NkzQ5omCKERyuNKS9xqEacpvfUVoNKFSF0lGQ5ri0oQrcMsqQDjiqBpqCsP3lCWBZubGwCVGFwQOnw7lBTkxTDoaXSMljEqShB6Wyd2uUWPErVVFyHYN2VGORph8wIwaK1pt4I18riqUlcohMDiq0BDjJONLZ0HQJVYVqJ/YyzW2eDeRnBZk1IiVRIqYzB2avt6gwCW25o/uW52/PeT/f0vPZjynLmYhZHlno2S5cyxntsd4pjzeYltHcbwi9rWmyJWkvlW6AY+k0iun4uZiCS1wdKjGcbe+zN07saaqqcaOyeHdtxsd7ddTqHW9gNrXEjag/pwTvcN/+zmFf7P0S6LnaDPOL8j9FWlSPDME/v4xb/5Nvb021TO5OMJre1QTo4ppE+OOXGDBg0aBDyqREMp8M5S5CVCSOI4qUThwcXJlBmmtGgdEamk4qa7QJkaB8MV5cRbyjLHmXIcICsEalvQ76u+GdvHIBH4bUGXrENJayuhtg0z0c7jhQ/WrI6grYBgGws465HC7wgO8aE/BR48tiaKVZal9RybDHa+wgN1kBgSHiGDfsQ4i3AeITVaK6SXeOGxFrwNHcWF32bNS02jqZsSVnQiAm1MS4WxJcvLKwz7AzqdFkg13k9bMoIHbQ9PBq98qxnhrbfexu1fvZOV9XUeuP8BrIjZt38v5ehuUJ6JdkorjohUoNZ5V6KkIDMFk9N7MXmJNSVUxzNJEowxeB9saKMowrnQ48Q5D96OY3iPJytyBoNhOC+kQlZUKiUEiAhvDVEc4Wyg0Xkpgn4IGXQy3iNEcPRSIlSOyrIkz4aYMkN6GzqUpyokSaJuprfzHK2PhfTnpo877zBVDw9jTLBHJqxLK42O41AJE3WvlXBu+AuwX8D5xPk6U2vKZap2H1KF8wRmU8VsqrhmJqZfOu5YKzkzNJwZWoYmmBXUyz/eMQKkWjIZSS6a0BydiOjocN1rKZiMtr4j3H8e7z56+gSYHpAbKXIz3TnsbfSpCxHCe6790EkO/dMaL7trk/c9Y5q/uXwSfx4oVNtrPAfXJvn+z1/Ly+64hFbxyMnMGz7zLG47tMBmK3+YpRo0aNDg/GP31CnvKPIMKRVp2gFcpZHwOFtgSoPSEVGk0UpWWoOqn0bdQXsb99w6R24sJviDhp4YIlCzrPDjQBxf05HCj6soS8H21VaVAEK1onRVVcIFEbVVWAG1zxQEhyApQk/z2lvcUVUDHCAd3ocgL7gLBR2Gq8Xc1lXcfMaBvq+yJ0/Vyboqbyipwkx0TR3wYSZej7UaCikcCoVzUJZB11L3J6lpV1EctAhxlDDY7NHbWCdutem0WohWjJJ6vI+e7FiiFlmvrq7yxS98meW1ZQbDAufg4MWX4krD/n2zCCwb/QHHl1cQ3nHtddcx1Z1mmOVESYIW0M+GCC+rZHGrS7kxJVEkCda0kjzPkTJUuhzh+BdZzmCQheBcGoSsm+QJtAgWyVorkjjBOYfSSdDKsFMrEZoqloxGm/i+w9tgcRxHGiGjOqupjrXb9vlAhQv/2vHr9b/BVS0kwsEEIDQOkSIkNd57VBRVrlR+HCxvaTYYB7kNHgFC0FKCFx1IxwE6cM7J7wfHsNvRjSTP35tgfUxhPXeul9xdVToKWycdov7vYeF3/o9ICq6YjnjefMJkLIl2QYd7rJe2r9f9dBIDC1BrbdR66wKqWewCHpKBAefZOyh541dW2UgUn7yo1mydO1cS2/7/MKvm0qUZXv/ZZ3Htyb3s6bfD6w93/lVfNttvo9wFUtJq0KDBNxR2nWgURYmUcaBJydDTwuPIixEIRRolSK3Gs/KiqnTga3vXsJ46cLJV7wDnq87Hgqo6URV4fWioVs3nIoQfc+addduCuG0Bn/A7aFW1oDpwVWuHqZpmBbLiyLONtkT1/UEIWgcppqqQBIF5rUdwFafYeYeq6RWBWVb97qvqR0igqrfDJLoDhERQ9c0QQbchhEBIUbuwgrOVwLqF0hHTMzNEcYQpc4aDjP6gT5xEJK2USIeu6Dt8aJ9whAO7trqBEJJ+b4ABjl56NXGs8dJxZmOdxcVVWp2UvXMzvPB538Tsnr0srmzifE6r3SEfDfHeVBQ3SxRFeO/HCYKvOr2XZY61JUql1TngKMuc/qAPUqOlpDQmUMtUcA0zNg8JsI7QWiFUChUtSRDOGVsWFHlGMRjgTdCNqChCxNEO2tNDFxWqqoPfqcEYJxbWYqwJs9VCkMQxqrJmrqlXYwvdOhkXfutaGFdN/MOw179xUe8R6T3XLGfcmBkuWhkFGqSH4WzM+pHOOCBzUlB09dgCGx46WJNC0NKCZ++JedaemF7hWBhZvrhUsFE4skp0DueudNTnQ6IEk7HiSFdz6WTEwY4KkzXnYwc8DO69vsPsqSJY2z4N8bRJMgiTTre/+jAHD6+S3tvnk92IW+fT8TZMZAmJUVy8NMPFy9N44B8vO8EDsxtbz52HQa4t+ze6zPVqqlRDhWrQoMGFjV0nGnGcoHUlkvaOohjhnUerGKnU+IFdz7yqqgu2q4Kk+vkbuObB4claE+hK1efG3NIqE3Au8OojraGaucZb8KE79xZ1xYbgvArSpRJ4EyxudU0/EYGeEuoaVXJRVRqkCHqHEN3XFiZbvPp6ykiMqSwAYtxZXFV6D1mNsebRO7dlJ+lF2B5ZzV6Hl6tGglLinEELidIarSW2CA5DSiryoiTOR1ibY5wjEoI4aaGjpPoOT5EVDLI1HJKJqWniJA4ag/EIeGJmM6tcTSIQSjEajThyxTW0220oCxYXT5FECTOzU6yubrKytIb1gmc/+7lYBzqOqx4oobJVlMG6VlRVCAhUM2vt2OGsMzEZGuyZ0GxvOBrglUJUTSSFFPiqxWSZF0Q6QSmN0gqlE6TSWO/AW0wxoBj1kD4cx3Ya4X3Vy+VhgsCzXaTqhMBjnMEZW7lqVUYBSpFGLaSq6YFUSUk4V3Z+T9DzOFFpPOqL40GJSIMAT6CsHOgbfuCra3z7/X3a5c4+C04JbCzHH7CJYvVoBxsrTjx/D2sXdSoq3dmXSf1nbXM9lSimEsUlkxGD0nHzcsGZoWElc6Eh6E4+I0qE6sXz5xOmkrOrF09UmFjPnI+6iv6sJjlpnxZVDV+XDJ9mwuV6f/f2p9y1/xCfvGfEXUtwdHmGK4uIF911EVednmOu3yayisQEmu4L7j3M/+/FN3PbocWH1VBIDydnNvmvL/8sP/bp5/Cc+w+gbegXdIEf0gYNGnwDY/edwWXNWc8ocpBKE23rdup99ZB1jtD3gkoEG95DhKZ0SkqsdUEIXhaACxx67wJdqQqktvp2iHFDt7pLtrF2nMAE8WzdaaEuA4RkRlRVBBBIoYKWQm6JppVWVdXF46VHVMuMZyerykUdCLrxttZ7JdBkBApR9W4QbG/nVlVZqgTEV63CvfVjsXeVPYVtqGhdgpAEWWeRWuGsxxlHkRV45zFlECILLxCE/hntiYRWp4MpSowp2Rz0cXjSVps4SdCRGk/7njfdRk0984611TVOnT5JqzvF9NQUCsvS6gJCak4tLrGxsRG2WcF99z1A2m7TTtrsP3Ix3oRalrW26lBeifmtIYr0WKOR5TmdiQnipEU26OG9xZiyatpYC6sDTU4Kh7eeJI7wBFvlKI6QKgqnhHVYCoQ1pHFUJcAhwPfWV+s5O6HY3rm7TiittRRlwdjaWASzgCiKgpi7omgFLUZVFRnTrdi2XreV2+JhTM/z450dktudQfTTFbtNlx7ubK3X8azFjP/7MwvMDYOL0YM58dKBzLb2W5RZDn6lAODALWucevYsd954kKKjd03V0VIwnShedjDFeNjIHXdvlnxhMd/R3O+qmYhvP9giUk989eJcMLHknud0mVos0eXT49yRw5j4znngnIy3CxfCozPNZX95Fc/84hw+VxzYmAiTMVXiVFch6jzq2cf383998MX8p1d/kjsOLoN/CLczIZB47t63wq9+70289I6L+YlPPXdMoTon6vvL+d/SBg0aNNgVdp1oGGMYZRlCSKI4Qgiw3qCQhP4FEutCP4hIqapKER5qQtazLiLw4ishrzEmaBoIQbWqmtg5HxyinLeY0pIXBWXd+M6GBnl5llGaEGh65xFCE0WyshetmvuJrV4cYc49aDMCbyrQvxBUVCVRdZrWKC3HBi0eMMLgDKFzeTVvFYobwZmoDm+9J8xWwzjBCQFw1XW7utuPqVTVS2NSjLdIGao3UgWxNzZoEYwNehbvPUVROTD50MfECU9UJUVCSNK0TStpkRc5pjAMhkPidopQmjgOFaiHw0MlIuOZxnq5aoOsdSwuLlNkGc951jNJ2zGnjh/n6ksv5sQDx7jkyGFOa8XphUVGZQFKcfrUKQ7u3Qe2JIpSjAnNFLVWOOcZDYeVjWtIwkajETqKSTuTKASjilYVxzGl9RS2GO9PpTxY0FEUBOReEMXBuUkIwFhwhk53MjR29PV54iuLTcbJwIPD4TrBKMuSoigCrc0Y8iKj0+3SSrsoJSvHqgf1xdhWjAiVLzPep1uptRsnE46qwOY91jtcRb+qm0A+HaEexD07slnQLh3XLWZctFmMg8rT3YgvHGhz/1RMpoO2CrZmjbefoZ3S8dNfXB4nGbtyL9p2juvccuTzy4Dntu85gtO7ox5ur3REIlCj6iZ/bHtvXyt0r9/+mScL9f5a2x9x4uoWF39lwDnLNhcYfGzZPLCBHmi6Vp910KO+oLUsGe698K6FqB9z+O8uJhpsdQT32+//2zdEhHvNwfUJ3vzRb+b/efHNfOniU8Gi/cHLEnaDcpJrTs3zHbdfxtQofcTxOOH52+vuop8U43U0aNCgwZOFXScazpbgSjaGQ+b2zOFtwZmF46yurJLlOc+4+tkkLc0wGzIztQehJMOhJ45j0mjrhkvVPVu4kDRIHxyhnPUU+WhMN5FVclK6QEEqijw4VvlQ4SjLsqpmVMmKN2iVVMUMhbVbYurAPAmzyeBDE7cqxPfO16oQXDWjbpxHEXQS9ayTNRZjbag4eB+qL0KQJnEldqdiW1X9NERIuLbcqgiajKpfQ1WioR5gqHAoINj1htn/0C28FUW00whnbVXV8cHylUDTCj+hh2BZGmIXVcmJA6lodSaIOkmYqfcOWxqMMZRliRciCJ1FZUFcuR3V++6s88A5jt93nCOXHEVWD0NrDJsba1x55RVccvQwx++9i2uvuJhWt0Ono1ldW0Np2BhsQqFotzukaYpSMXEUo6UjNzlRpAIdypWkscIhKY0jy4Z470niOFDhREh864pB7diFC00MpQcvoTSOVjvBC4mQGuk9tsxx1qJ0hEDiXejEvj25CtWl7YLuUG2w1lCWYb8F3UzoF+OdQ0lJEqdoHQF2XHUItMGtKki9zh0WtaJuFukCraeiU7kqqah1HsaUFEVJlmW7vWwvOLzzwyd2/H2gb2iX4Wrc3lDOC8EP37bGyYmIO+ZSvjabsNzSfHUuZTOR48QjLAypfRx9IkTo4bDn3j7RyJJPyEc1i55bz1fXCr60lLNZVJMr2475F5ZyrpyOaO22nfd5RrjdCO6/ts3c8ZzuurmgKVQCgdCWuw8u8cWNId/Sn+Xq4QSx27rPzn4t4shNKXe8YYC4wHKNwj6I3niO4sR21D4kly/s4f/64Iv5nZd/llsPLdBr5eOEo4Zygh/4x2t5/WefHahX1frPQl1tFvCJa77GR5/1ZTrGcPlSwaFeuZVs3PDYtrFBgwYNdotdJxpKRgipWFs+Dc7SmWyTZUPuf+AebGmRCLoTKc5k3DkqQCguvugaDh85uCNYc9ZSWof1gQLkvN/6qYMtQClFaUyY9XVhplkpqtndLe57CMpDo7Y60Me7yuLWIGVUCXir5EUy5lELYcf9CYJGI7wuncMLWdcpqtluGwKhavbZeVuxabeaEI412F6hhUIn8Q672dpNybjQYM56F5KZbZqN8E9IQmqhaJq2SJIEZz3GhHUURdXYznucdVjrUFpy7IFTHNg7R6uVUhqHNQbhIe6kCBVqOgAqEvQ2enzsbz/MRr+H0AqlFTOTUxw5cpCJmTkmpiZpt9u02glJFIdNw7G6tMjhi4+ACE31Bv1BOE5C8uUvfYGDB/fzpZtv4/4Txzm0bx/Lm+ssra3TmpymjWFqdo5UxxzYf5B2u0VZZGgtQGhsWTI9NYGSEUurqxRFRp6PiOO46sBuyIYj1DjB8IAFynBulIaiNAwGQ1qdLlJFFX2t3hdVt3gf6HuerYrDODhk61zcCvJNqKDZEumCgUHox1FfQr4qP2wltPVhdVWvDKjjj+2VjmoMBKpY/V3h+0OX9azIyUYjbGmCobIrdnvZXnC4YvVse80xzelBga92nqMbBUc3Cr7rHiiUoB8pPnG0w59dNc1iN8IBN9zfY+Y8CJ07yxmXfWKB27/78CNO+3rvWcsdy1kQhS+O7A56Xf3vuD/PBYCsq7j/mR2u+cwm0l4gg3oErEYlfzuzwC2dTV60McvFeTvQimqW7AUG7+EDmxsMr76H77756kCX8o9M/6rfmx6m/N9//W30WgW3Hl7gyxed5vaDixyf3QABN9x+Ga//7LNIjNpa53aubz0zJhyoNbKZz3CRuIPf/sgAD7SMJ7HukU7vBg0aNDhv2HWicfzEPeRuBBLuvec2BoPN0AvD5JRFwerqMpPTl+KMY3X9NNZ4Wq02K+v3sXfvIfbNH0SphDiyGFPgnKMoiiopgFjHKKkI9J8QeIRGfQopPN4bdGwpSygLiJOU0Sisw1TNKLSS4xuvROBdCEE9oV9FoiOkCTPwQoQGWeDxQiK1QpdB0C2EDLOcQuB8SBACC8zjRVi3whE6Udd2uIw5UMY5UIJIBIqV8LVlrUVohTShCZ9DVEEoWLFNmO6rdfowxizPaCUxhS3HM+SltcH21gdFiHceXwruv/84ZTHkyJGLqF2bpFJ0COOvS/Eexdz+eb77B76X1cVlVlbXGA2GzO2dY+/eeSAci3w4ZGNlBWcNUzMzdCa7uNJRFjlrmz2GwxGR0njnGZV9kiTGSsV6v8+1z3wGhw8f5WMf/zuitEV3ehJMwfzsDN3WBPN794UClAvVirIsaSUJszN72OyPKIxhOOyTJK2grZCafJSR5yFYrZMCCPxnR0jONjY38FLTjnTQBHmLKTJ04MlBVRXy3gUKW5XwBZaSq6o9lSuas1XyG5otyqoqpZUKDQWVDtRAV47HFP7dWbmoWIRBv+M91oVqRaiMhL40WZZhTIGQBucLIiUrl62cdjdiojuDd57BYPBor/MLB+eYRT9XAPbgxMMD2sFMZnjtnRu86MSA3/iWffQjycvv7dGqnZ8eIqp/xD4GVXK//9Y1Tj9rmrWj3XEAZ5ynqHRgSyPHRuFYzSxfXS/JbT0RIXhoyiEUDloP9/1PMOqqxtr+OLjeXWBVgIeDw3MiHvG1dp8jeftReLLvAtti83Ph0XT29t5zy0rBPXnGyW/5Jz511X18891HuPr0PBetTO34Ho/n1sOLbLbyHQmTF559GxM46RjGBU6EflA1B26tM+KmZxzjuuP76Ob15A+0jCMxQLSKiI8j0jtALzOhekysMaYuh++4MCtZDRo0+PrEru/ZJ47fT3+zR2emw+rqMoPBCKkFGyvrTE3P0EpTFpcW0FKQpCm9zR6D3gYIz+LCEvdP3M1Ed5Yrr7wWpUJQFXjuBmvKkHxUM7y2KuuXRYEzFo9HKkdkBVJ5hHPkeYYQMXEsmEw0RQFF6XA+uF1ppSitwRoTLF8BV5pxxURWfRq2ggOPDY00qHUBUoQeIF5U3cWtDdULKYI9KkHgzjjICM5Jcvy3qGhRchx4She+uw6IbJU4RMhA/wlPBOpZ75rEK6gqIlIGwb33YF0VIIlQKcKx3u+zfMsJ5uf3opTEGEtRluMZe2DMYffe0+50aV3S4dAlR8/iA49n3h0UowGbGwOO3XE39x67m6k9U+w/eJDZ+TlMXjLZnaC3MKCXjTiztMLRiy/iGVdfRV4YpmYmadsWUgh6oyF33X4Hhy+6lMOHjpJnwyD4BoxxXH7pxQyzIQ7LaDBAakWSxHgUxoTu3MJ7nJdjnQNVVSfPM7ySxFGMk8Hq11mLLfOwv8aaD7dVUXAOU1H38BbvJcbm5FkOUiKVRGsFqFDhkkHHE6UpWicheHAekw/BF/iqG3xwHtqyu7XOVuJ1Q1FmjEbreF9S2pw4amOMod/vE6XBZU1YS5x0gJi8zEkSzdzcFK24zcrK+qO5xr9uME4+vGd+aPkPnzyD9J52GZKMpY5mpaV3VCMEcLhXkJrQN2eLnrXFZ/GE14WHBy7uctt0wt0nR5WgW7BROJZGFgEUzu8QetdW3g+HQen43ELGK448jGi3wUNDwKTRPK83jT5PTlT1rRUB0ji6ixl77u0zsTBic3+L1Uu79OfTLb3OLphmiyPLP5zJKK2nbOXcdmiR2w4tkpSaVqnPqiL00wIrzxbnx0bhBZRqZzYoBHzhkpN84ZKTTGQxe4aeI5sl33JywDXLWaBEiRLkll6tPs+b1KJBgwZPFXadaBw4sI8st2TDDK1i5ucmOXnifpJ0gmE+4szCGSamJnFlgW4lWOvoDzMuOXqUhcWTnD5ziuVojc7EJBOdSbyA3BgKEyhJQkq0CK4/oVmfDRUOFWZ/rBMIDJHw6NjTFRFKCaSMEcJh2xGLiyPKsgzVkIqi5L2j7s5RI4oihJI4X9nNhur2uHqhCNUEL0XlFMTYPUhUY3SVSDu8bxG2uqFXmgypKlep6i4f6F8hWXHehQC5GqeESksRqCDDkakCGI9AUpSWpdU1rA2BcT1LboxBKoWxDmMNwodZ8ttuv4NveeGL6LQ7Ff2nTmBg+yNnTPF4yMeQqPMY0u4EabfLxGSbk/fezb6DB+lOTQBhf1hvGAwzhtmI2dk9XH3l1SRxymC4hjOOe++5lyIv6bRTJicmmJroUpZ56MKtFFmWMzExSRKnrG9uUuYFSitipZBKkhcFo9GQJI5riT3eGfI8oyxKNjY2GGU5e/ZMo6IY4QVaCcoio5bEWO+QhOPufGgEmGeglSaOYqIoBQS9vgEVjAW0jlFRgtQRQkik1pWOJRxn72r6VTjGoVJSaS68w1Q2vKUxWGMpbR/r+lhvKE2BVgJPifMZaeKIowQZpcRJwvrGGtaAVBFpq40pLCOXMRh9g3f3rRLlbuWgJIB7ZhL+w7ftZ6m985YmCILz1HiuXxhxzXLGNcsZ3dIFYXqlzxhFkk9c1OV/Hpng5H3DnRoadlZYHo1rm6gsvkfGY2sr7N1+9tEIRb6e4D0mlgyvS4g2BHbR8dz+NPNlss1o4zGsdtvvqnQk/ZLZY3323b7O/J2BTia8xwuB04LFq6ZYvGaKlcsnGE3FD3ncpPVMHOvzN4Unr6VC2wZZRJYiOncJ6Vz33kK78bsAncKiK/rVdGa5dinjm08OuHijYO/AoOqqqWRrgqtBgwYNLhDsOtH48j/dzN79h+m0Yk6fOkGSdJneM0crgUE/Y6I7y+TMJHfddRvlesl0ZwpjHKfPHGd5eYnClGhZcMtX/hHnBfsPHETrhDgKdrd187rQodtibAgCpNrGd7Ylw6FBSUeaRnS7CaYokUoTpwndRHPsxAZFEexJg1VtsE71dX26umF75/BOIKq+GTWFyTmLNLKawao7nAMCkihGRhql1Nj9p15XrZ+QMvRMUK0UWzVok7627XXBitURXLGwzM+3GPRyEKAjy2gksC50jA4NCt24K3lo+uaqZCdUeqKq+gFb/TuGg2Hg+tdOXX5bb5DH8AyqEybvQeiYfl5W4w/rjJIEpSKGwxFTM1NccfkVCKFYXV3l1ttvx1vHt77gBZw4cZLjZ05z/NQZku4kBw4epJ22grBbx+zft4+iMKyurjLKcpI0oTCharG+vk4Ux1DZDztvKIqcOE6qruoWFcVIHREJT6QiWnFEpBWm0qmEwx+qH4UpyLViamqKVpKO9TDBYlcRtzokrQ5xlCDGPTW2NDrUOp+qAuWFx1tLYYOOw9qgC3G+IMuCeNzYEYgBZVmihCLS4dw33pJEKaPCko1K3Mjh7YDhaIBWMa1uSpk77l04SVGWWNcEEg/GUEuW2hord4ZuHjg2nQDw1bmUyHouW8/5xc8scqRyuZIe/uyqaf74utlxl+XdJBO13qametbfd64k5YGe4YtLOVd3NNfcucHM8cHZQbOH4Z6Elcsm6M+nGC0eUiuyfYyPOh95GhD085Zk9cYu0+2S9D2C6/vTjz3JqOhRwnvaqwXdhRGXfnKB7mJGXDuVQbjRVcdSlY4Dt65x4NY1BnsSli+fxCaSM9dO73Alm//aJtMPDOjc1+d/fMdBmIwfs314fVimcssr7u1xoB/omM9aHDGRB+pmZD3dwu78zDZHtgYNGjS40LDrRKMsSzbXVjFFgpAOFQnyfMigNyKOWkhpWTxzP5EWaJkwGGxSliOKPA3aCdlGSTBlQWk8D9x/jG63TaRThNQoqRGV/sE4H6xehcZ5R6RD5+ThsKioQ+BLT39jGGYJtWRj05LImDTWOG9xViCUROCqbuDVzVhuNUdzPrjdxFIhlQoWop6q14bCliHMFzI8gKRWKB3GiRNorVFKBscnEfQVUaxIkyTQdioalTOu4tkGZ6dWK2LvRMRoOERrR6QV1gnWVnuURVQtB51WqHD0+2XlrOWw3lL3nKgbgwVbXomrLHydr8d8dtDz+CDQkUbrBFB1XSHQj0rD7OwUF118GUopVldXOHH8OJddepSjL/lm7rnnBB+56RPBAllrBJ40CR1zjTFMT88R6ZiFxTNs9nqhkiAVHsvG2ibey2BHXG27MSYkHoTXoyTF4tFS00oUkVKh6R6MaWM1bc7Y4FjVbrWCqByH9duaMUpFGreIklbYhyIkF0KIMVcaP7YKCILtbEAx2qDKk0EYJiYSRlkPT07aiugPRiHpEYI0TcFrNlZWkMrR6aSUJiTHvd4mZVmEniKJpL/ZZ22lR1FYhLB0u93zeEyf/vDAZes5/+LmFf786ikWOtGO91ul40C/5JqVjBedGHB0o2R+WO6It514eK7+2V+6Jfx+/nzCpZOBGlM6z90bBlO9r4xncb1gZBxfOJ1x38gwfdsai9azZ2S5cjXf0vICrq0ZfK3H8Yu7/MPR7kPmBDOJ5KIJzZGOph1tnbuPtAllItiYj5g9XTyGDOXJgy490cdHrH7IcjhPiXd7cM6luRCQbpYc/cwiF/3jMtHIIirzj4fkQ217vbOS01lZAuDiTy/uXMwxrojtFuc6psp5OqXjm04P+aHb1zi6UZ7lwjYe0baxXaCHr0GDBg3G2HWi4aupv+FogLOePCsCJ57Q42A42CBJEjb7fZwLwTNSkniF95buZDdoFACyEc4HTn4Sxwjhq/4RHocFYXB2iDeeVquF0h5jodPuUnv0jPIMW3hMYehOxlgnGJIzO9sizQy9TQOCIOr1hroRoIDK6Slsj/eeqZkJnAyz2UtLS1gDaToZFqwKAUoqlIqhahLoASVD067SAzikkkipAiXLg/Ghv0jdULB2iSoLS7/viGJPnpcYk9PptIj3dXBGYp3AFIapWUlpY/K8COJw6sZ+VTNC7zFV9+xA0XA4ayp9iUQrSWHMtqfR9pDmsSBQ3KI4reyBtxAlCZdffjkWwUZvlVPHT3L1FVcwPTPNTZ/6Bz73pS9Q2JKytLgsxxiDVpI8z1BKo7WmMAUnTz4QqGciHJvN9dUgwK8oZt4bjCmI4xhjTKAsCYHSCi0CFSFWuuqnsv389ZSmRClFHMfjCpfzQaMhqJoteqjtw+oEAy/Gz3bvHMYUWFPiywF5Pgw2tOUGnhwpY7SKaLdi2u2ISE9S5EvESqEQDHNLEscMhznr66t4ZymLHuurnumZKYTXlHmB9zDMSzb6Q7SOcDY4trXbXbL8G5w69SB4IWiVju+7c50XnRjwjwfbmKr7tvDw7MURe4eGzkPMBDsBr/3aBv9wuMM9M8murg4PxErw0oMtrpyOxt2+PXB0IqK1lnP4iytMPzAkPjHAOc9/e84cH750gv/ygr0AxNYxUezk6OdK0I+rc3dgeCicGsBtqyWzqeT58wlXTkdo+cgjzzqK05e1mD1VXNBRajKy7PvSkH/aiEJez8MM90GTKar0dFZy8J50o2T+a5vMf22TznJlC11VLnaNbcueQ1LxiBjfdavnT8t49vfL8ZzFsxZHPGM557rFEVO5RdcaoCahaNCgwdcBdp1oCF+SDYdILUijtLKa1ZSuoMwLSkJTudIG3UUQOXsypZnfu5dDhy5i0N9kOOyz2V8nidv0NvuMhjlxqrFGBCtSW5AXmyRxCByTKALhkcIhpavE44rSeKwp0Vpi8ThhsMYDCd1uiyQZ4o1jfcMRdzXDkaXIQ+KgtEJIT9pusWd2jo1Rn7XFU4yGQ0bDnMmpCdJYkbswC66kRkURstJa1JBSV73/ZNVcDlRldVs6h/OCtO1CfwqqZoLC4V3BcOAYDDKkCEF1URikBB1BGsWUpWR5uUen3eWiw1MYa/AiVHZU1f0raFA81N/vPA6BtaGFoFQCX9Yz8Y83ydgGRaiYjE+OQBFSqk0njShHG1x2ySUkrTZ/9ud/xl33HWez1wuN5mSoBAklGI5ywJAkEdbknDq5EHQwLpw/6+vrFKVBRxEQ9A6RdKRxtNXksNrnaRTT660xkoJOOxnTmyAkkM65oMWoExQXdBogqmNqx+uTPrhMBVqcwDuLMwVlnuH8COGHaJXgXYa1PSIVeqcUA4hTycRkF5MVLC0ukyZtprozpBMtlNIsL28w6mcYaxkOcuJYU5YlCM+ZpRW0VUgNe/fOc+b0EmVWIjsxpjS0WjEqkmj9yE26vpFQ6ye8h32Dku++a+OsZTziYQO3Tum4binjvukY9wgOVXWV8EhXc8VUSDI8QVS851if6RNDLvrcEq31nTbE3//Vde7ck3B8MrgFFUqy0jp7JvxcM9cPhdXM8rGTI2ZSyYH2w9/O67UNJxVZV5EOLBdkCFsxxkxgC9FXhsUoZ3+xdd57AZOnRxz88ioTZ3L23NsfVzJU6Zg4Mwq3PO+3rHyfQLcl5TzPXMpY7Gjyyv0wMQ7t4FCv4HCv5GCv5PqFES3juHhjq6KkHMjtFueNK1SDBg2+jrD7RENphNRYb7EelDPBnlZIuhOzzE5NkxV98oUzpDpF4jE2CCE3VjcYbN5CaXJGZY63Aq0yvA86BGcEzniUsMF2VnhMGTphC/okSYSznm47RauSPC+JI43sSqamWhR56BC9vjFkNOiRlZ5ESVTs6HYgiSBJNadO9ylLS+lzWklCO5liaeEMK2vL9HsDimxE3EpxzjG7Z4Z+b0Q+ykmSNlJLbC1elxIlBFrJIF6XouJp+yrhkHgsShdI52kphSEmjiyg8C6BtqcogtZACIs3Ficlw35GHDm0VigvEdZiiyEOTZ5nmNIgiILWoOrrYUoznsH31mC8JS9M0KCwPZB5/A8wt02MXpPQhFA4I+i02rTbbb5w8xeYmZnjE5/6JEqnTE/PYgXEcUyW5ww3e+TDjLvvupPLLr0E4Tzra2s4LEIqvMspTIGOIrzIg7uUd7TTmDSJq87YPuxnb/GupLexSq/fJ2kHE9Gxda0NjRjjJNnRRM240KwRv9X0sO76bUyJNQ7yDOldaNIoBN720bGjFSUUhWWUDciyAT5t0W61SUtHmZXkKmPhzBIb/T6J0nQ7HdyiYzgcUBaGorRkecZgY4COJEkShUaOXrA+GBK3YpLegCzPmZ6cImm1EK0Ohc8x+YjyPDnvfL1hbL5wrvce4bPKeX7sK6s4AR+9eIJRJB/2My0tee58QqIE8WZJeyXj0k8tMv+1TaQZK4K3vtl7Ltko+A+fPMP/vGaa2+ZTznQjrNiyG33Y5OYhXo+k4IrpiMndUnc8rBxK2Ngbkd5bXUcXYGDbt5J/GsREXvCC3gz7ygcl11Jw4CtrzC8eG3OlfCXsPyee4G2Mreetn1/ipQ/0uXc6RgBXr+Ts75e0S0e7PLsUEsb75IyvQYMGDZ4q7DrRMIWhlchqhj8ERTpq4b1kz5555vbMcv8DA7zQeCwexVR7krTVwZYjbCGYnJwjsTmrK2uoJPTHsEWGFQJjS/CKOFK00hZFYcmKnHxjSHeijdawtrGJ8AIhJWmikdKRjwpiHVGOMmanO/R6GUkM3Xab5dUeUSxwBobDkm4rpZ+PyIoChGZheZVOmiAQaK0ohUZIgXUwGAy46PBBer1VVjYG5JnHOoe1Hi3COJVUxFEcaDcyzIzXegnrcpAFCkukY6RWmCLHCYu3BikkUeTZHIzQUpHqFqUpKQuJdZ6pJEIrh1YJeZGD8tx7721cfeWVzM8dAAJNyDpbea0LpA8Gvc4a8iILVZAqeN7C43ugqcp5azjKmOi2ABEqDVFCHGs+9KG/xQALJ+9genqOvfv2s7Y5RCpJVhqKYY4ANjfWaaUJcAmlKUhbCUVp8ASaSStNycoSrEFKTbuV0Gm1KppEsBF2zjAaDVlZWaY/HCJ0jESHRTwYU6K1JkmTisJmgXAOY4PoPi+LIMguCooiYzQqiBPN9OQscRIjhcB5S5Zv4CnxrkWWGXq9ZUajnLWNTVpJgZv0rPX6DDb6GHMSY4PLVSEEvf4QawymNHgdUZrgSpZ2WuhIYvKCfFSQG0eRO1LroVxFRRqhJKqi47lcIEREmk48rmPY4EGojAAmCsubv7DMnpHhPdfOUqqzKyB1XPiC3PLCr66z77Z1Js5kpJvFw8+cizBNf2Sz4Of/cYlMC+6ZSciV4EOXTvKFA21yLYitP7sRnQiJ0DOWM9K6XwiC2Dq+yXpG330YHz26bub3PbPDzOmCONvuSHcBoEoUNq9MmPlnXZ7/zjbPOj7FQ5kp1UPPJyLueek+po8PdzBE44Fh9r5+SP6eqGC+Wq9ynuedHvK808Nt46tol7vsH9OgQYMGX294FBWN0D9CKIHSwQlI6gSsYDQccnLYZ31jvZrdtnhfMuj3sLZEKUmkYzqdKeb2z3Hy+H0Uecbm5kbVFK0faPEiJk11cHcSnqLIAoe+KDClIFIRQlhsZS1rnSDPS1xbs74xolNCEkdBMI6i20notNPAp1dDuhOTDAYpC8sjpqdncWXJ9HSXTkdz7733IaVEeUFZ5DhTcmZlgbXVVbSMwUORDYjiVijrmxDNlmURxNcimK5GkSRJJV46iqwk957NfoYQBZ2uoMwNnSRGSomWEiXC+Gq3IikFE92EYT8nUoqN9c1A6hGaI4f3EMetSmQCou69QXCk8hKcC7058iwL3cKVQgm5s4/G44AQgiLL6K2vs29+Bk9IrvYd2s/SwilW+5uURcmRQ0eY3bMHqSSRliRKMhFHdOdm0ZVWRUlNVuTMdeeqxoQFzpZMTnarvhah9UWaKNppGpKmqpO2tZbBsI+xjk7aZlX1UHGC1gpbNTaMopg4iijKMnzOg6rco6wpGfUHjIYDBv0hRZEFS9woIomnUZGqYoOqt4spGA56pGnExtoiSsOwnzHojXClZ2N9xEZvhUgFOt3s9ByLi0sMBgO8CgmaqbQX3Yk2rVTTabXwDtazQKMQzpImCoGl3W0RRW2KfECnM40toShGdKe6zO3d97iPY4OdqK8M4T3f/9UNciX5h8MdSik409XbYlfB9QsjfvofFpjJ7M4VPML1VVdcBNAuHdctjoCgIVlNNXfMJVy+VhCdo2u38J7ZzAZL3vEXerwU3H4g5di37kXuwveh1gVszkWcvrzF0VsHXFiZBiwfTrjzhRPIrmRvO0F7gZMPP8KvvfwgD3zT3FkLSeu54mOnufSTC4jKvnZHpek84lz31wtnrzZo0KDBU4PdJxpSkKSK0lp6/RFSxbSFxDrDxsZqFdR3mNuzj9XVZYo8pygzimxIkrSZnZ9nbm4Px++/h7XlFaQMDcza7Q7eh+7Iaazw1rGRDRjlYYY50RGjLPRbMALa7WA3urTSR3hBt9Nmc3NEnLQoMkssFKUSDDY3mZ1pU+QFvWGBMeDTnFhL9s93WVpeYHpmL7LqvC1VhNSe3ORo7+n3hqRxwkx3mmHuyPOc0bDElI6JiZlqn4CzLvD9HUSJAJkxygxaaLyNEKqkNAO8Uygd0W4rHJbSWhKhmZ1KyApNPirJigJrPcWaIaFF6XI2B0Nm9sxQGsNEZ4KJTptIKowv0TKpEq7Qm0PJMHtmjcU5E+x5pRxTu84HjLGUeYbStdjao5RkstPh1sU1Rv0h+w8eYHZ+Dq00Es/c7BR5tsHq6gaLK8toJHPze5ienqEVx0RRhCkNWXWuKCWxNrguqarPhgecD3Sxoizo9ftYC1IpVJrS7nRQUVK5gDniOEHK0EyRqhFkpZ/H5CUbG8s465iankIKQafdGVsdUyVmtX1pWRbkuaGV7sEUJXkxYrLVZjjsEccRcazo9dcrWlf4niiW5GWOtQYlNKbIGRahqhYJmOp06G8O6Wc5WsXolmT/gb1EKmZxaYmJiRkkEUkUsX//IZy15HlOnMSkcee8HMsGD0JV2Yit48e/ssqP3LpGpgV3zyQg4ExHk2nJK+7tbVmMPobrqg72BZBpyacu6nDTRRMg4IMEcfA1KxnXLmXE1o8/4IXASMlaGnG6m4IQdA9rlq6aOrsK8gjfbyPB4kUJR7463KrEXAjwUCYSp0IFaP1yw8FPe1QhHhS1e8Soi5Mp9714hoVnTIVL90Grc0pw1w37Wbx6kukHBlz8D0skvRJVuics4WjQoEGDBlvYvesUgmyUhfk4B1ESMT8zC8qwvLxCaRyzM1MkSjDRTlk3BcXIkuiIrCw4c+YEedZno7dGkQf3Dx1HSDGi1U7xeXBoElJQlDmJ8kzPTDAqCorCkmcFUSshLx2FyXDWEemYPM9wToBSDPMS5y2tNhgMJ89sYD20Oy2yYkivL8nKktI6+oMRSm1w5vQZpqYn2TM7h3Or5FlJHEsQnqk9cwx7Q7SGMk3pttuAI45TIq2Di5YUFaUKhAwOUkVeUmLpdB1SKvo9S5JGZEWJEppWK0LLgjQWKB3hnSUzJcIp8jxHFtCa8uSZY3rPDGVpkUoyGm1y5z23E0cxs5OT7Js/io4jjBVY56mMvsA7lJRISbBI3YUbzW4hgGc9+zo6Y4vV0NPjxOkFlleWiJMWe/ftQ0mNc5bjJ05w8vQpysKxZ3qGvXvnuPaqK7n+mdfxwKkztDodvLOMhiOUjBAEvURZlgyHA5TWRFEUAn88WZ4xzDLSVos8K1FRBM7QbbdI05R2GpKN0DXcBgF/rcswBpMXrCwvMxwOmJmdJU1jijwkvYHqLbE+D0mNC95akU6YntqPEJLe5hrdiS6gmJqep9dfZTgI1CgnBYP+iLTV4dSpJYZZRqQ1Ok1QLqawfVSk8FKyutEnzwuiKCKKU6SW7JmdQ4qY6T37mZqaJR9lqEgSJ4Ey1m51cFD1MGnwRKAWlksgcp44dzz3zPBBS/iddqOPGsF979h0zO89b55b5lOsFGgrmO91mFP70GqCCYZc1u9hO7eSJT1Od1KW2wm5kriKirU42SKfjR/LEFjbH3PqihaHvzq8cOJtAfuPZTzwjDZrB2KOvXLEpX/VpnNGnbWo2jjI/c/P+dqNLZyU5yzMCMBpydpFXdYu6nLq+lnaKzmXfWLhQXqaC2UHnE+c4z7R3DoaNGjwJGPXiYa1JdZqpJAoBdOTXWbmpzFFnzxTWOOR0pDlGwyyYWhK5sNst9KKvMgoyxLngx7De0HpLEUmGQ5zhPQoFUFZUJoMR2imV5QO6xTGGYxRWFOitCJSGqTECU9hShZOrRErTR5pomiaSMVEkxF5buj3R3gUqxtF6MjtJcXIMJQjhHZs9Pt4B+12jC0UAkkcp5g8q0TCVBQwQRy3xwlR3adBek0cK6Ssm8+VGONJY43SgpnZBCU0G+trTMxNMSp6KASrgxzhS0aDIXiBsRKBRimB8wolHHNTXQajHqX13H/8bpZXTrNvzz4uuuRSLAn79++HirZlKpG0dUEHUtOlZOWEdT6qGsNhralxle2rx3vB6uJpRsMhF19+OXHcJhv2WV5ZII5bdFoJJ1bXWd3cwBQFCwsLGCm47JIrMKWh1x+AEESRIivrJGNEnMQgJUppnCkpshxrDPv2H6Qsc7zvIyMNVjEzNYnWGqlCEzzv647dYX+MRiPWV1fAOeIkwtgEcDgrca6sqhguJDpeBytkv9VzhUpI3mp3UFFEluU4SvLckA1z8sJSmhJhFaNRhlIRsU6QSiCdp/SWJI1xHqI4xRiLkA7nI9AdJub2Mj05TRzHgaLobZVgQd1usc4XnWuihScDddJBZbqwflGH0XTMgVvWtlFwdiy9C4TrxeeX0TnzzXzfp9u8rnonMoorF+ZIS42smjJ6BEvJIe48+tFgwLxdOO49jB5sNL37bXNasL43Yv89El0+gRqGR4ntZ7fTnvXLS9pnJKLuplgvJ2DhGVM4VcLDDH/7y3k3Iu9GfOkNbfbeucmhm1eZOjEgrftWXCD74DGhOidDvyENXkJ5CFxCnYUdn91gaXIwroC98ikcboMGDb4xsHvqlPNEWoIQTE50GQzXuPeuHh7LcJghhGQ0GrJnzzT9YZ88C3zzsgwajYmJDtZZsnyENRDriMwWjPIMHWmklJRmI3wHFpxByRYCgfclRW6g3cHbgqIwpGmKFB7nZEgCBEjh8N4w2OwTtRLKYQ7OkaQR/cEQJQVpKvFecune/bQUdLpdhoXgxIkz5KMSIRTea1ZW15nbu4/uZJvlxTMUecbRIwfZ6A8BjRDx2H0pUJTCfV4pgXGCWEoiHSG1JYlblGVOd7JFWYzICostctI4Ji9GJEnMYHOE1BqJJM8sUx1PZyol6cYYlSILw0R3mlbaZW7+MLZwDAebGLMXrYJYGG+xjvBABpwXVdKxs3HdY4aHu+78Gh/8mw/yg/+vH4X94eElhEDGCUcuuRQpNJvrS2TDEZccOcqdd3+NMwuLZHmOkhG6JegPhtxy621cevHl5PkIqUIyUeQ2dNA2OXN751jb2MRnBRYYDgaU1jC3dx9R0sYPLHEc4yx4bKgwIXA2dP72PlgSOFOwvrzCyuoSXghmpqarxMtWAXtJTa8K+o3gZGWdD45f1lIaQ5ZlDAY9suGAoigw3iGlJm3PoWMDmxuoIkegEJWI2AlFnKiQQEuNFDHYIQcPHyIzOojfncCUOUlrApQKFSk8XoQTKjgT+3GiE2KJp3Ew9HRCdX2XqeLMdTMsXz7B5t49WDY4+E+rVVf4AlS/iuPGbUEfedWj65kdXcELVh962TLKWJq9i7Xp46GC8TBrfkwqCw+nLm9x0e1DJpcfQ4OIJxoeTMtz92uHHPyHBM41xEfpXFsv5nTo8n3m2mkmFkY89z3H6C6OqmD9Cbq+Hk/zVFFZND8Mcq1wAuLesxG9F4GPwE6A33rMd7MMsVKMnxENGjRo8ERj9xUN58kKg7cWWc3+ZjbH2NCtWgpNUQzobQxRsSBNEnSkKU1GlgvWN9aJpAxibeVptWMmpGa9JxmNcnQaqDbWGJI0wtqY/sgwMdHGuJzJyRQhHEJqTFGQFwXeOYo8x9hgnao1aCHZd2AvvUGfqck2xlg8JWWhkVIh8USRphMnCJezubGJ1i2ecfk+hNIM+jknTi/R2yg4fu89XHf99YBmfZgzM7KsrfaIdEqrZVFKo1SE1ACeQCv2JIkBn5AXAygNQjjStqLMVegabUMFYpQXWGsoBjnWQ4JAAT6WaA1pK+bU6QWyzKDjFnvm9+B8ySDrs2/vVSitOHH6BAfm9yKURnof9iGeorR4PKWx6Og8BRECrrnuGv7+po9R9NaBi4BAz5I+VHiWzpwhSiKe/azriLTmc1/4PMZ5htmILMvwPojHrXX0egO8K5mcnA62w8ZiTU53YpI07dDrD1HSkmd98jKn1ZkgiiKED93eJWBdiVRUvQ/EOPkzxjDIhti8YHN9nThJCTJwhxBB3BvaaMQgRghUJWz34ASLiycZDUYUeRHiGynRWqN1RKubUjqHsQbnHWm7jfUR0WiDwkKUtkBFdOOEOEnQUYxUoSGlt4ZCBmOFKA7WxDYfIl2oWwTT3mDT6aoRQxhsJRlp8GTAe5yWLF05yd037Gdzf5cDnznM9e+9gtZKC29CrxrkAPQCQg4Rnc/iozPbqG0PNcXuEVN/BWoV339JmHmuv1Z4et0FTu27hX5niWFrddxc8qwEU4BcKVD3jDBXtB/9uSHg4F0j2hv2kZd9suDhzKUp/dmtR5OX4FToN7EdTu/YdY8K213Eevta3PK6i3jW/7yP9mp+fisb1f3ICYGXgkGkw/X/KMc6iBTrSfSwC23Gmun1y7j81HcQla2tD2/b2KlhytSw6cPToEGDJw+PoqIh8EYAElsE+lItfJVeY40F51FSoqwkyzKyIghdh6MeRoApLWiIZBDKDl2gtoTExeKMI221KEob1hMpdJRgrcEWhjIzZNYQR6HJWZGXlKUD4YkTXfXekAxGAzZ6Q9Y3N4hVhIoskfLMz05ROJidmGJyboqN9TWEl5SjAknMZjZidbPPwlKPLCvIi2UKF8TeWgpWV1dY29yg319AaY0UAiVU1fHbMTXZYnamBcqE3gnGkZUZrbbCGEUk28TdiGJ1gSRNWV3vVYlJjIw0SdxirdikHYdu4hubA7JRQZ6V5FnJyuICl19xFZddegX7Dx4h0pr11TWsLdFSY7xAifAwKkzVgM57rLHnzXUKIYniNnmeUz/BrA0VqMHmGkiYnZ1mc7PH33/y70FFtFot9GBIWym0lqhYI6KIL3/5i1xzzbV478nzLNjcpq3gyKUjlJBI50i0xKUpURyHjtnDPv3NzapRYpjL9YCvjlU2HLC4uEhmSw4eOEx3YsQoz7EudFZHVEGi8EjtaMn2WNhfFAW9wQqFc0RRStxqYZxDx5VTWJyiZIQtC6QN9KYoipGtguFKMApQcRDq1gFnSBBEaO4oKwcjIZEIIiXIEZRlAT4di97D57b966nSjqc5veNCR7Xvh7MJt7/mMEtXTeKU4PAnj3D1H1+HzvW2/MGDmQYzHc6/7GpE+2bofBb0enVWPvhYVemCzKH793jXITdX4YTHScOp/bewOHcnRhXbln+ohEUgMofYKB/bpgLddUNUOPx51HE9HghgOKUpEomsnHdXryq577tGXP6XbXwl1RAO7n3NkI1LzeNKvmth/uolXb74gv08++OnmCrLx1/Z8EHH048jVloxq2lMoQSFkphaT/Iox3ju08AjnWbP2sUcWdjP/MqVVZJxjvFfGIe4QYMG32DYdaLR1pqJqudEpIJFZ380oigMubUY7/FakyjwQlIUOYH9YUBI2krRShVeKUZ5Tl6WCEmwh3UWX0R4AaMsJ44VWZ7jK75pGinSThtXOoTRlFlOno3wTtCNUzKT4yuL2H379pFlOaPhCK08BRadSTptzXo/B6no9RZwJ04yNzlDlCa02ikWSTlSLJ1ZxwtBMSrwQnL6zAJKSjqdFKcUprIuKrO8MiiSWBeaDHY6EhVFFEXJxsomaSpptyOcB5PnSO3pZYIoSRgOMrSEbGgQXU23k6K0QmqLFdDPCxwGgR43LDTOkiiFkilKhmRkz/w81DP51uNC/ExR2kC3qTpgj6v2j9fJ0nt0pOlOdsYrU0pzxdVXceqTp1jd2GQwylhZXmT2wD72zMxw+x05gzxjcrKNNZaolRAnaahKeU+eFyihaKVtsmxEURToqKQswrngXAsnIqSO6Pc2WVtbI9I6OF/ZUKEQ3lIYQ7/X59TJ46yu95jZO1s156toUhXXXkhBEgfr22yYMxqMsM6E5FZFRHELrSU66QSxv/fIKEapGCoNSBInaLuVzIXEgtBBXqlxc0CoeoxXNChP1aRLgMfhpEPoGFPmsCOxqJKn2i3Leax3WGdCz5kGTxhMorjtu4+weM1U9YonXW0xONine3ICWVZ6gQfbHLkWvv8icB3EzJ894vcIUbCx72PcMfsFyuDigJMG/ygoWF/XqDZfFYKoL3bE5l7A3i/FtF+j6B86P8lGrhRnOikT6wb5WFa4jRqVacWJiRanOyml2lnBeCxmV+da3OPRNuaqu1/OnrWLETvKO6KpfjZo0OCCwK4TjazIoQ+R1hRS4AjORloQHIaiCBlJbBma0XXbHUpjQHmsABXFSCXIS4O3HqlssGG1ISC23iGFwjjwxpEXhlQrsrwkHxW00pgoEkE0nijauk1ZGLxwRCIIe6WEoijIRqH/Rm4k83PTrK9vUhhFMciYmmhRYhjmBWuDTfIVQ9pOSZKUzd6AdjdlZXNQceTDLLc1oQ9DkRXMzs6SV9SkdirpdhOMd+yZa9OdSFE6paWi0JNDgDMWlXQpxIA4ipDOsrq5SWkN1ki8FOSlZbC4zvTcNGgY9AukL5CRotX2lE7RG2Q4ucJdx+4liiOmpjpMTM7grUPqYOOqlELKMO+dW4MWEqtqceD5Qf1Q9mL7qeNYXlqkP8xx1jIY9Lj4kstIWi0ipRBeoqSkPdHFlSXSSVIZ9BDZKKPT6bJ37x5OnzkTmhaWhjwzaC2DdS+gdMSg32fQ30QphVAKW9kKewyb/QHLi4sggruUx6O1xosglg3VBYXSocN9PhxR5iXdiQnanRZCVNogH4EwpJ0pkul58qJEWIOQAueC/sM5h62mGGVVubAI3DhRCN3GZc2rH4vKRUg05FZMIoQkShK8GQYNhnU4b7EmNF60zmJM1VTQeEylUWrwBKBODLVk81CrfhGAe157B/d91920ltvMfWUfUT9i5s45Omc6xL1kax3REqL7qV18WbiSJosMSYpVehxMfuNqcM6OjO3AM3mzZO+nYnhQHD35gCbeFHDocX5rlfjPjgouWR8gH4uWwnucEKylMRuJZqGTMtL1gP0Tdkw9njzukSc91qaOU+rsQUuM6yEIIPMphd+iYB2s6K8NGjRo8ERh14lG6T3Sg7EGYaGf56RRgjVlFXRJpIBIS6QQSBlm/JStS/8WTYSUMUmi8TisVRhR4lB4Z0N3besYlRZvLZkF5XPiSJHlhqwIjdaSqjGbEKEjt8PjnUArzamTy7Q6KdYajBcsL66QG0uW5UxPT7K60WeUFYBDSRUoTrbEe0uv38M7j8AgZaAESSnwHsrcUIwMSTpkbu8e2m1NnpdILO1ugjeOVtJl0BvihWeq26UwJsxe5znDYU6uytDXwkeYrMAjGGUlLSTGlKGrdObITXgoaO+w5YjO5BTCOkw+ZHXlFF/88oC5Pfu4emIqdCqPdLBmtRBVTlimKFFSnnfxsCf00ojTreDKexiORqysrtKdmGF6ZpY4jqpGjYo9M9OkcYRzlo1BwcbGBhbB1ZdfzuREl4MHDuJcgXWGPMsoS0unE1zFjLWhQtQfgjMkSRJ0DErjipy8LOj3epxZOE2sJK1Ol54aECuBUhKEwv//2fuvYMuyO70T+y2z7XHXpc/yVQCqgEKj0OgGG2Q3yOb0kBRJBRkyI85oXmQiRuZZetaLXvWiCIVCilAwxGlRnJgZ9ojTtO0bbQAUgEYVUL4qfV5/7LbL6WGdezPLZFWWAdgg7gdE5b3n7rP32vuY/Xff90mFTlOUUHR9R9/02ABlliGVigRyLDLkBNEiRUrV1PRJtR5ZEAgfk4fgI9FciehQL4SIMrhCRFlUIQhhzRdZq16dJhWItadJTEKVECgRMJSsjiuMsVTVCuNtdCpvIhepaRu6vscYg/Mujiye4SeGG1/foS/jV+P9nx2bW5ZXFyyvLgBQvSKb5oyvb5xu8/S3XmSyu7/+7ec1YfgkiB+Sw0byF//as/x2jQC6m45rx55RKfnGYus0KTjBcGqZho/RqA2ccl6AUyWrtLJc3V2Ren9P3u2hlx6TjL1BxhubY4r6PPkqIyfgVM9qcMTJCoNwD+TcBAIiiHVn4v1rCMKvCyf3hvJ6Gfjjq28SLt3ES/OucwPoQsY7/eNRxRE4dpss/fB077/x8c70DGc4wxk+Nh460dBCoNfjIRKBE4HeEc3UrMMHR2s9yspYt1WSQCTsaqUQCDrtkErgbKwQE6LalMQDMhqweUg8BKFi9bgDoXJsiKRqKUUkmfctxgaKMqeZLRHe08k4JmOtoWs7AoKKgBcSKRWz6ZI0S6KMauvRWrFarUh1ihAG0xucD1zY2eTundm6Oh6VqPI8wfuA0pLVssHaBOMMaWLZSifoQmN6z3y2YjgpWa3iOI7WiiwtKAYFudaxWu09g2yDg8Njgrc4o8mUxjpQaULqPcJDVmTUTUvTthRFgnU9XVfRdpZvfffbBJXwzFNP0ps+SruuRy6MMZiuwQVHCCfyth8w6vGBuGdUF/8fb2neB6SKXR5jLHXdxMq+cyyrht4Yds7vUA43SZIEiQI8x4d77O3vcvfOLsY5iqJkVCaMN8/z5S9+gZ1z50nThLax1HWHtRaldAzgnaeua4y1KBFQeY4SGuc91lvquqYzBq0Vg7LAWYtzgNJ0PmC9R4rY2ZjPKvquphwMGA836Ox6Bj5oQuhBpHjRIUjwtFiXoINnbVkfp64Cp+NQp2NOa4+O+D6OBoOszQxj4BAvu5QCJQRaEMeqnCcYQ9fV1MslzeKImYC7d+7QukiKd1aczlkLIU6WgpDv9xQ4w2eH+ZUSl8j3meDdHxgGwKWe+kJFfaFCeNh+Z4nyxx9rNkYQSPy9TtjHCnE/jYrR+lgmlXgpfvLSrh+1VgGrTvB776TsLj3Qv+tvLw/mfK4Zsm3ueYZIH3jyux13vpZgk4dQ+opEJ8r9Aee/d5E737hFP45cs/Ovzrnw49knzg2vTUrulJs8fv3rXDx4DunjrdVLS5stOXl156M7zMe3ONh+cz0iJ07XJoJg+/hJrux9mcSUeGExSbO+BILrV77DfHx7/Vvg0G7zp/XXuW0v82GKVO/taAvxsd9pZzjDGc7wifHwqlMhVm51UNhgCUCiLc46tI5BODKS9GIwtK7grIMzGxzWOEIXg68QQGuLlArjeryPnXHhwItIXNYyjkp1ywqHRGeCoKBppqRaMh5PaPuOTGlEJrHeoEPkLhjrCR7QCiUFWZYiiCNa82mFVJrd/SMkirLMmMgBBEc8CcXOuSF7+1Ok0tj1OJhUgqaukMqwd1CDDCSJYrHsyXXG+fNbDEYDTG9ZdR3BO6SUDHMHytFqRZIlLOc1RREr/hKFt57hqGS6WOCUIHiBQFJXFYiEZtXiQ0Bphe49zz33BXQC16+/yiOXr5CmWaxx6Rjoeu+xtouEbakj+fojXt8YQMegeVW1WOd587WXGY8HFFnJnb1Dvvj8F+k7g/eWt976EW+99RLz6Zy6d+zeuYMTjsl4QCI1Xb3ieHpMnpd8/qmn+fVf+xV++MPX+dGbr7F3sKJ3Irpfa4l1hun0GJ2sPTNCwBhD0zYkWpNmGaLrEUohg8S2Le1qhXUOnWZIZ8FHgz0QpEoxKkpSIVlN5yzncxSQFyVSJ6cBe0ykDAINGAQJAQc+JWpphlOjwPuDzvjOvifnEsfrFVookjQhTVNiuu3w1mC7jnZZsaxXVHVNXTd0XYcPgBRI77kwzvDeYmyPJ6CSEqnX3RIJWiVILZGJQuv7RnXO8JnjsT87oBtq5lfK08eCuudhIV30WBnutySNQ4TA1e8dc/Gl6X2O0w8HFeBLhwtujQru3DfP/7C78JsJ7vHik83jB7jxpZKL11pGRz8h3s/JOJoUtENFM3x/klwuLO1I8f29hOm2RLQO7An3CVIv+eXlFlv2PapLXpK+dBHsFJIPD55PhheTKuULv/klzn//IoM7I179z15i4/acz/+rO5/qNLfannm2oiqPTpMGAOUSBvXW6XaDeptBvcXB9lt80Is23bhBU8wY1jv0umE53LvvdM36eycgCNy1F7lhHkF8wH5OhSjOZGzPcIYz/HvGw3M0TI9SCi/XwZcA0xpSremtJUsTnImEV4Ij+ID3oJTEW48QHh/Ah0igFgR8CFjXYa3H2YBzlpPgTsrIO5BKIUWUrXUoRKJRKIrhiOGopPQFR4czyqJEasV0NsURkBKEkvTOUiY5w+GQxWJBW/cMyjGz+TFeSQQS52IXRCpFCNB2LcvpApUovAEkOOfxLqBkyvbmiI2NIbu7+3RtT5ElXHj0Ubp2FSvbIkQZ3s7ihOW4W1HmCaOxoqkqjo9XpKkgz1MGo5LlsmVe1SAE9bIlzwqCtAyGOS4IVguzDnjj2hazJV//q3+Vm9fe5ODgDlceeZS+96TEkTTjIp/EOY/3hix5GOfgwNHxIS//6Me89PJLOBRttU+ejanqGmMbvvP9P8d7y3R+zB/94XVkEnAWIkMBlNYcHN9FeIHyKY9cfYTHH7nEqCx59c23+N4Pf0Dddiil0Tq+rqbvWc6m2D6gpCLg6U2PFIEiL+itIQgZPTO8p6kqqmaJVhKXJAip8M6dJhlKBYbDSFS3bU06KBhPhjRNu1ZJExBkTGrxSCRedhByoEWKBMTavAsZfSyI4xRKCLxQ8T3qPNZ09G2NaTtc3yGDpT7e56ht6ZqGqlrR9QZr47gVIZyq+wQBUipECJRZyvZkyHFvyMqcIhugswKl1u7zSqBUFsezzroZPzmszfl23liwcbOi3rz3uTl6ekyzkSJ84Pwrc5LGUkx7dBfH2MQ90s3HOx6QWceTs4oLVcv1cck0TzHyPeZ870UIBCVwnx8QtpJPXJ92SvATiUVPEgwl2H804/rzA/pM4CQU047NG9XaXT2w0p6uDFzZsFx6HMyh5/pRzt4PoTWSx7qS56vxuw371oUC2UuSSmIL91AdoWf+62c59/2LEASXv/UYm3duUdoXSdpP5xC+0RqetUte2/wB1x+VTOaPMF5dJDEFItyjlgtABEXqM4K2pwwKEyKvzElDVR5Rl0e4oGhDDgQWbsx+f+70/Dqf8mb/JE+nbzGU1XtSDcFILkmE5dhtcj9H48DtMHcTPllmeoYznOEMHx8Pz9EwFps4uq5HCEWiFVIIut4QhAPnSIscHzymAy1BJxrvA0FJeucRPlZ6Tr7irDX0vYnE4hBASRQSIaJMrhASJwJaC1zvSLwilYLJeJPBZMzu/h12toZsbmZ4YLlsUFLTdg2ZSuiswVuLLCV1VzMYFOztH9H1Fu/BAT7YOC8vS5ABUxtWdUPvAqPJkLbpqTpLUB5C9IO4desOjz72SDy/AJs7FwkOrJPYvqNrW0bDIfkoZ7WoCNZjU8Wq7uj7lmKgqOse6wRpFvXVg3MIJSnLEqU1CYFyWJCkOTJInLd4H4nOb197nfMXL3Bue5M///af8ILvuXrpKTrTI4SLpHhjIjnZB5y3pxW9ByEEwc1ru1x76xp39+8gpSABlvO72AB7+4fs7R9xfmsHYz3LVUOWSYSEZdWjRCBNEubT67jg+NVv/HU+9/STvPrqq3z7u9/l5u1dmt7gnMeYhqqqWVYLgg/RY2I9KhWcpywK0jTFOYewgeAdwVqcMTjTo1WC1glJKggSamfJUoUXKdYaTBe7XDrJEEoBJqo3IdbvvwD3jQ8IkcbkgjQqfYUMsFHy1jpsbzB9i7cdzhi0kqRJNAjsVys6YxhPNpnt73Lr9k2k0ggUDg9Bxve1DMggoimjSpBao3RKmqXkiWbezQgqZbR1AVWM1lyPdYIi5L1xNsKnHpk5w4dgHWjq1jG+25w+fP/PH7T9pxo7EnF0adhbnj1aYqXgoMy4MS5p9AckliEQCoX98gj7C8N1wP6XDyaTvPTNCccXE57+3V223lkyPOiQLqD6D/fueD5VzMaao7bk6f2ryAdkQ8M7ms//swE/+N8vPtjQD+LHRsL2y+c4/+LFdcISUEYxuqMRE7d+/T7hVVwnqLl1PH8456D8FjcfH9Mx5PFbf4UL+1+A1BCefoegPAMCLyTPwaBeLy8wtxUORx8st9oDrpsr/Lj9AnfsJQISGxT2PuM9LSxfLX7AL+bfv8+35aNhg8aF+99TX/lk53yGM5zhDA+Jh0408iwhSSSJzgg+GssRPKgYYNeto+4qvI83EKUl3ren5n4oifJxpKrUKV3waK0p0yS6WXuH0AJrHSAQTiK0QFiPlpqAxQWHUhnG9PT1jEGRrhWK2sjnEII00+TlkDzPOD6Yslp6qlWNMYYu0eSFpqkNCPDB4Z3AF5JV3ZJn4tQ0sG8dterp6gopBD5EsrrA0/eCd96+iUgEo3HJpZ1NlqsZ0FEkkuBzQnCkKuX8+THzeUPb9KyWPUIKtrdy2tYAgSIXZC5QV440FWxuDEmzyGvZOrdN3/cU6YTptGJ2uMR4hxSGb/3Zn/I/+Nt/l/Pnr3D7+g02BucpRwPwAhkCTWcR62vvnD0djXrQzVQI+NIvPEdVrZgvFly/9jZN3xIEdJ1ntWhZzCtmswUXdy5TVR17Bwu2NkfUraWtF5TlACEE2+cucHdvj7fefIPlsmO2bOj6DmsNKEmWFag8AxRpoulMJDo3XUs5HJzyHrz3VFXFcrlkPJmQJBqX5wRnIidFJlhjsdbSdYa6mTMYjRiMxrRNgyMQZaliECGEILCe7UOCCEjh8cLhXaDvKzrT40xPaxxpXaOkItEpw9GYwfYWOtUURYkE6tawrFbMZjPSLIsqW1WB0in6ROZWJqACPhhwHq0UOh2AgCIfILUG4Umqjl4kNE1Dv2jRSkUzSm8phpsopSBEMQStHqZD9fOGNadICvqB/sC0WgRIa4vw9/31QXyKn7ZXyclYFpA6z5VlgxOCtzYG9xKJ02RT4J4uMV8b856c+WMjCDi6kjE6Xl+Xz/i8hQevBbp3bNys150f8ZHHUb1jRzl2ck+mHqCyJuKorerukaPvu0Sn0L0jWyiu/sFj5MdFvMhBRLPF9OZncZr3Xr8QuFC3bM4eo6n/DkUziYwKqxG3rsCaTZHyyLuevpHWHG9cox7fQuqUy/oOg6LGIbluHqXxBbv2PB7JFX2Xp9J3uKgfwhzyPdDCosWZat0ZznCGnx4eOtHAB3AyunOLSA4OAZzxkYchAlpFwqrWmjRJSJIERMA5j5CSuqopyhzvPcrHm6exUX7U49FB4/1amSMIWFf5u97g+4BKFSFYhDQxUHeBkCU0raVpesy6sg0ghCTPcwbDAfP5kqqxjMUAlQqyLCY3qRa4ILC9Q2WKqjKU5YB84BEo+q6h6wNCuTjXL2XkbEjwqLWSkmDVdtg+sHNui2rRI9wMdEI+zNHC0/U948GQpu3ZPzimbQoevbxDmmYkmaaua4JZIYVAS8lsuWBrMmZVrbC9A+dJlWC4WdJ1BttDY2re+PGP+If/43/EaJDHMRwl8cGBFHR9v74OJ1X7j3h9RUxuvv4rv8TOhfO8+uMf84d/8i3u3r3BalHRm0hsd1YRNh0heBaLirZpAc1kWFLkG+RpTOCm0z2QgrQUDNqUqi4Y72wiEBRFzvkL2xRZStsZ2rYFIcjzfM0vicnDbDbj7bffZjAYMtnYxPtAkiYIux4R6zsW8xnz6SEET56XaKnXRf9wj7x9ejMOa2dyGA5KsqxgejxjuphhHDhn8SEgpOfK5ce4/PiTEKLEblEUJEmCC4G27+m7nrbvCQEGozHeOXSiuHT1KiofxaRGCPpqSdtU9H2L8wEVFEIn6CxHJpIgBMEFjhYVCEFVz0myjN5FPlCWJ5hKramxASmjI/kZ3oMAzUbK29+8wN3nNz/QgE64wNY7K7bfXjK+25BWNjpBnyQef1mMENcV8iurhirR7A0y/HrMKEiB/eUJ9ktD1tzmT8XrDVJw+5mCR35co/3HkW/6aCSd5wt/tkDZEW/8zUskleXyD6c8zEFiUSCAMpjLr5O+9YunvId3bwjybkv6O8dQKtxzQ/xIky96hvstF348Z+PmisHdEnn3G/dlI4CaI/JX38W3+tQQEPwAtfxlhu32+rGA8BIWwwc+LWPEpcMLnFMv0OZzFsM99s69wmq4zzl1CAJMSAgIUnEfUf5jr/svyXv8DGc4w88NHt4ZXMs1ORVssEgEUoBKNFKA1AqpBPiAVgqEIMs1zrk4FaUleIHpPVpGBaPgHc4GkkQjRfRFkFKAULgQyde96dFBAZI0ScmLhKpukEqRqATfuThpLyWm8+RZhhBgbVSIcl5gkXRdRyM6Lk62oRRM5wuKsmS5qrDW0zcVg2GGMwbbxeTpsccu8Oa1u7S9J1eagERpxWaRY63HO0/XON555yaDVDMcFrR9S56X5IOMNEmxXUeicoK1FFnC5UvbbI3HlGWOt5553ZHkOemgZZAWjDc2OHpnTq0t1jYIoKpjEjUYDTl3bgfTO5Is5eaNt/mDP/hd/t7f//tIKbHWoaRCCInpmziWJKPU8Ee5gou1IpXWmmc+9wSPP3YVKQR/8Me/w63uLl1f4Zxdj8xBoiR5ktB7T5ZA5wzPXDlHV61w2ChJO18RgkdKjVSCsijou5bRxpjBcEjXd9RNs+5OSJquxTlH0zS4tet2mqZr34moABWkwHpLUy9p2wZno2Gf1AnOWnzwaHGPRxTFdCRCCvIsJS8GaKVomo7ZcoGUjj5ISBKSfECSpiilKMdjNrd2EEJydHzMsm6ASE7vmhbnHVVd44Mny4s4LiUEOi3wvkboAd5ZmmoeldD6BikgiAyPIMknCCQET9d1rJqaXCvSJMdZSyDD9C2JAheW626eQieKLH34+sB/2AinQWO1nfHK37vK3nMbHzx+vo5t7355k93n4zZpbRkcdjz67UN2Xl+Q1Bb5rm7Hv9+gTPnA54+XbLY9b28NqZ8ocU8UuGdKwtqj4f5TfZjVftClqceKa18e8NT3VzHpCnz6c4/ulBQrxxf/eM5rXx/x+t+6QlZZtt9arj+YH77qE4lo5IMT6yBgfDNh/KKlHlToN2rOX5A8+RdHFEfdKXcmiBVh8G3C7O/FwsNPZPpwPaDqRtA/enIS8T8PeTm1SxhW5xhW5zh/9AxNPmNv5zXuXnwJLcx9uzlLGM5whjP8bODhVac8mDXpNkgQ3tMj0VqSFwWpALGea5dCxY5Gpuitpw4mSo1KGBQJo9GItu9ZzBdIFQgiELxHqEjsjURyh3CQJwkCQWsj58B6j7WOQVbQ9z3D0RjbdgQvkFJHUnEi0EpQVy1pllEUOcuqoWosu3szrlzdRAmF9wIpFKNRSr3qWVUt21slQnZUTcdbb9zFeLe+8UpSpdFE3wXn7Vo5KI5SzYXGS8FwWLI5SamrmqpqIDjKfIi3BhccaMW8blj2HRtFyXKxwsvAcDJEBMHu/gGJlDRVi9AD2rajrnsI0Pc1eT4gKxMO9/aompY//853eOzRx3jhha9wfHzIj370Eh6wXctJdKXUgwjE9zwego9dJ4g3+CTR/I2/+U12djb5t//u3/DaG2/iOkuSaoTQlIMBm13H8XyJ7XvauuEvfvASXdty+fJFdJKysXmBjck20+kRrrf4Vce4VEzyEcoGlqsFRTley9JaTN/hvWc8Hkfp2q4nhMjdMCYa+Z0Y5yVaY7Wm61o8seEGkoBc/xwQSKRIyAclRTGI3RLXI5SgtwakRGc5l8/voLMcLROkFPR9j3AtvTFRhhlwxtIbg0sTsixlvligkoRUCLSOajg6SQm+RiSSYFa4viFNU+rjXQgGrVPSNCMfTBAy0LcLUm3xXY0G+q4hG20Q+haFJ9ESpUDoAMZHrkjt6PTP+ejUfRyVfqC58fVzXP+Vc7Sj5F6V/724/0ERCdDdMKEbJsweGZA0lq1rKzavV2y9vWS43yJt+GQk70+D97hL11pxUGb0QiBmFrnXI3d7wobGXclOA+awkUB6j0AOvHtEbA05NWDvPR6GGj9QvPMLA1abmid/sGJ0HFUFRfiEJPfTBQABlA08+ydLZhcSXv87j7Dz6pRH/+yQtLIfuv97vLIPSUYE3DpY8rvfvcHkUc3t/RXbdcdTuea5cUYgsKN22LRfJ3SP864MI6TgJgR9yL3hq0+KqE5H+wX8/G9D+ITKcGJ90YJA24zEFNE/494Gn2KNZzjDGc7w08fDdzSCj1/GJ6P+CLQIaAGTVJApS+8DSiWAIJGeTAomkwH5xZw3b+yxImCcZ1G19F1LdElW9M6hg0CKWIH3HoT04ONIViAq/VjTsVw0KKlYVDV953BhRXDgrKPMC7q2xQJSSLJMxeDSGJRM0FpinefoqCYvM9q+j21tEbDOEAQ4bymLHCkSDg9n4D1JokllwsZ4hPOO2awiYCnyNEqRSklvPIvZEmd6EjmmKFKMBSUF25slR8eGxbKmX3myLLpTN8uGwSBnNBxyNF8yLAqKJKOarRAKgu0piwSCQynNqjbcubGLcZaAxBiP1DW/9du/xXQx54033+bV134U1bz6nuDuJQ/GOrSOuu3yJKFY37NMZzg6mHH+4hZSSbx1NFUHSpMPhuSjTb743Je5/ubrTJdzJAFnOqxt0UlMJoUXtHVNXmQ88/QXSBJFXuQ0VYMWkVR9uFjQHxhefesOk8mQy3//Cta6+JpZw2Q8YTAa0HWW1XJ1aoSnlKaqKrTWaKkRRPnjJM2QnUWmCaLvAIcgoHRCUZRonUIQTGdTrLWRQyIlUmX01hKEICsyysFo7XS+DhZQkYjtPc6EaJbnPb2z+N6j0wzrPYPBAK1O+ACeLNdYC5kWdKbCC4UJHZNhTt8GBoOE6fSQznQkxYBhCnmWk44TUlGyWoJOHRcuniOYwLgZslo2KJ3QNTPwDqk1Ovt56mjcqz57LbGZxBSa48eH2Fxy85d2WJ3PY4D9MSgG92/mlaAbJtz90ia7X9xA2sDgoGXrnRVb11ZsvRMr/SdStg/e6SclStzbpxeCWZ6wSBPuDHM6Je/J6s4MchZlaMPak+V0Fxsafy7FX76XfKhrDaJ5N0NavC/RUNivjnHPDth9Iuf4Ukq5iJ+N0bFh607Pzu2OtDnRLf+Y53YyuhkCm7s9n/9u4K0XtrjzC1sMDlsuvDJn89qKtLLv23cwkKw0yd2nP3hsan3pXvcH/PGtG6jb8tTn5veFiFQMAue15R9sWL45PEcZH4yw24TqlxCT3+aTB/DrdChIQv01wuqvgd3gwRnvh+5qDYHRLbvnf8zti39Bl63Wj57hDGc4w88eHjpiid10EQN/4XFIUgnjTJJmAoVGOR8lOWVAiSh5Oj2eceXCFr/85Sf4kxdfZdX0OC+xzuJNIAQHUmHx6ADeWbyLvI0gFcYJsH30xjCWjdGQrjWYvkcgEd7hvMVi4w1URBO5JC/wAbreQJCM8pyN8Zit7Q2kTkm0Z+/wkKrq6VrI8pIQJKkucd6SpZpz21tkqULphMVyhhaKy1cvc367iyTi4MjKAqk8wnuOplOqVY8zjtobrO0Zb47Zn1X0naEsBihVY2zANDVqUBC8xgXJcDJAGseFC5u0rme1qqNKU+cwxrExjCpMzsWxICE8SaJoup7DgyN+67f+OT4IfPCE4OhtTLCkUlSrmt/+7/5bhFScO7fN1772dbIiP9VYV6lmY3sSK70hBjHFMI7waBfIrCGIhNZYrJM4JMvVCiFT/v7f/w02RiNef/U1Nrd2kAJGk02ODna5efMmXWfwSJK84HIqQQ+4tD3hq1/9CnXrWKwWCA/nL55jMp5gvaOqFvR97BhtbG5EfgxrN/AQonu7EmRZgahjR0PIKIFbliVKaabViun+AXlWkucJvfHQ9aAlqc8ROonStUKdmuF5T/SDEWCcpapX+CAwxqKkQklJlmb0XUeSpEgpccFHt3chcbZH6YDQKVJ6BrmgXXXM5xVpniP1gM2dlCAFpq9ZzixHB4fkZclwnDEMWfRm0ZK0yPA+dmWC8OxsP8L06Ji+70kfwpzsPwisg8Z2knLza9vMHhkwv1rilcAU7/7q+jRcBfGe31wiWFwuWVwuuf4r50hri7SBnTcXqN5z/pU5aW3vJRYhMDjoUNavydkP2QUIgSAETkrqRHGcpywyzXGe4u/vTJz+8J7Hwr39iGODPDbwWvUQJ3xv3WJhSf5kBlpgPzegzyV9Hjtm83MJtz5fcu5Gx/N/OCdt3Puu1kNjzTuZHBi+8jsz3trKebFJedmdJxltIcv3JHAiytyeO8j4Rl9yQdxTLLx/BbfDnH9qv7/2rQnvGhH18RTZNYf83w//MW+2b/O/3P5PodDsTlZcOZ6QdU8izAVI9vk0CPVXCIvfAJ/dO9+PfNK9swmxfcRyuM/+9uvMx3eoysMHuoif4QxnOMPPCh460ZACkGEt2JHghSFNNHleABohO7wJSCFIlIueAyF2Jo6Op4QAv/q1L/HH3/kxq67DGUdw0bn5XqMkPt/LKHfrQwBOJD4dIQSO53OkFWunbDDWEnBR+jU4TtrOiGi8J7zn4rkJv/zCV3nqqad49fWX0SpHiITx6A5vX3uLvrd4JCHEoDMEh04kEPkeRZai1AbGxATo/PYmSZpw+84Bk/EG+4d3+dznniAvcxQJaQJ7hzMIjrZq0alGCEfbtAwGBV3bkek8mtMJMKajbyzzxYJqo49jVtbhcAyHYwiCputJU8V8WXOi1kSA4ALOe5w8bTVFl/a+i/v2huA88+UcrUsG48D3v/cDvvDcc2xsTgAwvaWuaiBHSUlVN2itGI4GfOmFL/PUF57GEfjW7/8Rx9MFb77xBk8++U1+/a//dba2Nwne8yvf+BVee+V1Xnv1dfb2dtE64dErj2G6itfffIfD4xnT+Qys505ZcvXxJznaO0Rkmscee5IsyyPPpHe0bUfTNUghyYsSRMCaHikVzrs1yVsQ3TAkmdIUw5KqalhVFUorVnWLFQKhJTpJoTf44FFCIYUiKwY472NXbe3gJwRIKfHC0/WGpmlI8xKVJDjnkTpF6AS/vrZVvQRvEDJBoGnqjq5bkQ632CgLmuUS5zqKQcZ4c8LyeEY5LJB4Egm1aWNHJPSYXjKZjAjes3f7gL5zCCEYDof03mBMh5SCQTmgb9rP6OP/lxT3VfhnVwf84B89TrWdEZm2nEaaP5HQ6z1F+yBjtwPg5td2ALjxV87da7QI0L1ncNghrOfc2oNjcqv+qMPQCcHNPGeZJtSJit+BH+ad8cCdfQZXYp1gv1cpNQAHj2bsPpHx2I/qT+i5EXdqg6Cz8OPbmpe/DZXpP+J5ME07bpy7zfPVmG8stimcetfr7/E0wZwe5r3JxsnPIQR+Z/XH3Da7/BeP/q9wMgqVYLfBjSHZ++TcFLtJaL4KPgf8KT/lgZvrLq4TaLMVq8EBfVJzvPkOVXGM1d299Z8lGGc4wxl+xvHQiUYiFYnIkAqC8IgQSFRKnmpsCCBTkD1CwCAvGGaS/XkHwaNUYH96xLmdLZ54/AIvvXKd4F2sFtuAkJIkEUgEeE+q1doROZLBgxJgo8mbCY7eOPI8QYRYhh5PBizmq5goSBhkOXmaM5xsUK3mFHnJ7f0DnnzyKfLBiD//7kv0zZKnn/ocZTmm649RMkFKRaIllmgElyRgjKHuzFoFBfYPDtlzlqbtMM6xd7TPZJRTzSvyNKFrLC6N18VTsLM5YlEt0dkAKVqEVGTFMHp0VCuUgtWqBudx3tB3sQuxMSpZ9T075zeYT6eUiaLzGhcEWmlWqxXOSvzJRMNJN4IASLwL5FlCWaYcHB1x584+gzJDCM2l559nMhkDcfumanjtlVcohkMuXLpAs5pTlgVYQz4csnv7He7s3+XNt15GywwpAl/80pMUoyF//Ae/y7yesbV9jr07+0itmBQDHr18heFwyHe+/adM53OscxAEaZbS2p5/9t/8Fo9evcRTTz2Js57edNR1zcH+UfRBydLYafDgCfgASuioSOUcIjic97T1AttamnpB3RryImNUDoAlWiWgE7I8x6ERUpAkkbOilcS42P3xIYoTaCHWsroBLSVZXiKlxntH1/dopejaDmtatA6Ydk6qAoqUuupx3ZKmWkQTymXGfDal7RpGWwO6es5goOm7iq7rSHTO1taYrusQQXBw55B5PidRmqrqaJoGsMync8aTMTpRtG1HbxwXNrYe/EH9mUN4329BSw6eGXPwuQnHT2yy8ePHuXpnxK1fu051abUekfrpBGAnR1lTislXnrTzDKeWjb04bjk6NszOp7z5tSHTx4dI69fV/w9B51F/PMPf6d8VOP9Uzuq+RC4MNeaFEe7zgwcGx1Ed+uOvLOaEAR8Eu0vJd29rDipJa2Oh4GG7UK10vDiccbnL+UIzPk2GwjqeLzLNuEzpjWO26tYd3/f0qdaf7VfbN/m/XPu/8gv7z/E3rr5AeOx1NsIdhkaTu/eMmT2kX41NZ9Sja0g5oWhH8bGko02XCASrwT7LwUHs0gvPdHITJw0g8NK+K7GIl+QsuTjDGc7wHw4eOtH45a999XQWVQpBVa+QeIo8pe/aaEiWNAgnCcHhEKQ6xRpPCAJrDW9du8Pm9pg01Vjr1vrwaydtAlLqWKmWkCQJ3tk4PmUDBAchQfjY4VisWkbDAUJJVnUHUiGEItWKLz37Rd689hbzxRE7mzsYYzmeH/PW22+wdf4iWilkPmBQJgzzIW3ZY1qDlBKlE9JIEgEFPkjUSZVRaYQSVF1L1TU4Y8E7mkTy6utvcuHcNlpleOHxUqIk9M5yPF1gjWE0HDDZGkdJ0+DIdUI6SBlv5DSNYVqtWKwqHrl6notXzvPSD97k+lu3UQqOrCGg8R5kRlTAUgoho3W5XBNchQgxYVOeLIsSvdOjQx6/fJXf+Nt/C53nDAeDNck5MJse8/rrL/P2jVdRMrA3fYy7u9eYZAMu7FxmURu+8PRTbA0GPHL5UVbtijt393nx+9/ld37vDzg62uPpxx+jXs6Zz1qkKphsbnI8nfHiD77P229fJ88L0qJkezyiLDQyGZCnCdZaTN+zu3eH5FjjXRzNOOFHIAWDwRgXHE0Aay1Nb6iXM4y15MWAgCQtB1T1ktY6hLVsSEWSlqRakWUZWqdIHTDOIawAFFpqUhmlbtVaJS14T3CBYHpcb5jt30Uicc5gncFYh3U9wVnKQUoiPKaxtN4RpCDPJUINsMYwOz7A+UBZ5qgg6NsOVaSUZU6eFSyWK7QuGJQbWOvZ2jlPU9XcvXuXNEtwPgWh2dgcQ1B0dU1vLZubE9Iy/wl8Ffx0EPx71h40oX8MgkIEWFwumF/aoLj9FBdua678S8Xo5hhhJRuvb/PSf/Ei9fkqjs/9lJKNk3Bz+07Pl/5gTlbHzub9ZOvJgUH3nre+OqQrFe1Ifuge9V6DODCfAQn5IXGqviQgk/jzKaFQuKdL3OMFPGAVgvVXr324oHt9sPVzBdNG8r07mjeP1mOwJ/v9mKfsReB3Ng8oveaxtgTACM/uC0f8w+2nGQ8zOuN4+a1Dvv/aHsv6/d0SpSRXz48o8prq3Ku88/gROoFbYUDiS0pzT91KBThXd8gPSDYEMOkMygemecL1SUmVvIRyr1HWWwgERjc0+TxeDeFPR6AehLPk4gxnOMN/qHjoROPqI4+wNR6TFykhBA4PjxgOSrROEFIghVw7Vxu6ehGN75xgtVyyWs3YvXMLgmG6WPDUYxd5+bUbOOswzqKlQgSJMIZEJgQHnTEIPEqC8QEl1mpASqD1OlHBgkgIa4ncECQCSds2DMshRVkyKQcMNzaZTvd59rkvMp0t2BgNGBYlz37hWYR7nWxccuvaTZSOHQ3vExASYw0yRPdpKSRlmmK8i7KLNrpYSwQCQdf0yBCVgoJUsTLdVxy8c0jbNAQcOMHGZIxpexbLJVmeElbgXZw/3tzZol2smE5nICBNBcYEhIzjAq7vcM7TugzbW/q1j0OaSaR0aBVfTmMss/mU//I3f5NvfOPXsAa++KUvkg5HFEVOoiO5HmAxbzg6XPLs48/QNpbpqmaQDnjl9df43o9+RF21/PClDc7t7LBatewf3MF0ipd/eIP5csVgkPNLX/krICS/84f/lryQzBeHXL91HdNUjIdlNDqUAefBuhrTVyyqwGiwiVIZIcQ1B+8IxNEvZ9dO3roj4KmWc+7evUvXW2aLOeWgROgMqTRKS5IsI/eBLM2QOkVnCc4HAg4hA4SwdqtXFHlGwJNKkMLRH92Njvamw/QGZzqcszSrY5yxMckw6/E8b8iLnM5ogrMMR8OoitV0tE1PPkwZjScondCbls2NDW7f3sOHQJqmbG7scLh3GIUOgmRQjumallVVszg+5vIjFynLnLquWcxq6lVDNiyQqWZjWFCWaTT/+xmF3//fvecRCX4A6/n74QyGP/6AJ0rYfH2LL/yT53n5f/09+mHPTyNGPwkP08bz7J8sKFbrTkUk+Jxup2zg0VdqlAu88o0xNpEPXFoIoH68QvT+J59jnCQYicQ9XeIvpfgrOaFQhPvY5A9choD6Fcvyzw1sfuTBALBe8Pax4vZC8s5UUffrQs0n5tBE/sVSWV4eLHh8nWi8dvWA3atzNpI4Vldkml967iLPPLLBO3fmXL+7YF7FkcMnLk+4vDPksYtjlLqXBJ5MS/VK0L/HCPOo+GB1NxECiY+KZEbJNZ8m4HXLfHznfWu//98znOEMZ/h5w0MnGt3hFJ2XOA/JICfpDTaskDJBZQlWKvJySCICQeexfZ0rkjQlLwv29nZBOJr5inMXLrG5UXJ7f4EFUqUZDlKcsUilCd6RpFl0afYmzuj7QHAOLRRJolEqJazZfkoqbB/n7WUGy6pic7zNslty/dZNniDw/Oe+hLdQ5jmTzS2unD/HxnjCsMipjUUKtfaciAGCkCLuf61SIqSg6w1FmZNoTZZndC1r9+cop3uSfDR9gwsW7x1aJ2RZIEkkWiuSJJBmOau+Yz5dYqxlYzM6TgcbRzGWswrbWYbjkjpIJIELkxKPwkvFwf6Cze1Ndu8e4pzHOkehJEWZgI88mq6t+NErLzObTfnaV3+R4bDgD37vD3juuWd56qnHT6vBjz52hXM72xwf7fPKGy/zyluv0vkVRZHw1s07HOwdM59PuHV3j66z3L1zwFOPP0nbdTzy6CV2dgb80Xf/gOA003rGlQ3Fqplje8NiOcdah3Nrt3il8MHgbLzOFy89Rp5neO+w1kbys4iJhrUO0/fs7++yXK2oqpqmrtjY3EYqhdIpUkmQGiEDSZKge48QCikVSmuCC2itSLOEIQFJTFib1THVcoWzBmSc3zbG0LQNprdY28UBNKFjEmSjc7lbz7E3dU1epwxGBUIKzl88x+HuPn1vqJY1h/tz0ixDa5hNK1zv8AJ29w6xxjOdHnPp0kWWxxUHeweUeRHHwEwP0xlNnUZZXx87J1me0Yv4PjBtj0p+djsauPEHP/5Rcdg6qdh++Rw7P7zA3V+5RRDhJxbAhVOfh3jcJ15aUc7d6e/vw5rsvH27R9oAyZovEB6g1KQf8Phns/j4jxCErQR/PsU9O8RfTN/FAXmowwfIH1OUz2m4Yx74pBPOwe5S8ec3NbcXEh/u45t8ynM9STbE+mWZTWoOnt1DJf5dOw/Axijnhc/n/MIz509VqE68hN7bVxDv+fdhEISgVx+UpJ0lE2c4wxnO8F48PEdDKySximPbBq00wjs8Ft87QuexQLucx+q0TtBSYcSJKo/Am6jK0jQ9o2JArivwAhU81oIJkszHm0lnujjK4gPBO0QQCK2iPKnWCCSTwYhVs4xr8CC1QKmCqmm5sF1QjnKaYsBgvIF1AS0VG+fPcfnyJbbHY0AiJJw7t8W1G9coy5LgDWWWIpME21ucV/GG7T0e6K1DKE2apQgEOkkQQpLmJU4oOgfWhPVok8baOSDRWUZZZIQkp28M3jrEmivgXIe2nto4lsdLVJpgQ6BqLV4GRqMBVdUw2RlzeDiHIJmMMm5fj6NG3lh0PkCh6L0h0RlpnnP1kctMRiNmswXDrZ5f/at/lY2N6Fot1oFRAG5de4s//ta/4/rBTfb35wRvOdibsmoNrg8s8g7vAm1jSZSKiVeqmC1WZAXUbcNq2ZPIhBe/8zqLRRXlhVNNcJFsrhLJcFgAkbwevKc3lt70eB99MqyNhP/OxQRT4qnaBuMhTXP6rkMqTVmW0QVcrflBbk3y1ookTcjzlAkF1nqkdzSLBdVqFcf9pEQJiV9L2Yrg8R76vme5XCGV4PIj50i1oKoasjQjzVJWixWrVYNKFONxgfOBLE0pipTRKCO4DVrTsn88wzpH7h2pViSpYGNzTFXXgGd//4C+M9y8fpc0kSSpRiWSVGsGoxylFH1vMdbig2NjYwfTGarFEonEeMN4wz/4g/qXHZ80FotFY1SvePTfPcHu128TVDjVfvhAfAy52/ufE9bRbDbPKfcGnPuLC1x4p0bqt6BYwslM/ek4zHpxCJLZhCt/+DgHz09pztW4xK23uG8hQmB+eYK82yG6D34tT5MU+ETKQ6FU2OeH2OeGhOKej87HvhwB1EigNyXcftAOAiEIbi0k37qWcFi/Wz477ufdo24fxKM4eTwS8h+8Uqs815/YpynePx51/7OkfH869fHOf52whZjmnGRMZ+nEGc5whjM8PB460RhePEd57hzGG4SQ6HJA6A0EjxcBgcJLSAUod+LWLbB9T913SKVoW0eSJtR1TZYl5LnGhQ7nA03TkucpaZ5iTQc2dguk1vR9jZQCrRKCcAgXkEqQ5zldMHRNg5CKIGJ35MK5y+xsb3LxwnkGw4Kj/X3avqHres5nOU9cvkKiJFIl6CRl+/Jl9A9eIlUpeZFjveXipSvsHRywt3cY1ZCcRQriGJSP0r1SSrROUFKjtMYi6NueLE+J1HaLVppl3ZJmEjUquHFtj+2NASiFA5JE4QPMjldUXY/QCmU9wjgcHc4LZvM2ckZmDU3TEQS8/sbd6D+iNM44rPekWYYqCoo8RRDY3tri0YvnefSRp7j8+NMUgxypxDqwt4Tgcd7Suo7h5gb6cJ++ajEE5tMK6z0+CLq2o0wFy2rB+Us7DAYlQ1MwHBaMN0Z0R448i/yQ8biI0rxNt3Zv16SZQMhIwg4QSe/GYI2h6zqM6el7g3UeIQKtaVFSoRPB1vYOy1WH9xYbHHlR4NfETgkkiQQnyPMMoSDRmmpxRLVaYHpDcJbgotyvlzHgkgoIkesT1gGiIKqLJakmWEfTWpIkZTgc4IVjvFVy6eo5lIpqVweHU9IsReuEw6MjtM5oWsOgzEnThDRJUTL6dFy4eA4R4PXX32a1WBGkYDguaOoOGQR9b0mC5OojV1muFgRClNa1jno5J81zhhtDjvan1HVHVnxCM7CfdawjvOGtMZf+5Cq3v3kDYQK6c2zeqFDmXtC+uFjQbqS4ZD0m8xBJR1gHk7pJuPynV3n8Xz5NflQg7Ym+8vNQLCCfw/Z1uPhjoknkemn1FuqNX+Nz393mmf9KcPuv3eCV//wlvHp/MhEGCr+VoO5296ucnq5E25Srd19gUO9w48p3WQ731l2DB7YUOA2Mc0X/q5u4p8r7L9snw/rJ1p80St672HjM61PJv3srpbMPJnmnj0rSS4rViwY+ZPpPCMH2P8zIHo+3pxBAvlyh36kZLCQ/ci23Hj1aJ2A/sbYQAHv2PO/0j/FLxfeI39hnOMMZznCGj4OHTjTq6YKwcx6pVFQ3ch6zWqHyDD0o6bsOV/UkgwHeGXzbI3WC9nH+XqiMtFB07ZK2aRimG/HmHwARCN7irSaRCSrX2OUMj8C6qEolRCBNoDMghSBLFVpBkWbQdjgkLkh8iDfj7a1NtrYmXLpwns899TS37t6mzEq0VJzf2aHvW3oLeV6wOZlQFAWewGg0pusarly8xPb2NkK8xv7eEVrreB5CoKQ8FWnPkgShYrDRti4mHw7wnkACIqcsY4C7t3uESnQ007OBre0tjmdTlIyV9SJTBL2+uXqBNQElYVWtyIuCg/0FKtUEE3DOIUXkhwQhEAoGk4KqavC2x1rHrVs3KdKCR58QZElgtViwmjsODw/51p9+i6armc+XdH0NwbOoKowLTKcL7NqRWyeSyXjIhfPnOdid88j2I/wn/5P/Kf/sv/5N9qcH3HhnTustgyRja3OEC57RJOXWzSNmxzVdgOEgZbK5wy++8HWq+RFvvPkW145uUlUVzhnaJpo3FsUApQWh69FZhuk7VCKQWqOCIEkSkkSTywKCx/Yt3nhM20bXbNMx70002OsNUgWUkATvsT6QFQOUVPd1dBxKgVaSRHrUMEMnmvmyQirBxjCj63vqZkWiNePRhESlVMsVo8GAwbBkPq05mh5gCWRFxta5TebTBTs7m1jT0TuYzyo2xmOkiJ24zY0Ro9EA5xxFkTEoc5qq4o3X3qLtW0ajMW1j0ZnEeo+pW5Z1g3MeZz1N2z3gU/rzAVsamnM1Se149l/cZueNBdnKIN29wZh+oKm3Mg6fHrH33Abzq+W6W7EOTN+beKxj1qROeO7/9RUufvty9G1Ydzd88Py4foPvH/2QsHmLR7Tj2UHKuUGGWlfOw+AI/+y/5EfXJK9e22D8hMZXDjWWnLhcnyYKnlMzvcJFI8CAoNWSIMBLx3K4j9Udfbr6kKtxorMLIZOEocY+P8Q9+WCC98fByfNf28o5XhqeTO3pdRNAmcLNeexkdFa8+0nvQXfd0V3/aC+OEGD+R4bNQSxQNG9Z2r8IDLOEWrdwrmKoAk0QnNC37+8tfVYQBF7tPs+ePc8vFd/7DPd8hjOc4Qw/P3joRMM5S922rAUoqeZzfNVQIEArVgcHuK5noxzQ1xWrgyl6MIjkXqmpjcHVPc7U5DrFe0+a6BgYBxnN44RkOEiZLWcgJFla0PWGYFucErigKbKEPM9IU8XGZEh/OItVfQcgML1hsVgh0Tz2yGNRySgItjZ3cL1Zk4MBKZDCIzWRX0HAG8doVDIeDsiTlEW1YrWKBljx5noyBhF9RdI0OkynUoPQCERUfwoelYC3iixNmLdznBcMh1v0bcX21gYH0yneG7pVRzosKMoh5VZBkg1oW8titgLVUxR5LMEHGWXfjaNaNASdkmQpwTqkjzKwi+kSay0r75ABiiLj7v4NXvxex8uvvIhA0XctXd9RtS0HB1MWi5rhsKAoM1bLBcfTOcY6xpOCaukxxnFwOAWl6JylHI4Zb2wyX9TcuHUAzpGPM4wK7N49JIRAUeYs5zXOtrgg6TpBlmR89YUXyNOMpvkt3rl5Bx8Cztg1kR+k8OAFWieAxCOQWhGMQSPJkpS+qTF9R9PWNKs68jVUVIZCBJq6ozc9IgTSLCGsnZXTPCErMlwfAx0pIVGKYqCitK0q0VqhnUYnEpUIvAwkIqHvEpz11KuOJIO2cWjrmFYzus7hHVRVx2K6oio7vLJkgwLmjlt7d3DW03YGYwxZnjIZxSRlMCjwPnrHLJaOpo3V7cVijjGexbKnKAdkaUboYTzZIksb2upnlwz+qRGg2W6Yfu6YwaHl4stTkvY+gjZACKSVJa0sGzcrHvuzQ+ZXS1wi2f3SBvPLJasL+ZpVELdHQLJKefqff+E0yTj5863+Lv/N7L/nW9W3aX0HU1DXBLlWPLk94MsXJzy5PWR36bk+Dcy7gqSoKX+zJPk3ku3/KOP8sxndliMSv2Dr2pLHrx+jOs/IWJSPxOIfb4+Y5ileeo4237nvxO/rZayLMyeJU9hJ8DsJ9ktDwnYaeWXiMxzxCSCeTbjRj3npt1v6255g4tXbyAPHjcD5D5erff+Y1AdvGHkUAbPvOPgn0S8m+JgFHpMSgBeGPU+PoA7QrTkYxyGw5/1p4vFZnHtA8Gz2KrUvPqM9nuEMZzjDzx8eOtEoRhNMZ1CJQmlFMRzhswKkwDpHlpd0CIxpscEj0yQGkiJQ1w2L2ZJSe/COcjIgSMX2OGexWBK8iERsLZgva5yFRGmM7dHrADQEtXZnjk7NiU7wXvLIlUfYPTri8PAYFUAoQW87lnVD19U45whB0nQts+mS2fyYpllRN3WsnM1m3Lm7x2x6TOcs+a2EPMmYrSqctxA8TVMjE4m30bvBe4cJPqplEXDO432PsaC1xgeH6OJNUxDQKqPtWhaLOVevXsD4wOb2mKOjCgGkg4zEaTa3zjE/qsAZpLAUeUbfOjKVMNrYQCdwsD9FSIW1jiwVa1M6gUfQ9gaFQAtNmiussRjTM50fMXQDOmto6x4vBdZaylFBVqSY3rBaNVR1T1akbOYJaRY7L9Z6lsuaprqJ1opXXnmJf/z/XDI/Oma5arC2J0wrikGCImBD5GRoJZF5gUAilaD3NT94+UXG4xGd7fEBTNcyHGzB+jW1IY4+SR87DYlW9Mslpq5YtQ3NqqJpa7q+izP02QAXAkLotaGeJEsTAhZrAqb3WCqqxlCMRly4ssNyOl+T8z3D4YDFaorSMJkUrOY1nQsURU5WRtWq2cEK10d1tKa1TGc1tpeUOiXDolKJ0ppUSxobWCxmlIOCa2/ewLlAXdUkShJsHxPpzQlSJ7StxdQ1UmlMb2irNiZWQeJtoG87gg9ID8N8gB4mFOUQZwO3b17/CX4l/GxA+MDn/82d9ycZ7/0ZSBrLzhsLAC78eEZfaq7/yjne/PWLeCVBwrm/uMDn/ukXGd4erQPmaBj6385/m/9u9q+ZucW7hKZ8CNTGcWsuaG3CW9MheZKTyoRJbCYQHPS7jt3/sqG4At94wlAc3QDnyJbmXaNeAJl1fO54xY93xixT/SFsgBBVpJ4q8RdS3FMlIbunpHTKTFh3aU4I0eLESO4T8FaEEAy+klB8XmOnnvnv91QvWQ6P/ToR+CSEmA+GQNyjvXAvSbmfny+BgYCBkBACm0JwQQje8p75Q/pffNQqILCpZqTCPMhi5AxnOMMZzvAReOhEo2kbRsMB3vS0dU8gBtVKKOqmRqoEWQ7ou+jgTJ7StR2dMXGMRULXNZRZDggWyyVFUZJmORgTiXs+ksXzPGWxqMkyhQkBL+K4lOl7RKbIswypBE3X8OznP4+QUFUtVVMTkPR9x41b15HKc+v2be7u7nIwPaZtOqz3a26CY1IOSMuc+XQRxx+U5sbtWIE+MXEzvSFJEqQIa3K6w/YdeIcTCSrTkR8SJKtqiUIglaCu61glVTqOWonoPO2l5/r122RZSpKknL90juOjKdII7h4eYFpLkmpGkxGdgcWyjj4MnWVzZ8x8VuO9RdgeKSRaEXkrwqNVoMgL6qqlWja0qkUI2NraACSud/TGYCzkaSAfZMwXnsGoIEsDs+mUzWGKF+CcozMWKSXYmFz5AFW74mixIgB90+N8QApPIlLK4QCdKDYnBaYX7O0doVXCYKRZrBa8+L0/xduA6WPnR8hImEdKCIJ6taJdj1NF34g+8nRk5JU4YsDvQyDRmiSJZnoEjxSS4AN971kuWrx3bAxHDNMhi3oGCIZliQoOJVs611AOCpzvgMBoNGI0nDA7mlKMc5xx0SCwsYjBgPlsyeHelCA8A50hhMMIQb9qcBK0ViilUVrRd4a+6kBCIiNfJ89Settje8PRsiIrS0Zlju36ePpe0XYtkrgfAoyGI4piRNt2WLdAKkVVV0itHvxB/XmBgGaSEKR4l5/FB2/7LlYyuveURx3CA2ul53azQhf/FrHVEoTASMEyUZzfeZ3/1I8RvF8ta/k0lBdGKKGZ3Q7YviPYnuWNAu/upQkKz/B4j6ybUQpzqgr13nUBJN6T+A8h+4dA0BL71TH2hdGDncTXCYU59sz+ZUewgY3fyEgvyfiHj5Nw3LedLARpodj5n+Vs/49g9W3DwW82+D58ohzmk+ClPcXlsWOQxiTKE0+nAvrPJMlYI4DC8eXBy8gkEKy4Z+J3XyL0cBCfKMF7qIN8jAQvFsp+Wq/UGc5whjN8jETDLlc0Mro0O+vwgJYKLSVd32AQaBHwSLwAb3o6F2fjrYvSoFpovFIcHU8JSJLMUWYZPoAWEqkVUmpEcKSJjkFXsIgQSb9JotGpwvo4cmONY7lacHS8T9OsqOqGKqw4PvbcunWDP/pzhZSS4BwIQZqkuBCTiFjlC4BHrCVOZfB450GsfRzCmi9BwNn+tMKYpnmUYw3RvTzGOS4G/OtL6rzDe4+UkdMhhcI6z+HdIybjEavWYF2Nt4KmbikHBcJKjHUMRyN2Nre5deeAJ7cnPPfkI8wWCzqn2NcpjTEELyBRSBFds5uqo8gTkBatPXXVY5Cskprj6YzJeIgXniRVOG8ISOazikDAdpaqaikHCWmWM9waYK2lrlucgcWiobM9k2KIVjl5pnnmiSd5/Z2bBBHIUslgmDHZzHEW0iIFLEmiyHKFThK6xrFctFFlSUq08njXYbqWql7S1A2m73GmxTgXO1hakQpBEIrgo2N3liYkWqElDIcj2qbGug4QsePkDD5Ew0VBQAuPInpoKKkZj0cslw22hzt3Dlgul2xuDtnfnTJfrNg/OmY8LsmThL53lEmBVp4gJWle4IxBBUGKR0lJMigwbceyip4mSimsMSRa40PAh+g4r5REesF8umRjMmI8HIB3ZHnObLnCCdAygeAIOC5cusKjTzxDu6qYzw+ZTB5Bq5Tp8RQffo45GgKKo4LN13d456+1XP3eMUljP/p5ACHgteSNX7/I29+8iD+VKBVU23PeuXKLehwIApwQ9EoCEwomH7iOHOAw/nppwyJUIDhBV/r4+VxD4/nCasmwOvH+eH+CgRA0WvL2ZMBx/sH+DaeytTsJ9vnh2vD0ARDgqsDh/7dh9Z0oS7v6gWXyzZTJ30jRE/nJuhsnuxcCkcHoVxLUSHDwTxvMgV+f3qcPYj8sX7i7lPzBbc0XH+2wwLGPxz0Z7P2sDh60xD9ZMPoFRZ+eR97qEJVD7vfIox5siLyYkcZfzt59cAFyt0PMbfRLaT2cqFd91OFPdkCAXBFSQZgk+ItpPP7t+PkXvYfGnSYk79PXWj/Q9Y5l3fPOnTm39pccL9rTbf/z/9Mnv0xnOMMZzvAweOhEI5+MGWwMMX0PqFhJVALhHaIVKOeRwRNQdM5DIlDC4k0HzjMocoTxrJYVqZLkRULwgURphPAkqSYvNXmaENAkiaIsS5wX+LvQ9x0EMI3FEyA4+uM5r7z+Bvu7u+tgTVLkOYG1pKqSSCVx3iBE7DR0rQHvESrBBUvfBLqui4EhASECSkratkbpBBA0zSp2D7RCCgl4nAsgJdOZQQqJkgJrOxxRllRIiVAKIST4gA8xMaGQrKoKhMQGjzOOySCa2pGmtMowO14gfCAvBtxZNIRrd/jc449Rz6ZgIyFdao0kuoFnOqPte2bHNW1roiFfCAThaeuGOzfvMh3kDIYDxhsjBqOcxfGCetmQFTkyTRjkKUJA07QktaRIMh69ep4kLzk+XrJaLnnm8UeYziuSJOfcuR2U0njXk2YpSZpgTMBay3y+JEvzmDSiaNuerjE09ZLBeMj2ZEiSJAipkUqxWKzo2jbOnXsIxKRMJyk+CCTgvMV5hxDRN0Mniq3tLY4OPLY2MZHMEjwK6hpClOF1axNAY30cqXINVWXxTtL3PZKM+bRjb3dJ3/fMF3PauiNLEpy3FEnBua1NlJKkWtC6aPo3mJQMNkZYBWUXPUC6Po7xSKJBYRAZ3gUI0Uk6TXOSRJJmCdZb2qpBSEleFNR1hxCCtukohyWXL1+hLEqKLGdrZwt8oK5WXL5yBSmTz/6b4GcFAoqDkq1Xdqgv33r454VAkIKbv7jN27924TTJED6g3m7Q351z0ClI3j0C9FAhcwBf3fsqTdOTjkQMFgOCt5IBmXdsNfdJsoZAEILDImV3kLNINd26W/XAoSktMF+fEBL5gC3W2/nA0X/bUn3Pns5SuYXn+P/X0t9ynPvPCvTG6ZDVx8cJ71vHkSqA3f9HjW/Dp0pgThIMpQNl6U7UQt61zWhiKbZ7rn1AZf5TpzgnSUap6H99C381J8i4Vz+J5ylcgNYhGk+YaJCCoN5/ZOEC+IBYWMTCIecGeav78EUGCJsadzmP74+NhDBSp8cQIcDapV2sHGIav/vkXoc8jAmlODKIxoELvH17zp+9fIfDeYO1P8Oy2Gc4wxl+ZvHQiUbV1mRNSqIFcj2rLACkRuUFwdooTxoECId3gbptSZIUaPBCMZ0tkAImwzFSaRwQsARnGI420IkCbwkh0K5n1E3wdH2P9B6cBSnQiQbiqNLGJI76SKVIiwyVKJyVSO/QWlAWKcW5McvVHK1SFssF586dRwpLIFA3hq43aJ3EMZw+oPKUvncEE+Vjl1UXzxsiUVm4+HOSUXcGGSDREqklPni8tYxGQ9q+pe8MQSuKMqWqHIf7FXmqSbKEYpCRDYe41uB9T1036ETSWcNisWAkBHXT4nrDwaJFJgmLtkcqQSIVUnhWXU+WgBSB3nQoA03XE1wUxlIi0PeWqu6Zz+voYbIxJC9KbB/Y2ByCcCznjtmsYbVqOJquSKVkNBkz2ZT0pmE4GnL99i5SKLaGAbwnUYHJxog0T2KgXffUXUc50MggSVKJTiRta1FKkKcpwUTzu/MXN9BJgU4yyrzEdh3GOUQAtya7bmxv0lY1iBAlhKXHOYuUglQrsmLAZMPRtS3OR4NEKSJR3zkLIl2LDMSAz7sGa1u2Ns9xeLCPFAqhIgeoa1u0gixJER7KtAApkTLF+lgVDN6RJQl5nmOdp1lWdMIxbXs2tkesFg111SK1QJAgpcRj6Y0liJj45GUGItDX3drQLBA6Q3DRVyBLU5RK6HqL854sTei7hr7tyIocISVp+jNs2PdpcUIGf+aI8z+ev4/n8FHPHe827Ly5ZHG5pBvpOEU0UIShRsws95tyPHTQ+r6pGPG+n60UGPnByUHmPANj6VQ0CzVS4uQHrCOAe26Iv5h9+PoENK85qh8Ygr/Pr2LN11h935BckGz/gxzxKXLW+8P88lHB8AuKxQ/sJ6OBnDR2ZGA0tpy/1FGUJ0NR79mbYN0Z+GzHf06UwUgl5q9t4h/JP7hrpARhoAmDeyv4wJUoAUrgt1PYBk8Bv/BwPZfwAclu5A4JSNaJz6aEzfgC+sfv+06oPep2i/z9Y77/2h67RyeCJmfjUmc4wxl++njoRMPUDSvnYkU9BBIJUkX5WRcCOk3pm5agUgzgnUGh8D4Q1jPnSicM8xTnHVXdovKcKAElSDJN2zV465BSUbU9SZIADh+iUVyQoHNNUWQ425INNUfLyBdABhIdGAzifH2xM2K+WJKXijQJZGkWA+ZJjkpNnKlHo1PHaDPlysXL7O7tY/tA39coLfBe4oXEIaK6UxB4IgdlkCVkWYrz4LEEG9BCEYSmw5BlGX3fEZRCyYS6dfigQAjqtmGcRvlavEHqhPFkk+quZzmvMMaTaCidoOscvW1Ie4PSSZTGBbKsQKJidUvH5EIrhbUCRySyIqIzNkphvMd1lv2DKdY5Nrc2SFLFbLpiPCnRaYIA+rbDC0e+OWJje0RTV/Sd5dbNW3hj2Bht8MgFQApGkyFbF8dY19FUjtV8hfPQG0uXwnA4wK6N+No6tvu1jkPxzgVC6AnCkacZSykRXsQOhIguJEU5QAtFVS2RMjqwiyARQmKsREnN9s4Ox8dHhD7g194Y+LV6jRAIpdc32EDfebJ8zOZ4i8VsSqCnHOZ0TRcdzIOjKAZY25PlBZPRhLatsdZjbTSrTFIZ+TcqoTGCqu+puiWDyxPKIqNterxPCEGBiEGeFDLKlwaLMwLvwBqPcfG11xLSJMdZi9TRyBEpqesFSozjZ0JJsiwnTTK67md4dOrD4qyPqPQCBBE4eGGX5uItnvsnR8iHrdKuhRk2r6/4xX/yNiZX3PraNte+cY72Uob7OzsMXl3yyL/aZTdN6ZW851rxKeKzAMgQeGTZcL5+z+smYoV63BnGneExWUeuQaJZpprbo4JKa07GYrTzjOYd087Rl/qDJV0DIKH+scUe+RjsvvugQCB/SiNSEasRn+D8To89M+jvLFB7HVdkIDmfMj1M8OvRsYeJbU+7GCpw5bGG8YZFyg88u/ecx8Ov8wP/IO79VXqFDJLhaof2KwntU3GELjLQP6Bb8aDzECBdwMt72ee7tn3IYP9htnrQfsNA4Z4pEY3nbywe47///bc4mNXvFgU4wxnOcIafEh460VCpJh/kSALeR/cmbx0qjQFgCBJcwDpP17XoNEHpJFaJ1gZ3o+GYdrWgdT6SfF1UBrp89TJJrpgvZlhjUFJSZCnFIKWqbFRS0pre2ejwnUiCSnBVG/kdUlEUOUJIsqIgG0uqriUrMpqmo2sbdJqhtaIcFQQbaOqerY2SgKFIdaxsK8fk/IimEpSDAtN7gte0dU3w4Dw4FwmCQiWYtZO1VhqdRedyiSQN6br7keK8p+1iFwIfMMaghEAIhTWKvEjpq55OtawWUS9fJglSCHpjUBp8gK43JC5grMV5SedaBIIyzaJ3iJAopeisRQiJs7FCJwHpo4yvEALvHPNZhUxSLl4co4F8oMn6HC45NjbLGJCdH3Lj2gEHu1NEqnEmKlDhHLkUkRTvHW3T01UdNgSyYYoWit46tFSE4OmMYbGscM4zGgwAz7JqGW+kSG9QUlGUJXKqkNKdGvEFosFfublJ21SE4KN0sIgu853tCcExGm9SFgMqPKbvUUqTJBLvFDIEEpUiT+SCVMbG1jaT4YQsBws0dUVTedK0wIeAUoq6hqppybKUVGmk9LTekamc3liU1KSJwnYt2nq0TFitOpwLKK1wFvACgQIEIngEkX/kjMU6h9QJmUpxPo7CBSWQaKyLnROtNG3VgvVMJptY10VhgjRZd/R+RpF/kCdEAJOD+/Dyuksdu790hxt/64d88V/cYLTbrMeCPgareR0IZivDU7+/y8WXptx5YQubKg6fHjH9O+eZ3G7ZExpxrYXev59X8REI7/nlYtXy+LyKYy/v3c99v6sAhMCkM0w6w/m6Y3eQo0JgozXIEMjvelZvHPPWNy+y+6UNXHKfYeD6FO2hp3vHPoB//J7Q+1PEnMIGkj+aom5GGdpEwqWrLUoFjg9SrH34nUsZuHilZbJpEZ/SiO/+M5TAexkvARhJQelywvQqqhuzdfgMymvyboS94bm1usZb/8PXcOnHMOmTsHltxdO/t8vd5zc5fHpEP0zw6qfrJn7S+bBfHLD1+oi/+UuP8tqNKdfuzqkaQ2/OjAfPcIYz/PTw8BGLEAiZ4gnIPME6h1Zxntj3cRxGlDnKWqRweFTkVay/9q2TGKupbNSfz7MEUAgsEsF4OGJ/fw+kQynN555+hmVdc3f/CGM9WQIql1gs1sbEpkhzNja2mB1PaZctUgqODqcEApPxELH+X5qmJKmiX7SIVNJZi/fR5Xu0NcbanqODA1KdcnHrPDfrmzjtGI3HpHrI3sEezkn0epxBSkkIYP1ab3FtHJVKQSIFGk1vTHSqlhLhNR5o+wbnO4LMOZ6vOD8u8Y0i9IE+1AwHBdWqRSVJJMKv/UaGRc58vqDuDV3vSHQUyxFB4bEIJEma4ILD9R4X+shlwOMtGDwJUcYTKUmzwIXzA2zXEZKEblYxHAwYTwqGkzFPbu3w+u479E3LZLPAO4/rYtBcpIpRqtE6YTIekaUpXd2SqoTlcoXpPaNByeakJC0V/UGLWjuEWxMTLtt75m7BzuY2OsnIJwV6fxfrzH1iLgFQbEw2me7t47xfq1PFYE1LjfcBrTXloKTrW7yKBPwQJCdKnrWHECR+Pe4tUKAkRVHQu4bOSYQ0CALBxep4XhSkaUrbWUQWGGQDmt6QJgmdsRAkQqg4yhck3gfqhVl37bKYjK/J4s5LlI7yzX6tkKO0Jni/Npc8+Xitg2ApcL2hrZeMJpukOqGpKyIvyNL1CikfQBb+WcBX/9n7HgoCQrMN3RCAJU9iw/DeBgIOfmGP+RNT5k8dc+X7h+y8sfyYScZ7IASEwOCo45l/dxeAbphw4+s7XP+Vc3QDjdrrSX7nGDk3D6xsv79iHqKHjXWkzjPsLY8tatRHKWPdvy6AEEid59FF/b5NRrsNX/6vrnH1xRFv/vpFjh8fRh5BiA1itwqYo7jekyr2CWQhGP2VlPwJ9a7Ff7JGU0BU7w5aBVAMHHrmsfbh1dGKgWM0OUkyPhtclZJtIRh+wOsmAak8bN+FV78E1XlOrkK6git/+CjX/+O3PlaiEQRsv7Xkwo/nbL+1ZPrYkB/8J4/TjX66nKrTzoqHUWt4ZqPg6rkhvbnMazdn3DlcnWlOneEMZ/ip4aETjR+8/AqjcoCQgaYzLBc1Fy5skkq99o6wIBVZovFCIwXM5jPGky2KokClkZgdbBw7sq7HB4En0Jse21u8NTjr6ZzjnRs3aZqW3sRgPsQ4l1wnjMcjdvf2wHkOu32ECBSDktbWKKEJMuD7nuFmSVO1DMqCqqvRqUSogHICJxzlMGdUjLizdwtnHVmWcOPWO1RtR6oz6q6jbXo2JkOytKBaVfTWoIXEOI8NBjzoVNE7RzAW4y2SQJloZIhk7bZryPM8BstonDUIAb0TBGspN3JCF8hlgp5our7D2kBnO5x3pA46E2KcjUeIBCU0UgWc9Sgl8MERBFhnAIEXFqkkNgQkkiDk6bxxnmiS9XjakI7OeBAr2sbyxIWnaVrL7RtH9L1gUGrQkslYUq9asrLEAyZ4hAtkStJWPcZXlHnCpZ0NPvfEJVSqOF7OWVUp1p50Fjyt6RFeYU2HTiTSBwbDAYOypO2ad41ke28jHyVL6L1BK31f4CSJA1aKnc1tFss50ioCBiGJibBOyFRAyNgl8cGidOwqJLlg4Ev6tiXRKdJ5RDDxvUxMYEQIaJng8CQ6Qeno5SKDQHaedD3q1QmFISA0IAXCRyf7GLfEBEGsu1gnXgYnNU7vI0k5kkskQkpwUUo4LwrSJMEZS1PXSJXirMX+DKtOXfvK+4OuZqg4vmyA6fr3v8Al7w6FAgGk4MKPZnzhX93+aEnbh8F7AtBsZXjmd+5y4ZU5f/6/eJr+Ykb4j7dJXlwgrzeR3Mu75+cBtPfk1rPZ9qgQ2Gl6CuvQ6zV+YCfjY67tvZA+sPPmgs0bK+4+v8lb37zI6lxGCILsUcXV/+OA2e/0rP68xy0CIUD5Rc3W383Inlh3od+zz7S2yPdwXsxA4/S7uSX3r+zEvyNkimYmOLibspzr09Gph0W11Fx7o+TRpxryIrL3Pi0HYyIEG/ebn7wLAXAwfQoWF1l/KE//FD7BoUWA2SMD3viPLrH73IRmK6Mv18qFH/ct8AlPPxCfV0w7nv3t21x4bY4LYJTk1WXP//tHe8yan2PDzzOc4WcA0YPtpLC9FhW6DyFEy4Gw/vkvOx460dg/PKAuKpTSNG1PXdUIDFIoVKIRzuNRTEYDpJYMhkOc8zTLFW1T0zUdfVvTVQ2mbWKvQVuESpiv5njhMJ2l63sIPV3XkuYlCAtY2rZH5JLWNpjbt1BaY4LBmQQtAlmRUGYjqmpB0JpG9KR9ijGWqq3pWkuR55jO4DpP3xvmixW9gbpaB25CEHz07CD46E0tPY8+doHhYMLR8QF37uzivcB2AoQkywXjcUHbxCTJ2J66MyRpQZprmtbQuYBrGgZFwappsOvRoixNIfVRgQlolhXD0YT++JAsT2PSIESUS5USL9Zke2JcGkLkIXgh8EFh+g7rAlJLtFQoAcF5rHTkgwSFRCtJOcrpnSVNBBZNwGM7SV23tKsFj1++yEjl7NWrtVO3JSlSht7hvWPWGp7MB5zfuUDrPdlwh3EuGCQdj1zYZlAoWhcoihIpZwgcddXinSXNFEoJJucnqCRDKIXUisnGBtP5DCvXssAEnLMMyiGbm9s0XRu9N9bxe1KWONcT8AwnI/I0x5ie4Hys6PlA23sy6U4Dh3gT9qSpJs82OT6oESguXblIng9ou4ayGOKCRymJEoHRaEyR59RVhZASYw3OWKSx9H1PZzrKpmG5WkbTQWfpgz/tUBAEKkkQUsXxqUD8gljzNpQIeOci2V1pRCqxyiGlpmsrtJjgXcD0hhAcyXoE8GcVr/6V9/tRfBDeFcwGQAo2rlc89y9ukVb2ZD7ks13cussx3Gt57M8OeONvXsKfS+l/Yxt5syV5cYHYj67zkcDtOFd3jHpDad37uxanJOyfQP14vU/Ve66+eMS51xbc+PoOh0+PT70ezn8F/BegveYJDsoveIRs4Np9+wnRxDCtLZNbNdnSvGu9s6sl/TAhCLjzwhZHT47uyQJrgfmNbdqXWo7/IrB40+H6h+dm3H8qIUDXSa69UbJzoWdzp0epT55sPNSzuhHc+RK4+26DnzB/PSmAHD4z4vCZMXI9NiZtLCo4HNWsZXZzFUclfeD6n+xSH7WoVPHENy+TlhqVSnae3oi+UgLEybV+iEtxsvSksTz3L25x8UezuAYg8YFfyCT/h6e3+K3bC1ZnClRn+A8ElydjXnjk8unvi7bjO9du0tqHlD3/S4Kt0XnObcTzUHLtp0VMNOR7Eg3vPT7EeMy6qMgZi/eeul1Rt0s6E0Vy/jLgoRON4AM4j1SgFSQ6rE3UPDYEEhEIIrBqG6QQJGkag+euIbQB7zytMbTGoREonSAEFJnmaGU47I+j9CzRMyPgaKolQUaToTwFLyBPU7yDLC3i+FKwCCkJwdG2cewo0Zo8lQTvGJUTtkcT3rlzB9uBVBqpLfSB6dEKtgLFKEELBV5Qr2Iw7JzD1D0y0SxWR4zLlizRgKbvWrz3DEYl4zLD2I7z5y9weHBE8B7tHctqReEzpJRkSQp4yuEAlSoOD45xKBZVyzjRUbGrs6xWK5TQVG0LssfbQJHn9LbHmI40iyooUkb5x6YLJGn0lDA2mglK4cGD0hKVpnhrwXn6pkMg0UrhnWAxa7C951d/6Xk2N7f40auv0XQ9q8WSH9Ud12/v0XQ9iRIkRcbG9phq32OswwpJYz13jxpWXY3pPE0tKC4OaJyjny7wQnC8aGgagQga51uStVu51ilV3RF8x9Z2VH0aTzZIlMIqiXUWcNF/RMHG5hZ7B7unJEYpBMIbjIvKYUmaMplMaPvIWwnBReM87zhYLZCEtQSxxHUGOVJcvvwEzaJh0bZ4H+jbCuejEIAXIiqXScG5csjGaMBzz36B8XiEd4GqqqjqGrNONOqmIc9SmqbGGMuNGzc4ODjmeHqE9y1KJSil8C4gvEAShQ1kiA7tUsbE1lgH3hNC5HVY46nDiixNKcpi/cUCbd189t8EPyV8orBRgO4cj7x4RHnc3Xvws8Y6WJTOM9pvT8f4ghK4xwvclYwL/3qPp75zQOk8qfP3Gbg9qHL+E8b6mNnK8PTv7vL07+0+eNs/e8AuPJyGqfcncCFwbu2oDnDxRzNe/TtXuP4r59Y8akGzB7v/3GMO/f3L+YSnIeh7wd1bGc4Jzl/q3sXXCO/JAj6M+bBz0s340IMGSJo4Y+TluiDhMbrleHIDb/v1cR/mBCBdpWy9soNqNBe+dwkz80zrGS+1r/Bi80N2pwes9u/77IZ7Ccqr//I6AGVR8I2vfY0sydAbkvJ5TflFTXpRnT7ng5no8XHV+9jJeGW+FkCAxnnq9VjolSLhf/PU1sOc0RnO8DOB569c5P/2j/5BLBAD1nv+t/+ff85v/+i1f88re3goqbm0/RhlPvzojSHOzj8AIQRCiEXvpq9YVFOOFnvraZd/P3joRCPRApUoBmVOWFraIGJFPTikcFjvkTLg+0AXAnoZ1YW89xjnMaYnUZp0NGSUS3wA6x3eedJEs6prFAKPJASJ1hqcxbkQSbIegnP0nSBLY3DfmZY8K9HC43yHc5ZUZ5EM7WQcr/CWN27cpKoqyjKnHGYkScrO5hjrDUpYQhBIrRiNN+jbI+pVhw8OJQVaJBSDAV1rWM4rfPAopaLHhje0bWA4HJInAiF7grOUeU5vetquY1BmDIqc1vRUTYsSga3NTYIIaC0YlUOUkNjgUCrBeU/w8Y3n18G26ztEsvYtCVHFyAUVlb3WNxOBiAoyCJQUeCEJxkbvj+DwXqC1QEtBX3fIRGCNJU8z2taCTAiu56AOeG2ZNg15otnZGOG1p6samtZRZDm2rbn+1ttMD3YJUlA3K5wL5PkGw3SE63tEmmF7yHROZwyTMuPqYzvcuHmANeBsYFAIumaJH0U1pTzN194WIIUi0SlCSMoyJ00S6qoirAnbZZZSphl911BMNtjZOsfB/i5SqUhI92tPFCFABILvARnlPRcrhqMBOtGkXuG943i6iK+PlJQbIzQqjp1pjQ7nuPi1r/C5Zz5H8I6+72nqmvnRlKPjYxaLOau6YlBkPPnE4/zSC7/AbLbkO9/7Hn/xw7+grpvYCtUBgsc5TyI1wTqUVPjgcdaBkpiuw4U+ksJDwNoe7y2EaAYppSAyPX6+kFSWSz+crn/7jAP6+xSIghC886vneeuvX3yX63YASCXlhmTsHcqu5Zr+sij4nMiwfpKK/Ad2h9bXQwr6UrM6l3P7q1vc/fLm6TH6O469f9zcl2R8NtciBMHBXuQhbZ/rUTog1qNNW4cjiibl7pUpXvr3JRunVX2xtk59AK9GICCtCM/+W+rpZaq+YBo8VvXMR3ewukP9K4X/tU3ETkJaW+qNFNF65NSg3mmg97gnSvyjOUGBLSzHz0YHx4Nf3GX35WP+7f/5Oxwfze4dXLw7QTq9Zuuko64b/uzPvsfjW49RpAV8OzA+P2L7bxZMvpmix+q+c7jvnAVI63n693e5+uIRIgR+b7/iB7OWvdZyrTIf+Fb93Y/xupzhDH/ZUCSaJ7Y313od6y6vkBRpFNTxH+b8+ZcIk8EWRTb4TPZ1MqadpQVZWjAZbHN+8yrHiz32p7cxrv/onXzGeOhEY5DnbG9t0TQVne3wCIoiozc9Wkr63iEQ6Dylqxrmi3mcCx4O8ErjaGmqBuNaujYHGcgSKApN23WE9Vy8lhIfAn0f5529A9ZjNMZ6pPKYzmCDi0pLwZBlOT4YxuMhiU7Y2JmwWlXU8yWN61gtOnrr8L5FShiPB3R9E9WDdIkXgaPdGYesePqxRzg6OmIxX6HTFBUSmkWDdw6tJaPRiK7tWFUrukpgskBVzzk6XDAcZWxdHHO4v0Kr6I7ddj0bwzGz1Qok5Klmc7JFmgqsh6bxZAq0SsiLggCUgyHOduRpRpEXCO9IlYrnLyxSZQgfojKwVATvENIjlIzqTEpF1Sup0IkgVQVFOWKyMaFIPFpJis0SpSKxX8uUItWc39nAWocNKRvjAeN8wP/8H/5d3rj5Ju8c3oku3V4zWzXcuHGLum9RSSTGB2B2vEB5y3iQkgnNatHy6JVHaVcNSeaREopByvTQoHQk3XgPddMwHE4oBiWL1RIh4nugsx3G9qRKcW5nh8VsQZJkCBHoektAYk3//2fvv8MtS+76XvhTVSvufPI5nXumJweNpJlRQAmBkIQsggkGjAwm2GBseH3NxcY25t6Lsc2LfTHGNoYX+2KDfbHBRBmEJNAowEgazWhy6JnO3adP3nnlqnr/qH1O51ZP0oyA3/P00917r73SXnut+tbvGyiKjFq7jpQS5Xn4gYcpKkdXkgKMQioF1uB5PsIKsjR1GSXWgZFuv8fWYIsqL9ht9lJkBWWVUeQJS4vzPPCnn+Hpx56iLEuqsiTLMwa9IUmaMMrGWGOIo4j11TWazTq+HzE9NcXBA9dz5OgRyqrEGAdStXaBjqUt3fclnaB8G0RhJVEc0Om0iSZuamkyJvBCPM+nLP/8cKx3Bo75eeDqpRrbn/cQqnxF3vQ48aY5Tt89SxmpC7LixGTxk/fOsefBLVpnk5elqfKKlz0HuHr76xx9+wLDhZhkKtgR35vEUm0a1v/flOK0nuCtl+ZkbK/GaMHa2YDupk/rIMzfa7EK4s+3ue0zc2wtjUlldu47PG/7EbD7Cpkl22KJZ7tjvMhHB4bV1jH0Rf0SgUD2SvyPbREJS2NUoGdriNwgBhVMjBzU4QR9W4PyzR2MshSNcvJ5aN0c4jUNbG7v4tXPkZiI94f5iMdXntx5vbZaQz2hiH7D486v3c/19yzizQbu0EOJxBIkml0Pb3Hgk2uUQvDoVsp/PNalW+jJRBSvHlD8F/VnugTOMMdpUi+sehwxTrOXbFsz9Tp/5x1vdhOKk9LWUGpN6CnS8tVBHbpaSSGZn9r9sq1fCEEUxCzN7CcO6xxZfvKLruu4ZqARhjGepyhKlz4qhaOWCCAIAjd4QpCmBcpzg+w8y8nGY0AghMCT2/z5BGtd8JipKsrSpWZXWPwgpKgK7GQbSlpM5QbijhLjQs20MU5cayUCi68UYeBTq8dUZYXQmlarQ284YGrapzQGT3osLrQIozrjbITJIR9XFLpESo/RMOXM6iZ5NUZb6wL7ZEGjWWOru0U2tjTbEZ1mjUApilzjKUOaFbQ6DZSS6BKXbm40ZWXIi5Ke7tJq1siznDzPGQ6H7Nk9h/J9snFGGPgkgwHd7hZxHGGQZEmBpaDSgsEww/d8otjDComSirKs8Dxns2q0Zao5RaPe5NTp04SBotKSyIe56Q7zCwu8693vZv30CqIZ0F0/w8bWGuk4xZMe9XqdIAqomxp5WaIrj5nZFr4J6Q16rK5u4HsSXwqSSuNLwzgdo7VGeTHKg6Vde9HFFo2aTxDH1GvTGDMiCiNq9TrDNOH0cp/19QGBF+4MTvzAJ45rjEZDPM8lhdtSoK2lzHKqvMIPPeJ6TKNZJ88ypFJoa8jyjCj06Xe3iOM6YRSh0jFCuLwOKyS+72O1S2D2lDMpKIoEZX2KsqICsmJElmcEgY80FoGiLDXaOv2ERnL02EmefO4wRjvNh7GaqsJle+DE57YqCH2PvCyoCk1lnCOY9BQHDh4gzzIGvS6+5zlLXwlZlhKErnNjBJS6wpSOUjUcDsmzDGudVbI1OcYasuylu1G/2upy80/CWvZ8bgNZGmcVanlBYnA383veYMtzPNC85fPke3exeaBGGamdToYVl5nnlzCeDWmuppN9uMxCr0jZK5y8q+zcdifnfPcpKUimQ46+dZ7lu6apwkmP3ljKZUN6RNP7cE7VNS4FXL581q3WQvymiNbXhpQtiZWWsR+ydVqTvn8GfbKPXCkQWyXC2B2zi7oQRBcfu3VAYmwtJ7VmJfIQ0ll/A5yTXl5YYliRC0GOh1zLzy2z3e0yFvXkCDGoKN45zdzJEdPHRhy5bZrw/j5fc9cBPq5PcfhU95rO08VgxFrLOHeBe4Mj8Mc/0+WZ3S3e/ZaDtOohZj7At4apE2NINY9NN1k38N8eXKa3DTL+AmC8JLVNzZFS4kv3uwh9n9DzQUArrrl/A2e6m3THl7Py/rNf99x6I2+8/SaeOXmGJ46epB6FnF7bRErBe9/0ej78mc9zz6038PHPP07xEgABN4N/XpcQwXyzAQhqUXMySbTtflpSVC/OTCXwQjz1hZ3kLJayKtATiveVKgrr1KPmy/47FULQacww215kvbf8sm7r4rpmoKF1yXA4wmiL0RJDSZ4XaD0JSbMWIw1lXiGVy7XQsgQjUJ5Am5Ja6BGEPhJLnmtCz0P6Hp164HIXqhKDph3GCG0IPA/f85BGEPkenhKEnsBXPn4o8VVMXpRM1UMs8Oj6Fp1Oh3qzwdnlVaSVzAUeypeEoUezMcXy8jJYTZFpBoMhUih6/SFKKpqtGkhLFMZsDbqUVU5URVhdEagAK6FRb7GwNMXymU3ac5J8lKH8CpNpCgxJmrF7zyJVlWE2h2it0YCnPHKd4kmJA2EFrTCiPVXHB0Z9TVSLiYUL3BtrQxhFRFHIOM3QtgLro5RPUTrKgKPduFixhdklDlx3EFMULuFa+ZxYW2U0ztFn13nggQexecmuQwfwMARhRJFb7ISf7UlFuzHFcNjHr9XYqjfY3Bjy2OEjhI2QdtSiGGpGeeY0IEEdX0JuDKYwnDp1lmYDyrxghpDZ+TZ4dWwQM7fnAPE4QegtsjQFLdACwOBJdigfrXaD9Q1HVDHWYo0mGY1YP7tCf9ijKjOi9gxB3KFKunQ311nZ2EBXGT4eAieOarU7KCkJgpBKOxDr+SGe55EkKaYq8CqfosgRfkCRpWijkUIQRCEajZDgqwg/8Oj2u8wvzNNc3wAlyQZ9wCCVR14UCAHNMERICMMITwmWT57CiBCJJMsyRuMx83PzWFNRljmzc7MopTh9+iRKBSztmmHY79Hd2KBMxxgDYRRTiyKMNgwHfdI0BQzmpXBceoXqC+25qiyqvHSpY29fYv2WKYSxBKOK+af7F+gj/MwwtVJ8AZ9WgW1E5wagSoKn8IAbH8konikZdRRbu0M2dgfkNXWOVbU90+5JHvnmA4zuc3oIYSYD9VdqMHdeB2KwO6a3p044Kpk6MSZIqi8IyLY7OVvXNakCyeptHcazEVnLPUi3j90K0IlleH9BsaydwhiuiG+268WelvRwxeB+SXSdIrpesXJvwerdBVYq2D8F2qKOpXgPDal3S/ZKxZwQXNDP2BZqG8thramweDsia9fhMMZyanVImpdct7tD4KtzdLSrHYwQCG2RJzOCj2wyurVGsb+J9yc9xFZJpxGya67Bc6d7WGt3uhZXPl/ntnHZ5QTIyiIyg1AGdSJFAxvSh7ozFSmzkmzikPbnGWRsAwMl3eQcQBwE+EohELRrdXxPIYVgvtXBUwpPKpamppFCEPk+i50pBIJ6GDHTbAJQC0LatfrOv+PA0fxC35/kR1l+4Jd/ng8/9vlX4Khf2fI9xde+9Q0c3LXAvbfeyMOHjxKHIU+dOMXqZo/+aMx3vO8ruO3gPh559hhr3f7zWn+j06IsSoosxxrDuCgYZjntONpZRklBP80QMuTGPXc6UfXkp1SZkv5ok+WN488bcEihmO0ssjS9/5qABlgqU1GUGWvdM2wOVi+7VOhFl7hKvVwlhGTXzH6GSY+suNQ6/eWqawYaaVqgrGAqjoixmNinUa8hDS7wLgiQyt1EK60JPB8TBWgLQlisdrNOUgpiXyEbFjXRE4imh3StDuTEZVAASjh7WCEEkgqBh5LbDFuJtDk9o6nZACVhptEkTzTpeINIetTqEYEfMBqPyQtN6AVMtztEUcSJU0OqyhCG/gQoCaJaSFVqQl/RbDTRVjMep1TGQwlDEPjkacap46tUWpONS2ZmphEqQ2hNEEWUa+sIYZlpt5mZm2V9dZP1jS3G4zH1Rsu5AFhLf5xQ6oJWvUZ/NEIKaLfqJEP35debLYoiQ1jNVMNpPKLQJ5SCvMipRT61eoMsLwjrDc6cXWa926W3tUkrVhyYX8SXgtIYbFHw0KNPsHtpiTOfeQAVBMRRRKc1i+e5QDtTwcZQMxhU4GdsbmXkRU6WVxhVZ6a9l07LozteQyio13zyykNqgY0EYSg5sBSjpEWZBjYr8aXH5voWeTokTVOkEky32ghrKQV4SpFlOXlREtZC2p0Wq6trjNOUdhxSiyJOHjtKdzwgS0fEtQZeCX6ao0SJF4ckw5Qs16TVGD8M8GTA7NIigScZdjcJAh8lpcsX0CWomGa9wXgwcgBECLJ0hNAag8ELQ3RZoAJBGIYM+33OHD9K55Y7sFVJlk+oadJ3VL8wcl0+KWi2WmijieOIWqNFiWV+ZobxeIyVAb1enz17D1AVOWmWYirNrqVdbKxvkIyGTE9PI20FokApgRBgcCYLca1GXK+TJeNJPs2XZrXXHbUkyAxzJzMujk2o9TXNrfKyXYILF61d8D+pLGau9oJm14W1xENNPNS012H3symjjsex19RZORjvWO1uA2LtS468Y5FkJuTW3z2Fn+nn6V968QBSXPTaxf+/zEdw4KKKFePZiOXXTHH69TOUsYfUFj+pmD4+YurEmOnjo52PV7Fi5dYOVjoQMViKGc1HlPFFj4Ltw9mmjglBfKPH3F+N2fyfGcnjFUpYXre7pO5bSgNHtxT6PFvbXiYotNNcXK6udrq2naiKZc3G/0iRkaB2u0d0vUfrzT6q4ZLbFTAnJbPTIbMJhNVFVKrJYD2rK07cWSPdETZY1HMpaiUnLzR//MAJDp/soo1laabOG25fYv9SC3UlCtZFOyusRZ3OMJkh0xbRq3Y2f+ehOZ492eXMJL8iDDymmuEFqygrQ3d4YadSIGjUAmqR+272LbRYmq2zf6mNv205fDEYuqjOBytS/tkAHaHn86YbbibwPALPY7Ez7TpZYcR8qw1Aq1ajU3PC2mYU0YxiEFAPY0LPc3QS3598vwJPygtnxV8EQDMTQPnnsbQ2nFpbZ/fcDIHvcf2eJRpxxOJMhweefJbeeMxGb0Do+8y0W88baHzzD38fb3zfO7nvf3yQ//bP/y2VdvrgS/bDGGbaC3jKP+cACQQyZLa9RBzWOXzq0Wt2ZRJCsnfhEHPtpef13QZS4auAPfMRST4inXQnzy/fDy/zyZevfC9kz9x1HF1+EvNFolBdM9C4bs9BZmOIygJvqo1Xi1F+CNJHhiFxo4kMApQfsPnMY6SbaxTG0NMVSamRk29aAZG1SOE0BtYIEMbN+k8aTHby0LYTJxcxsXD1sGjjBmACA0ZgK40pS1Auk6HKc4q8JKo1SDNLfzjE8wM2e0OUHDI7NYNSPotzgtnpklJrOq050jRDUNJsSsKohpBjPKVo1DJm56bodkd4StJo1QnDiG53gC5LqtJSJjmhHzPoJURBne5mn/EgQQWK3nCEsK6T02rHJOOUzW4fk1uSoY9JNbGSeKFPkhR4AioLUlg8AdpaslJTWkMNifAV0kZ4cd3lOpQVlbFkZclGdwPpKXwd8cjRY8ggpBNLwiii02nT29pkanaORqMJSqLB2dciMdqwtTVA24pQCtrtOlkgGacZm6OMYVKSDsdgSqQfoqTE9zRZOiZUiqaqc3DXXnrdLY6fXOY1t385zx5dplmv02pEHDt+hLnZOXQ+Zqnu0Vqa48zKAON7ZEVBJD1SnRGHIbNTU7z7K97Byso6Tx9+loaccrkjeYXvW9Al2uRIL0BhWZibIZCCta0tGq06w36XxcV5du/eRW+rS71eJx2P8KQkDEJq9SZnT5/FGEORp4zSzNnsGkugPGxZUZkKqQI8IUhGY1qzi0xPneb06ibS81wKvO+jhAPEWlcuG8Tz6Ha7hJ7PuLdJtHeJdrvO2soaVnmMkz7z07MMtzYIQo991x1kfnqOhx95gOlOg11z8wSexlhDkRXYyuAFCuUriqxk0O8Txl/cG9NLWW/4XUdYF1yF/nRZcfLF9RJ2dS7z4Gj0Km7/5IDdh1OOvLZBdynAbFuO4sDGmbumUYXh1g+eRlbmsvt0xb00IeesQyxUs9hqChEeA6EBAzIjIURhiURBGXlYCVsHGpSxx+Z1DQa7a4xnI6x0wENY55KVN33O3jHFyu2diavU9rHijuPCly4925c7/RbCvYqF744Z/fKYOwYZN87qncmfOxcvDLfrpYL1seTUQLJpLBVO11DkktHQ+4LY7PwHuskso8+VjB6sGH2mZOF7YqKW4PbPDFg8lp27li6zQiMFT7+xyeZ18QXfh76pjnouofzEBifODqi0QQDLGyN+75NHOLi7zVtes4eZVnTVq+2cENUiN4pz+zHZF9+T7F1skhWaN9y+xPxUjalWdMEpLkrN1jC74DVroVkLqMf+eat8/gPYmXbM/sUW1+/p4Kkvzszpy1lTjQY/8+3fw1SjuXO+Xk0DewHsnpoh8n20sVROaHrVTtaflTLWsrLZ42MPPca73/BaPv/MEV5zw0E2egOMdcyJ+x56lLtuOMh06xodls6rz//xn/COb/pL3Pqm16E8R3PNpttsBD61vICyotSaQVZQj1qXvS6EcJSqxem9nNk4dk3bXZjaw1x78QVdZ0IIfBWwb/4QJ1afJStSzn8yhH74Rb1+HYVqlrnOLla7p78o27xmoNGYbjFaO4WvPKLZ/ex+/etZ2dgiSVParRba90BKosjHRgHlJECuISW1QGKsnrBgLY4cJCYzXQKDu8FrxI6FhvtNGgcohJt9K5FYAYpJ4JQQZFXBoPDwlUdTGqJSI33pBsSlwZqCIknIAkuSDzmxPkDihJ5h6BNEPl6gCKKIIpeEQQ0hFVOzDTAwH3hgJcFCx4Xk+R6hL2nWOigVYIyG3Q70CA/qcYAux1jhBtBgsaVhOE6xecHBffs4s7pKfzDAE4pYAoEkLw2BD8ZKpHBi7kbcxCKI220EFokljOqUxuLJgDD08aMafhBRC0NnzQoIY0jGY/BDWpGi3ZlmemaGcbtJrV6n3ZpCSIlUHlEU4gc+tVrIa1+ziFIlfhAyHLWodIWtJOtrW0Q1RdKIGHRLEB433XAdVaHJtCBQHtIT9EYFYVin1mrwG7/3B8wv7iPyKu5+/T30e5ucPn6KWiPgjkN3kuIRBDl4Hu1mnalOm8BXSJOzurbOyvIGn3/0YQaJAzKR56ELTRAoWq0mVRVgtKWGYHaqwezsPP1hj3SUoKRi0O0SzM26pHCtuenGG1nd3AJtKNOMvMgQwmkxfKmoTEVWGmo1SV7mbG518cMRzXoDoUacPHGEVqNFsN5HBjH1RkiWpFghkRNtifJ8ms0GaZpSVpbAD0iShHziFBUEHt3ukKmpeaJGg1qtRlmU+J5k394D1FtNNpdXKUtNEHkI33lph76PNWB0RhhF1OLaFX6lr/6SF0zcv3oGB5eUEAhjmT5b0F7vsnYg4uhr6oym3C1TTO5dp18/S9YKOPTHm7SPR5dBFgKbHwLd4tybAlvsB7PtMmLBhmB9rJzMeMkEEZzgEX+J3/ePc+/+RzFvD6kfaFHWPKwnLxCzT2Zo3BYufvni03wRwDMllKvmsqgoebKk3LQXfFWmtKgzBq99/kfsJGTqXE3XDNM1w6FZeEhXJBPXiI3VkNHwmh89wLmBpLWW/ITm1L8Yc9tixczs9sDePVkue0UJSFreJc6w1pfoWxqUZYX9HQkZOx2CShuePdklzSr2LbY4uKvNiZUBo+ScY4sQgtfcMMdMOz63ocuyqwT7FlscXx7gKeeEs9G71KJaCcHR5T7JxOa9EQcc2NUiyUvi0KNZC3boVzvHwLn+1zYX/anjW2Tncd+lFCglqUU+M+2rg6YvherU6oST7KlXE8DYLiEEP/Cu9/FXv+wdjPOMfuKs08/2ttDG8OnnnuaDn3/gld7Nl62ePHaSZr3GDXt3ceTMCo04ZpAkHFte5cTKGqtbPf7oc486O/fnWQ9/7E/5O1/2dXTmZjDaYGKfE+97J+OlBaL+gOVnj+NXFQd8j/SDj1L0L9/9F4idnIovVJ3GDLtm978oepMQgmZtipv3vZZnTj1Cmp/T7zjjoC9uCSFYmtnPKB1Q6YIwqDFK+hj7/L+Ta6lrvtsHQcTsoVspBgO6ZUny3HMko4SqyBhvriEFKGOpsjHp1iplblz2hgWDxhomnGpnCymF3Jn0EbhsCOnOAAiLEMplPwuBlAKJdA5CUiCkRAmJkB55VaKKgrARs7p8hiwrCZQLtJttd2gEPjUh2H/rrXR27UIIw+rmFp9+6PP0hxDHPr4nEWiHPEPjxOWeJA5rhEEd34sADUIRRyGtRgNtK3zPd1w9Y8nzHOlJfGkIfIP0fJDQjnyiSJGMSo6cOMWBQ7vpb867ZG2jOXV6HaIQk+Z4SpIVhsoCFvKsYDDOmZmepR77aK2dc1LUxI87zDSbhKqiWYvQlcVnTFZVzM92UEGdpBBoK1lcmCeu1/CVcjNa1p3vMAgY9LuAoVZTGC+hFkgXaCebDJOEQVEQhh7jwYjQD9m1tIf1zQG75n2mmzOE9VlmZpYIopjRMGO0tUm9UaccF3jVmN5wyIOf+zRnV1eJgoDZuTne9bVfz2Bri2eeeYqZuRmKHMLIZzweoRYW6DTqSKHY6ncptWFcOvBTjxtkoxHtZhOFAanQVcnZ5XVOHj9DVZXOvSnwSUeW4/3jaGsYJWPiwMOPaijPPXST0ZAkr5CeJKo30MYSRhUSw9TsAmVVUFSG4XBIHEecPb3MoYOHmJ+fY5wm6LzAlxJhDUpCmgzZPb+X3Qt7kFXB2WKD1swesjIhGydkWYonxwyHQwb9AQhIR31uv+Nm5qbmqMUxZVEgfUnRr1CVYTgYYms1PE+hpBMhSqEYj7943Mo/17XN8a4sS8+lzJzJGbc9N+uxfj2cuRPEZHCfSvRGcydQ88KSlxntc6H4Y7tMA7BY3YBqgQPlOs8u/xqPfG5M+Iceswc7zDfn2N3e9dLZyRaW/MzlgQb6SjOxgo+s++xtK26Y1Sw2DI3Aci5L0k6EHT5yNMv1aZ11a8gqw5GVLaDYWc/zqe1jNiPLE0ckZ8+GzDcMdy1VNAOLr86BuYuP8dIjcGV2R+Bd1OWZcLdOrw05sz7iM48vY+yl52Krn/I1bzvkNB1XKGMsn3n8LCtbY/7Xnxy96hGfvw0hBPc/dgYL1COfdiOkUfM5sOQmZfYvtna2K4C81HzuyRUePrxGVZmdY1jvJqx3E544usF0K7pge9/2T66yM6/SOra2ynf+wr/mDdffxJfdeAs37drDdL3xioMOay3aGIy1NOOYuQmN6+LSxrxiQON8wP5y1eFTTmj86HPHMBY++/SzBEHAOHPuovXpKT577CQWy67r9xPGERtnVhheA43KWkt/Y4v+xhYAZVHw6Q/dx+u+8i3M793N/NvuRUjJ17/jTWSnfp7Dn3yMcpKH40LvHEPGYhmng6ts6VzNtBaR4sWH5Aoh8JTPvvlDHD375M5++d4XH2iAM1g6tPs2FwgoFWfWj7GydfLl2da1LtienuPmQzdw9uyzrB4/xvrnP4OtSsCQTZ5QBpwT1KRFfm6mxSInol9x3petxDYvUiIlYB3FygEOhx4FxoENrAtqExYP8KRLVY6E4uzmkNVBxiBL8YxBWOcKYTFo6/ZgNM5hUCE9KEufmdYSZVWhlKLTriOsoTfYIvBqYKHTnEZO0rs9zwdhScYJvc0+wlQ06jXmmxGhgrTQ+PWI6ZkOYRSR5xlTnSmSNGetP6Y7yshSw9T0HuZn9lL3F+h06kBFqJ5j7/7dLJ9d4brr97A4u5sza1ucXjlLf2OT4fomN910iCAK2eiOGRUFSaEpkjFZahH1iKdOHibwBLsWdxHVatQ6ewFFazZE64pmHCCVRQhNWeVgNJXWrKz22VhZpqoMaTqksAXhVERuUpf4jiZQmnY7xiIZjUqSLEUby4lTGzydncD3JNNz83TaHeKwweLiLpYWlwiDkKwo2NzqsmfPft5wd0BvMCTJEiJPUZ/r0IpuIS0VZ06dYLo9C1WOatR57T138zu//htYAXEY0s9zBt0uaZIyPTtFkad02tMIFHEo8f3YheiNBqRZ5sCftRgLyvMAxeGjJ7jxpgO8/tY7ycYFT3ge3d4QPwwJwhDfV6RpRVFqvCpn1+6DFLaCSjPodxkO+5w+c4q5hTn27duLkM6CuSycvW6a1Nm7uMDs9DwnTh+n3a6jjcEzktALUKEkGQ+ZmZ3H8wLiWoNAam64/jriIGRDSk6dOcP62jpZWeKVOb6waBOQpUOiMEbrkka7TpZ/8X2w/1zX5OEcpIYgLRxAOOPDkRkuGZ1faaxzpan2KywsJoIQT8FMrEhGhnJYUVtuMTO1QL5mrupk8kLqSgO1K71uLDy6MuQjz64TKks7Miw1DQbLwWAfN/g3YLt7oIxBe4RAgOU2scmT4bOMvRcePCkEGCvYGAs2xpIjm4pWaNnVMoTKcvOcphm6Tsyxdclz/z6j/jXQfEMwafPwhTHOeR2ac8ys87oJ1rK8Meap45u85ob5q65q28Bh2wr86ps9jzI22ddhUjCcdFOePr6FEIJdsw2nJVlsMc5KBzKeXcMYe4E71uRSIsnKnW7Jl3LlVclnjxzms0cO80v3/SE/+O738/1f8dWvONAAeOj4EQZpwm179rPUuXQAaa0lqEUcuO3GnfuK5/tMLcyeA9Fas3l2jRtfdwdHHn2SqqyYPy/5GsAaQ3d1g+qi9OuqKOlvbPG2v/xeonqNub1LyIkQPohCppfmOfzgYzz40U8yv3eJZDBia2WdPM3orW0S1iLyJOWme16D8jzSUcKw23PbW9tgemHugk60o0IWLD93/BLwUptq890/+ffZdd0+Gp0W1lqMscT1GnmaMeoNKPKcUbfPQx/9FB/51d90ro7Po8q84Bf/wT8jqrljvf7OW7n9Lfdw6K5beeff/lpe93Vfxslnj/DAR+7j8AOPcudb30hnaobVU2cI84jv//F/wpnnjvHE/Q9y+pmj5Jex3r3Y1erFlOtsdDiweBNHzjyBxaLk8+vuvlQlhCDw3cSD1tXLGuh3zUe4urZCLawTKIE/GjATS4S4UC1vsQgkxrp4dK0N29eNEJPWupzYUwqJUBIFSByAQHiT97YpVgLXxHAKDyklUko8qVyXQ0iSssCWCXkBwggKFEIYhNUkWUboeagwxIQx4zRDKoEfxNx00034vkdZgVAWU2mCsM7s7AzdXg+pFEYbCj3RMgiBCmKyUpNrAaXgyOoQKSxB6JFlJZ3M3QCSLEOaM7TbLSpjQUjOnl1j765drHULJB6x8RAqopQNVDRDFBfUm3spjc/8/CKzM/OMkpTVU2e4667Xom3laEDK4xOf+BTHtjY5cGA/K2tHmZ6qEUYNNnp9FmZn2OitMDczhy5ShkmPbr/AmBJdlRijKbV7EElPUYmCzOSo2FG8+kmO51eEYYhVHs12iMEwNdUmH1jOrg+YnWqhpEHKgqzKGQ4Tjp8ekI8zosOP02q06TTa7Nm9m91LC/hewNz0NL7n8czRo/SGKesbawRC0Wy3aHdaHD96mPWzG+zev5tnn3ic5fV1fAWWkrjm4/kCXVrKUtPvD4kDn7mpBWr1GtMz05w5dYIyGyKjiLjWIM8LfGuQUkx4snDm1Cr2dZrrD+whiAIarRpGCKoyQ1eGZDRGeoo8T/C90AGysnT5JtbS7XdJ8zHNZof5+TlazRb1uE6vv4WQkpNrPc5sJoT1JsIPMdo4W1phKfOcwbBLrdagzEuMEAgreebZYzRqPjffeIiD+/bz27/zm5RGMjMzw8z8DEobhPTBGOq1mDiOL/lt/kV9kWrnYXON4OLFloWZOOT//Io7OTFe54HTOZXdi7HiPF3AK1cWy9nBKiuDVSpTXSg8FmsoPn9FMFQP69yx61ZqQe1CCtjzOKbzFy20YH3sNCEAT60rpmuWhbrhyXXFKDOk/z1D9y3tdwQI/9yHk+MCayOEJ0DnF3jMX7A3F+2bEIKi1Bw/O+CO6+euKrZ2XXMuoT5d24FyKSnMWs6sD/ndT4xZmK4xSkv6o3xnv65pHX8G6n133cP3fvl7kNci2n+ZSwjBvdff+AWXW7MF/+yD/5l6u7nzOeWfG4rlacZ/+sc/zd/4qX/I7/78f2HXdft4w/u+4pL16LK6ZHB/5JEn+W///N/yff/yx/CCyzsjnT16grvf9Va+9R/8AEYb8jQlGYz4H//yF7jry9/EmedOcNc73sTUwiwIUMqjt77J7/2HX+F7f+ofOjeviXvab/6b/8Sv/6tfvGyHpL++xad+8w+44fV30F3dYOP0WfqbXUa9AaNunzxNqYqSqtIX3gOeb1nIxgmnnj7CqaePcN//+D38IGB61zwHbr2RW9/4Ot73fd/KW77x3Tzy8U8zdXCGv/R3v41f+6mf5z1//Zvxw4AizTjz3HEe+MOP8+H/8ht0VzeI6jWG3d4L368rlBCCVn2aVn2aQdJ9xYDG+SWlYvfcQepRk5WtU+TlC58Eulxd8xEanXP06JNECpplReirSWtWTWaXrOseCMAwybdw4MBdhBOgIRwXVSnP3bd1NdF+TpzMz+t8WAtlVSGwRIHPhEAFTECNNXjCIpRCWIuPRlnQUgEajATlo5oxnk3wUAjVxFpDlhUY7WazjRWEgUccN2k0pigrQ6fTBmspitL9qIT7YSMlWZpgpcL33YyFNhqtIS1dR8eg8KOQU6trZHmCrzzWV9bw/ID1fp9aGNEbhKA81ro9vNNn2VhbQUV1fKWQSqAkZKOEldUVmiePTcAYlEXJrrkWyajF2topjFHMdnZx5sxJkqQkViUnTj5DLT7GoetuYuXsJkVeMDPVJI4ivEBiq4qZqSmiQFBiyAuNNiVClyijSUuD1Za8Kgk8qDyLND7elOT6dkyVlhhVYUREvSqJopD8zAY33XUQKSFNE6TKOL12mKoQ+MpjtdsiG6YMUkP+qU9jTMmuhTlMlVJqD6FqNKba7N9/ACXh2SNH2BpsME5SpqcaLldiUCAF9La6WK3xPJ/pdoejR49x5LnnKMsCXRUsLi7SbHZ47sgR8iIjimPq9ZjhaMw4GXLsyJgsL1Cej6ckhTZIT1Jv1IjrDfwgwBrQuQN3CLDGUuQZ1lZUVUVVZWRTc8ShTz32CIKIMGqgVABSoI2hLCt0VWG0Jh2PadQj4jhC4LHV26IoDNpoNgcjHnrsaWabbbKsYHFpmje//i6qqmR+YYY8K1lZX2V+dh5PyEtmsf6i/uyWzRtMr7yN6UOf4jW7Ck71Sz532mN1JCeyjJdu8CgusgCzX2BgKoTkpvkbOThzgI3RJse3TpCU6aSTbXEpNQASUZuF8x6oQ+Dp8ZAb64sIBJ4pCKvk3IDjBYkud3acYS4Z5pYT3UkHXULVM2z+ZobJLfXXOo7/eNXjs/fdjX3N2wnRmI2noEqxukCvPeIELLrEphsTDcyFQEEIOLbc5/DJLrccnMbYS3GnlIK3vXYP/XHOVj97YWDjMge7rSU5s+743tvP0T9PNd1oEniv/EDtWstYy/Vhk8ZyD39+5rLfVxCGhHGELkuOP3GY133FWy4LpGR4acdk/fRZDtx20wXAxVpLmReMB0OwcOSRp9hcXuUX//4/Y+3kGTaWV+mtbbK1ssaHf+U3EELyq//0Z4nqNefO1XAsD6M166fPAtCamaLebjLY6GIv4/q0XZ/90H189kP3vYAz9eKqLApWj59m9fhpPvP7f4zyPOJGjdbsNJ/9g/v4w//yG9zzVW/HDwOEEIS1mOvuvIWDd9zMV33HN/Kp3/oQ977nHayePMPv/h+/QtF9aVkEAsF0c55xNsC7Rq3Iy1lCCAIvZK6zi05zlpWtU6xunealMl255iP0g9ClImMwQqKEh5n0ZIWFCuuoyNpgjNNk2G3xt7hwlqXQFUWWoaSiFjjnoG1wsUNbtoAUDLMMaSxh2Nn5Tbr1Ob2HUqCExQoF1iKso2FpFP2iIN/aYA6LN0rxdx1ENgKMdv4nJRKEwPM9ijwjSfI5klgAAJuXSURBVAdkWZ2yKh3tRXp4HkRRQKsREQYBZZ6xvjVgfXMTYxRpnmOqkiiMkWlKEPiUVcXa2hpHjx7BGGg2m1g0aZJhxmPyuMbqes7m5iZZXtAfDOl1N1he3QAhCDyPWuiB1IyGAzZ7WyjheHRZXoCA/mBAGNbQWE6ffRqtNVlWsrY5dudRjDlx+gGsMQghOX5mgzAM8aXFaMPU1IBGPaRRaxCFHq16nbjR4OFHnmCzu0kQeIS+h5IeMvCYX5hjvt2CKkPWDWc3VhkkQzxPk4wS9u2eIfQ8alFAPQpJ0pyCwmlZwoiwNiQdpRy6fg+H9i5x/OwGa+sbJFnCbLuDQNJstRgMhjTqTTqdziRBPSHNUzzP59B1eyiznCPjE/QGXYrjJU/mzyKlJK7ViWux67Z4Lozvuuv2sLa2SaczzfLKaapKs7bepRM3GAxG+KGPIqTSmiiOiD0n5tVVDkiE0CA0vhdMqH2CwA+o1Z3gey0/zeLiAsvLG9QbDQ4cOMCtB/czN7sAwiOu15BCMx44Tcag1yMrCwSK9c11hsOEbr/PcNhkY2OZ504cRUvN7OwU1x/ah68CTp44w2g0pB41KLMcL46R1+Th/Rf1JV8Cp/lYP0RV1Ojf9BFqrZQ33lSxPpZoI1jvKrQWCM9S1iuy9gu1PhZM3Twkmi4pS59njtzKE0/dhbHXNlMsgKjK0fn4Mh0eiWztBRVw/oMrQfDItjOTzlkaPEcj32JKr0FlnCPhRdu4pkl5sb3YpR0IU1i2fjun+8EcLTyOdu5itb4b4dxPkXvfApO99K57t/tglWGGp9ErD6GXP4OtUjcpNlm91obl9SHX72nje4qLH87aKtL6Dbz2LTfyyDMrbJ56AnQButjpnlytc3LVQ/1zBiwuLfvSALcvUikpef+ddzNczTh1hWWEgFGvT299i97aBtOLc9e0bmstp589xnV33nLB+Tjx1LP8h7/3E5x48lkslmQwoswLHvzoJy9ZR1Wco88UmbuXbOshgB2g8aVWuqpcJ6XndBlbK2vkiQv/feNf+goO3n4TcaOOEIKp+VmmF+eRSnH7l93Do3c8wOFPPPqS7s82bUlJ70WJzF/q2gYcu2cPkuZjBuOtL/yha6hrBhrNVp3+1oB2q4WfDqAs3aDMWEoMveGYQpe0opqbYRD2vHaYE6WJyax8oTVpmhL4AfXAn6Qri4mWQzpPKuGoUWgzWY9BSpfs7MCNBSVQSiFQTJQeGAESF2hnsCQFbI4ygrkOQVoRyhG+7zIQtCkok3ynpVyWBXlVQVU4sCEkZVkwHI9YX3eBblJBFMZgLYES4Cv6aYL1fdLxiDxTCCnxVMDB/fudYD3P6Pe3GKdjjC4oiwyhfAaDIVmS0OsPKIsUqbaIA8XSnr14YYSunGhpVI7QRqDwGSYlWhvyJCHIKqxUk06RxPMDLJrA97ZPOdoYEJaqKMkr7TLKPI+qN6Q3zlHeGCUFnms1gTHUmg2slYzSlKpMKYqCPClIu3VsEBAENWq1efa3d3Pm7Bl63TWMDSmLkq5OabVi911qkMpHCkOeVnS7faZmO6xsHoUyJapF5GnCUIKSIcNRQlxv0upI4rhJmWkacY3AD9GmoMoqxllCEHho47plQRgihaKqKrxYsdnbYjwuqKoKpQTX7dtDI2qysnKGRq3O8uljHM1glCSYQUFQa1KUJYU2+GFAIBXSU+iimugvCgqR0WhMuWuzLDFVRavRZGtzi1OnT6LTlI2VZTbWNlg/u0azWSeudwijGGMrPKWIopjB1hbaFERRhFCSvCgpy4ogqpFnmm5vgMFnfn4BX4VEoc/NN9/I+voGR448i++FzM7PMj019ZL8+P+iXqK6wEnr5Vi/YLm+wloWUOiK0s+grhFYOk13T/UWM+RSBnsShGeRNc1OSMm16BHOq+WVPRx/6mbEkkcyalGWbub0gu7JFdanmlc9kEs+uC2XyL0ax6fvRBjDnqnj3Nz8PA2vj14xZJs+5dijHKsJ/ejCjsLzqW2BtKkkJ6duYaV+3XnA5KLDk57bu6CBmrkZOX0j3u43UTz1a5jeUc6PBXz0aI/mwTdx96wbiJXGZ2ianCj3crZa5HS5G42CQ5Zobx9rNWbzKWwxwmw9i03XAYFNN10XZfv8XAOl5GKBr7jcAf0ZrbVB/5XehedV1lr+4JEH+dgnl/mrX/4T+OGlVuVCCKJazMqxkwghqDWvzQbWWsv6qWW+7Gu+6oLXG+0Wb/+m99Fb2+LTv/9HnHn2GNOLc6SjMdk4RVcVemec9een1k6e4Td+5v/H7/z7/8ziwb3c9Y438ab3v4sb7rqN/sYWK8dOMbtrcaebdP5v8aUCtp7yJyL1V1dJIdk9c4Bx2kebF+9Ede1Ao97Ekx7JcIiv1CTkbjILhKUq3aCs8gN838PaSQ6GNVgMWIlEgjQIoVzHwxgQEoN1XY0JtWpiIOl0EUpgK2eFK5BI4QbLCIEQCl95WCHBOOGkMRYrBEp4RL7CWoiFRShBjZLhuER5HkrJidrex1POLtcUJYPeFmWaUZQVURxihaTIHRdS6wrP9+n1x2hd0uv3iKIaZVGQYNBa46kI5SmMNQjl0Wq1KIsIXeUsLSzi+a7zUlnNzOw0va1Nsqqiu75Fo9VAa43VhnFv0z1UhUYWgnRUUqv7pKMxw2FCVSSOglZv0ghDF/AWRNSjOq1mHSkVZVWSJAnGOq2MEC60rigNZZWiigI3cw874A0BUuEJR38wQoKKWR1kbAwyDBZjDVKCLxVYwShVJOMetXqNJK8oj25hhbNzbTea+KoFOmMwKMhyS1lAVkJZjhmnY0bpkFD6nD3bx1cw6G9y9uwaqxtrtNodFucXKIoU5SmSMiGKnS1oGIa0Gk2CIGAwHLG6vk4YeS68rTBkRc6Tzx1hqjON5wfkecZwnNOMWySjEX4UoqsSXyny0YhRrwKUS68XAi8I0cZRArWtKLOKld6A+aVFPG8WgDSrKJOKKs/Jx6cZrZwm0ZZxXmKFQFcG4XnUozqzc7NgStI0JUkLhHDCWU96lGlGWY2oxxGbG1ucXj6L1Y5+1WjUWVjczbGjxzh58pPs3r2HG26/+0X/+P+iXmgJmDmO7l2HHc0z3j0gnU2I1+vUlxvuJ1RKpJY7i7/Y2nP2LvacvYssHJCFQ9KoR6992s3nCBhmq+gzOTwIhBViJj3XGQZkTePtTnf2RUiLmi0Q6qLBhW9YWjzNt3z9fwYL65sLpFkNYyWHn7uFNKuxsraL/rDzvAHMlRa+4FUpONW7juXhfhqNAXMLq9z4pqeYVRvUzBDfL+kdrjM4UWN82gkZTSUw1ZUf1heMCSaDhVHQYb2x32WPXOM+CyEQU9cjm3sx3SOTR5TjqhskD2d3sjK8C4EhNTFbegqDnDy7JhbBQkDUdvrDPZPuyYF3gXU5LGZwCqrM0XWzLnr9savvnfRRi6/DDE9jBqexWRc7PutoXvo8K96rruVLt9KieP6X4StYAsHceovGY8foPrvM3G0HLh20CoHn+zzz4KNMLcxdUWtxcVVFSTZOmdm1cMHrs7sXee93fYsbw1QVB2+/ie/5Z3+fPMkYdnuMugP6m10e/cRn+L1f+BWWDu7jfd/7bUgpGWx2Gfb6FFnOJ37j96nKEmsMRrtO3J+FbJAyL3b0Hb//H3+Nud2LVEXF1/7AX+Pme++iPt3ECzz8+YCsTNg4tkormHpRblHWWooym3Q0Xn1XrxCCetxitr30kmRtXDPQkDJidjri5HCAVTvRejtiPyHcjVJPAIixZqfVvX0rNzjLPSfkdpQri/PWlwiE9NgWeWyffCUlWmoX7iQnWbD2XGjftmLDdUskLo1DIoQk8hRCOSpPmI+JGgLZaDLONGVp3M1dFPi+QAgfbTTGVHhBgOdJrC4J4xq+9EnzFIyLt7c71rwKbSriOCaMAqzWYAWlKZFCYQwkyZgizykrQ6YNofSotEFXlTt70icMApody4GDe1FIstzNyBdlRjbqE9ZatNo+ShqmZtugLctnlykri+cH9PpdslGOtpbZzgxlNib0PTylUNZZ76ogIPACVjc2sELQbIRIZVEyRHkKzwOrBVnueKR5WVFVzqrPGgdCHBB0341SrpNijMVaqDQMu31CP0D4AVjLMM0oqgFrW30CXzEcFKwvD+iul1RFjrGGtc1NanFInuesbm6ytrVJLYyYmVvkq9/31fjC4+lnD7O21UdKwXgwoKgqiqpC6x5nWKUqK3yl6A36tNttwsDH9xXtqVnqcY1ut0tVlYSB4tabr2f/whKnzpxikFeYqqLdaWNpk2YZWZZRVaUDEKMEUPiRh841x44fZ5BkjLKM8XCIrjTaWme3LH3CoIaKAnY1WiAFoS8Z9IdsbG0glFtuq9tFqpBuz9kKW2EIlEfkeQSej5EBx0+dpdZ8jjAKQRuUEtSiiFotIowlyegv7G1f0bJgwyFPffdn2VyoUTYKqrjES3yCUYCVlvbRKTrPTrP06d2E/Yml6At5nkw+I62boKilU9TSKWAfu1fudJ1LLIWfYGSFQFD4Cb3WaSovZ6tzHDtJ7LOPn3cIAqjpC/dJGRpzHh3hU9W6bBmL87opKfyU3bP3IYBes8PnDr+Bk9k+tpGMQTIyjQt0HS+s8+Du7cZI+v0p+v0pnjt+M1GU0mr0ufuuT3Prex9lP2sUA9e5TdcDhsdrIC2jkzHjsxGTOS7yrtNbnTtuwTCc5ejMXWRe/cq7cYV9u9qsb2EDTpULF3S4JrpZLjzRF1G6pGI7vFFNXX/Be96eL7umPVMLrwXAVim2GGPzHmbrGUBgR2cx/WOXbPcv6pWpt4s7ePvUHfzUT/8ub/y3333ZjoXWms/8/sd44/veec0D0XF/QGOqRa155et6a3Wdud1L1NstGp32BaBkdvcif/Cffo3Z3Yt87d/6ayjP2wESg60ezzzwKH/jp34UL/BJBmPA0l3d4LnPP86jn/wsU/Oz9De3wMJ4MCQbJSTDEcrzdsY7L7FR3kteuqxYOe4G1r/6kz/Hgx/5JG94zzt53z/+Vm59++uQnuLxTz3Ah/75r5P3MhCQ6xRTGjozM1hjqfISow1+HBA1a0QNZ+Cii4rumQ2sdXS/fa+/njt238OpTxxF2Vdep3FxCSFYmN5Ld7RBUV7qxvV86pqPLs9zOu1ppqenKNIubi4Wtq8dZ2ur0dV2aJQ4B0KQ2G2bWutuvFIqrNieFBOAwBqNcO0KJqtwtKrJg3I7EEoIicBRs7aBi7bbDzYJQiKFIPZDtBBIpSjjENFoMBgm5LmP8kPCKEQq4zQeVmGFxZqUeqONkgHjUYrngzEVEo1BY3RFpQ3GgFI+aTpyHZYJ7UtIN9PuKF1grcGWOc04xBQphS2oKuOE7mVBkWdkeU6al3S3enhSEddi6rUaNoGN0RphrJBKgQhoNGr4vkdeVkRhTGe6hbaWosjRukJXgmG/T7NRxxiD0ZooTVCBR1Fp5hYW0GWB8EYMhgPCKMS3UOQGbSWmDGnWpxDKh0jhBwFSuJApbQxV5To342QwsV501m/CgBAhZWkw2Rjle0gpyPICrQ1FpUirirVeQj6hBwgpEF5Ib5DiCfBEQJYrkqxka3CC0WjITGeGGw7uZ+/eRTa7IzCabrdPrz9ks7uGUgF5nlLkmka9jSddyF2e51TWoMuKSjsHtLDVQVufooKlpd0kJ07RGw0YjUdE9SZhEBD4PvV6BHhUZUVR5AjpdEdVVdKqx5gs58yZUwihkF4wSblXhF7AwPfZ6HapxRFhEBKGITPTc1gBeZ6ivAA/8NmzZwkspFlCnheUlSYtK0zSx0rDmTOnMFVBu9UhyzN8pQg8B4Z979XXav1zVRY29oWsX6/J4wTXDRRUtZKq5iil6WzCyr1nOPXO49zxi6+jc2TqBXQALqorDDgEgrA4F/4XZS1aw0WssOw/9YZzFKpr2YSVSCuwwrB43utp3CP3R4xrm+hdn+fN4ad5Y3guC0BbxVo1hzmPTnSi3Md6NcuWnpoAEEdqNVehYJ0HTS443DyPWMtiPvTHX8PnH7uHhbmzvPH1n2J6aoNoumDqJieINlpg9eT5oQXDkzG2EjsrHyUtfuuzX824aF5u89dWUrHtugPueSRbexEqvHYdyctSFuHFCC+C2ixq6pB71Wp4CegPr8YaZRnGGPd8/FIoDTIRSAS9Tx7hD3/51/m6v/2dl4CJmaV5Dj/4KN/4d7/nmle9eXaNM88e4/4P/hE3vv4OZnYtXCIiz5OUqcW5y4KXRqd1Sffk/OWUp9h70/XM7Vm6YJkHP7KLsBbz3T/5I1RlBdaSpRlrJ87wcz/4T/iun/hh8jSjyHI2llfc36dXKPKcU08f4ekHHnaZa6+yypOUh++7n0c+/mnas9Psv/UGbnvz3dz25rv5wL/+IZ657xGEFOy++wD/+vt/lG/7Jz/A3huu5/hDh+md2eD2r7qb1vwUfuQ6H0l/xG/8o/9Ib2WTZ554mHve/q3c/VVv42c//mOoyVD85aBnvZjyPTdB/UUDGv3uFrPTHVpTM2ysr6DNFg5lTChSQqGkxBo3IHUOHed3NbZP3LmwPmeDa0DKSXdjW9jlPiSYTEtZi7UaIc79CLYfXFKAQlBZs9PKRoCUlgrjUsSFRDTbHLz+Fji5yvHjJ2nGNefeUzLpsDgUOuil1MIWIrCTDowLZJO+T1WWeJ4iCEKKPMfaCl3maGFoNDyX4SCbSOlTlqVz2FLKARNdkOcZopR4nu+cDvwAXRSIqnJuWdoihCVJUhJStDVYo8mTIab08IOQ8WRSLUkztBH4Yx9fBaAFtVqbQCnQmlq9jlIKo0vSICSu19G6cl+GLTh19lkE4HshUVhjPB7RaM5iK0WaZSgJZZlS9ivKsgJbIQFPeVgEp5fPUlmDFRIPiZKOZhX4IZEvqUYlCFhYWCCu1+h3+9RqdaqiJK7XKcuCqizQRqN8hRA+lR0wPz9FEHj40iMtCtY31sizAY1mi1aoiJsz3HbbLYxHYx5+9FHqjTpxXMdTTuRflAXdfh8hDLuWFomjiNFoRKPeRBvL0uI8w0FKa6pNbaPL+uYGvUGPO/fvYW5qltOnz5JnBdYW+FFAPfSxQDpOmZmZZn5pH6tnT9Pv95DSUGRjMm2QxuL5LlRR4fQjYegCHctSOyG+LyeaGUGeF1itnbUfIJSgqjRSWFqdkGF/jfEw5fSZZaK4judNriNjuYqL5l/UF6MEdFZLps4WrB6MJve/c53b7b8slvHSkCe/4xHu+ak3449fxmAmcck/EAg3U/YCnuHCSs5l71ka41kazDLdO0BS22Jt7hkE1URZB56o2Buc12K3sD84ia5q6HyGxMbUaWBtRNc4Ktbny0XSoAQsQ+OT2HgyOSUus8+TrriRnF3dzfLKHvrDDu98yx8yP7uCEBZrnYYONZn9Csw5AGIUx05ex58+/g7Ssn7RmXo+ZfEPvhvTP4HpHUXW5vEOvgu16w1ugP+Klrjo78n/hMROshRqMuGG4AiB3KZV3fbF272XoZ5bXWZrPGL+CuF4r7YSJYjU3S/e+dq38pO/+5u85zu/mfi8LoSYTI56vs/Swb3XvO6zR0/y8Mfu5+H77mdqfpYf++8/z833vObC7QtBo9O67Oc938O7goPX1QB0MhxRazaQShFOPh/WYppTbd78/q+kOd3hrtfcurP8+YPp5z7/OH/vK76F8qJsKCElcaNGMhgR1WsAVGWJ0do5QX4RcYm1lt76Jr2Pb/LIxz+N8j0+8GM/xDf/b38DhKDMc8J6TFkVzO5fYGbv/Lmx6HnVCqf4qz/7t/m1n/r3zMklmtMdHvvUA3SWZsjOOjtZIQRLd+0DYzn58HN40uX+vFKgwxhDWb34fI1rBhoCTVnkxLUacatOcdbsdC2M3Z5scxka1lrHSd15YIhzfGHJuYvETjI1ALZzNiZ/pBQgQQqFRWOtcRl+1oEPa8UO2BBKgd4WjDvg42g9UA8VQimMClEyZm5ujq3ugKlWkwlfiLKsKLIUP4gY9fokWYlOChCC/mCEN9F0WKHI8sxFxgsoy4xGR7jWoBhhpSCvcspxhef5O0F/aV4SBjUskvFojKcUQeAhpKIWRZRZipSCqVYDYy1VpSmrcmIbJ6m0ZZwlTE35ZJMskLJyWpPhcEjgh2R5Qq1I8aWHFILe1hatVhOpPKSSlKXFaEtZFWRlQp6XgMbaHCU77N51kCiqOVH2YIiQgrhex/dcZ8KY0s1K5DkVkqU9u6mKlLzIybOScZKQlzlZqhmNE8qqoKpKhDHMz80x6neJ6h1qse86QEawMLuLLBuhvJBud0heFPQGA9qtOkKUzLYaSFmj3x+wmm+QZRXF6TMsbi0ghCJJc3JdUS8FcWQo85xCa4IwYjgcsrLeoxbXGfS7RLUcT0pWN7cQWjDOM5JkjC5ctsjczAJ753cReT5ZWbK+tsEoHTMcZlSVRltDUAtRCgLfxxiotBPRWavddS/BFJrcuNmf0mhMNcLzPVCCvKwQUuL7AUaXVEVJkmUIT+EpgSed3W7ke9SCCOoSPc5I05RqwreOopAweqUHNH9RfmG45U+HVIFkc1eAlZdy/d0g3JLNJgz29Zl5au6SB6SReofadNWyAmnVFZLHdzZ43vLXsMy1lthBTgjgupNfRh6M6LVPn+taG0VQ1qkn0/hlzEz3OgCCskY9mdnZ9Pnzq1+vNLYxxGBZ1ZLn6qtkcZeTKuKUFwAGI/UFOy2ALT1FYmJOnTzAf/uff507bv089772fuJ4fIkwP8tiHnrsXk4v7+PM2b1UlY/LdLq2Q7+czl/E0wS3f4Dy2d/Bv+69yM7B85Z+dc0C2Mnz1xclh4IjvDZ+lI7qnbfEt7wyO/YS1eZoyPH11S8doJEJBzYQ3LvvTm6JjrJ+5iz7bj50ybLNqTbTC1cPgrTWuUiuHDvJ/b/30R3t7NbKOp//o09x0913njMLMIY8yZhevPw6a80GcaNOZ372kk5IGEe0pqcu+7n+Zpda69Jkdikl7/rAN/DYpx5g3y2H8AM30XK+eUGWpJfM4u+75RBf8/0f4KZ77uJffs//zmve9kbe9ze+jSLNGGz1yMYJvbVNkuGIzeVV8iTdCR3sr29OQgAHaK3JkxRjzGUzR15o6bLit3/ul7nn3e/g4O034fk+nflZ8nQCFq4wE2it5aP/9TcRUvAV3/p1ZOMEay13feWbuP9X/giBwI8Cvv4ffAcnjzzHSv8U7/vOb+W5Tz/J6UeOMVobYPV2F/WLc58pyoxSv3hr3+dhbysYDwfU6w3CMAIrzqWdCpdgaq2e0JcsCMfddw9bs2PhJbaJUsIleFtjkGo7PMN9TkyWdERiAUJOZnInlKkdhQiOIhX6LrNgMjAHg4sOlIRhQGI1sswYr68yHieMkhFGlwRBgFROjBOFEVHDI01SrPTwPYHR1Q5lqCxLdyxCkJQDjEnQJqPeiKm006coT6AEWFPDGENVlRgjqcqMuBajLQRhRJ6mpOOEuFmnoqA/7ON5NcZJQuD7hL5PFPhoDMkgQArH08/yEn8iY6mqCmMsea5pNDwQNZSUSM/D6oqiKNjY2qRer2GsxWTOXSKMIhphRFFzSd6FzkizAe3WNFZLsBo/8MjSFGljirxwn7caXRqUFzrwEkVgmxjt3Lis1lTGoKTE2JKi0hRZRr/bx2IRyukmikkXoyortoymKHLKygngi3TEyVOnCXwPUxmXV2Ek2liCQGKMIYpD1jaHzC8sUKu3yMuMMAwJYt/R2pIEFQVYUxLXInSpQUgqDek4YZwOaTRbDIcj1tY3qazrNhw/eYozJ88gBRhhMRpqtSZhGNMf9imSBJtbxsmQIKohrJtZMQDGXY95qR2FzBYEXkyWpZSlReaFy9MwBqMUzTCiKivyiU6F0iCMIIgUfhBiUGx0+5TWoHWJVJJmPUBXYHVFNhq86B/+K1Xp5GYvgYDtSYZX1+DsC9bEuShMNa/9SJfD9zY5eUvtsmADoGgWPPUdj3LLL91G/WQNBGhVstU5Qa91hjTqcflPbpdFWMV09wCe8XcGv+3hElHWAgFBUUeac/QRKyyFP54M1F35VYRXhRes9wuf+4na/LwKiho3HHsHTx/6MKP6BlO9fexdfh21rINfutnHSwDRJZuxBNqHrT0AHAAODKbB7qP0U0ovJwv7dNtnWZl/ksrL3cQUlrGtUdoAC5wq97DyxCL/87m/Sqm8HdXe9iaLMmA4bl2AGLa1I+cO3V6aR7LTnLLUxZgKj8yeA/iisZvgtd9/0XG92OvYRd7u7Ot5+/Ci1i3gLbU/5ebwsOv8fIGr7UupKq05vbVxTUF5L3dZayl1ha+uLPCVqYDJT7K2obn5vbfTaF/aYZBKMrdniVrr6o5TeZrx737ox/n07/8xo27/gvce+9QDfNPfcwY24Gan03FCa6Zz1XVGtfiS+4JQEuVdnp6my8qF+12mppfmycZjfuZv/gP+yv/+/ey7xQGq3vomD37kk5x8+jn8INix1b3jrffyj/7rz9Gcdvv4gz/3E/z6//2LLO7fg39Rbsj5wMEBLEtVTnKrRglaa4abXaqqYvXEGX7pR/8Fqydc13XvTdex/9YbGWx26a5tkI7G5OOULEnRlcZo7Ux9jLksrau3vslPfcffZf+tNzCza4FTTz/HY5/4LLuu20+t1aDeahLEEZ7voZTTkv3p736Ypz/7MH/rZ36cj/7qb3H4ocf4wI/9ECuHTznJgLbUp5o05zo8/B/v58u/9f3c8zXv4J73v51kOGb58AkO3/8oh//kMbaeXUfw8lKorbVsDlYxX0zXKd936YnWQlirgaewZbnzENpurVnrnJ92wp8E7kXruhNMHgXCqnOzYVJOwMcEb0iXF67U5EvSJUJYAt9HCc+BFYULXAsCRC3myedOkaRj9/sw4AlBHPn4SiLGKWLlNIFOaPgxgeesYEttkOgdDYLJEkpdoKsSK9WkTVdBGCKkc80SSiBlnbIsGacbSGmIajG6gjzLiEJBEBW46e0ALRw4UX6AtRbfi/E9n8GgRzIaE8Y18rICUZCXOWVZMLZMtCmSsrIEkaKmBEWREdVihJAMq4LA97BI8rzE80KKMsNY51Jbq0VUlaa3vklrqkM+HtOZmSUIFAJJqz7FaDSmzDPqjSnSccbQJPhKoZQL3fOVxfN8KjNpoRVO6xCEAXIiBtdikuCOpNRu4C3wicIITwWMxwmmMkRxnU6n42htQlLkGbqoaDQ6WOlmPyprWNi1m1oYYW2FUoqiKBmPxqRpRlaUVFWOLXPW1s7iSbVzc4miiFoYginYWutTazexxlBWOdcduo4oiDh58hR+HBFGNaQfccvtId2tNbIsI08zRnlGVRRkZUZVVs4y2PNIs5zKbIc4SqIowFpDUeQYqRDWOa9hNUiFQpAXBcZKPK/CWo88L1ES6mGIkoJMVxSF06m02x1C3yPLC4yx7N+7j9CT5GXKeJySFTlJmpDlGVLIiZnzl2Y9OAkbVMABpViaDNq/NMEGqMpy/UMjNneHjDqX3k4FAmsso90jHnvrh/EfS7DSTcZo9fxa0uPa5gX/l8bbCUZtjhZQOqA92MW4vkGlcoaNNSqVs91OjrMOzfE8M1sHaSRz+GXkPn9xVsZ29xkw0lL64wu2vdU5QR6MqLwcaSSz3YN0BnvOfWgyP3T1EpdZxr3glzF+GVNLO3g6YG3uGffuZPm6SABniDCletxun6SwPqvVgrOQvWiVtiE4VuwnNTFCWDITsaUvnJ2NZM60d84zfp9/ioYcIYAFb5XMxHRNG4lluVyia6YYjgsGYhbh1zHIc9oTccHmr1rbEGJbzVizOW8rH+Nmc4rPeDdzRC6xJjsvGmw8ld/Mlp7m9ugJGnJ8gWj/S722RqNXehd26hNPP8Gtu/ex1Jm6LNgQw3NA0tsqWf7sM9zx3ksF/7O7l5hamL0geO9ytXV2jfs/+FHG/eEl7x197Ck2l9dY2L975zU/8B2QuEx5YfAFgc3FZYzh2c8/zlu+7j2XfV8IwTu/9et47uEn+Znv/1F+/Nf/Aw986D4OP/gY7/2uv8KX/5X3c/M9r+EXfuSfsXZqmbd+/XtpTnd2zt1Nd9/J1MIcKydOs/fG6y5Z98X/DiZanW261exE7H79a27ls3/wMT7yKw5ovPUbvppv/4d/B2MsVVGQZ7nrmGx26W92+ZX/61/z5q/5Kvbdcoj+xhbd1XXWT52lu7bBYKNLb32DYbfHn/zOhzHaDcQ/+Iv/lT/8z7+OHwZEtZhGp0293aQzP8vUwgwPf+x+fviXfpq4UWew2WXX9fuZ37OLqflZvvPnO6wePkNnaQahBKcPH+Mv/9B3TaQAgnq7yQ333M6hu2/jLd/yHn7hA/+c0fqFwPKlrqLK2BysviTrumagYa0grkX0N9ZpNgKEp6CosEinJRDWCZ9xDlHuqbAdxOeePFoIPCaDaGl3BhhyQs/xpXMzUkriqwA/jIhin/5WF89TTHVm8KRLDBdKOoCiJFJJbti/i2dOnkEbg9CayPeQkc9QGIpajXq9Rd6Ypj4zjXdiGYNBa0mlK6R0WgqsRcpJBggaKdxx+UqA8NDaIK3FkgCGKGxgMYyGI6yVDgUTUgtbeIGHwEdbw9hAVWSEYR2ExQt82tPTJKMx1oLnhxSVxhhnHWu066JIKUizDIvE6BJjDF4wIggijIFGvU5eFEhpsRRoa9jcWEcKARM+vzYW00+odIk3zvArD6tLsjSn2WlTa9WRwiPPNWWekVaaOA6pSkN/MKJWj3fob9ZCUWRUk6Cp7RkFi7MaRliUUlgrJtqUCmOh3x8wTlLXfarVHO60FVVWYD2Fr3yyNEepkCwp8FWIQOGHNeqhJK638JTE933KLCfNUqwU+NLj1KkTNFtNBIrSlGgDlbV0ZmbobW1Rr8UI6wJ7sJrFpXk8qSjyiFoUIqWmLJ3tb5qOwbhE9CRJyJMxSZZipIRJZytNCrbGG6R5SVaVYPXEJtlz15ExCOksLfEEgYwxQmKtwZcQxBFoixCSZrNBo9GgPxgyHA1QQtHp1FCeYmZuiV6vT5ato6Qg9iVVZrBCYK79Z/uqq22SkAaOaU0mJfulRH4pgg0AC6durZG0rixGdRaoUO31EUc1Inc6nRc7r2yldvdbLN22i//anDqKPU/8vb0NC4zq64zq66zMPUlQ1FHGpzlaoD1cotPfi7SKYCIqL4OUcW2DU7seIom6FMH43HYvEJdbVmcPM9XbT5RfNUTj2uo8kKNlxaldD1H4Y84jLl36EWGJRM7+4OQVV3udf2zn3yUeY3PhgMqnpC7PHaPY5olNqiZTpnFA5IB/AiEgDSpS0USogJVqkb5usaWn2NAzCBygKa1/wTFdcAiT1xqkLJgud+jj3F0d5pBZRlrLW8onedA7xK+G7+S0mHXmKM8TcGzj+JVqgZVqgWeL6wnFuVDHb7/mNb1668TG2qsitE8IwXSjyY/9+q/wr779e+jULnJ/sqCG50C97eckz6wSXIYKq5Siv9FFVxoZXHnmeu3kGbLx5V0IBxtdPveRT/DV3/0tO+em1m5Sa13+d+r53s4A/VpruNXj8AOP8LXf/4ErLuOHAe/+zm/i4+/7ID/7t/4R97znHXz3P/0RooZLHn/T+9/FrusP8P/8k3/Ja7/iyy4EEFJy73veweHPPcqeGw6+4O9YSsnrvvItfPS//hbWGBb27p64RYLyYsIJ+JrdvYi1lj/57T8kiALufc87gPO6J9ZOJl1LksGIn/s7P8an/9cfARBEId/1kz/CzJIDNyvHT/HLP/6v2HfzIe7/4Ec5eNuNHLz9JoZbPY4+9hR/59/8XwgpCKKQ6+6+mYOvvwmAZx96nIX9uy9LcdtOMQ/jkOF517y1FuV7hLWQS+8PljIvsduWxJMlVOChPI8yL6jyCye9rLWs985SVi80APbCuuYRS2Ulcb2BKVPHRRNiZ7Aptq2nnE2Um7WWTvAtkSglndWqUvi+E0JrXZLnBYEf0GlPoYRx1CgpnThbOjFUrRbQDnzWu32MNfieQlrA87BSYZnw4cuCKIzJyxIZCOozTcoArDYEc3OEjSXyeovpXUtEa32S8RChfOcJbaAsS8o8p8zHZFGMnAAeT4FSCdBwM/fSYq2P0YaqGCJkgBQxWb6FNjlZajC2ge+HCDPCCkVRVZBkrsuBwPd9pJLUajXSLEUq16JTysNqg5CaqtSus+NQELbKEEq6mf3cDaiT8QhrDWHYwPM8PKkos5jcVJRVRn/Qo92eQluNP0ksz/KcPM+IohBjBc16G08JaEJR1hgNE/IiJ67V6Q/6SC/AaINUAiV9avUItCGuRSAkAok2TucyHidUVYFQkiIfg5B4QeCOzVfEccy2JXGSFQghyZOESkqKsiTPEwzu2rHWMp7QvbyJQ5nn+RhTTRygDIHvNA2e5wISi8JQmAIZbAc5SvKyoCgrkmREWeaoCaD14ogw8JG2wI+bLJ86Sac9T1UatDGMxxndzTUarQ5FWTne53jEsH+KPCsodIWpLJ60YCusrqgqiZYQCEmgPFBOexGGPlEYu/wW6bkk83qdNEnZ6HYp8pwgUASBoN7q8OzRk5w4tULoBxRlji411laU2tH5LubPfqlWCZw2Bg3skZLwVTBYeCFV5g7UfyGVvt4f4dWUAxovYV0AWASXBTDn9Z2xwpKHbhY4ibuszT6DMj7CKNrDJYSVDJorlF56AfVK7Kzl3EjZIui3znB0/6e46chXovSLSK3ffpYLQ691hpO7P0e/tcw5wu3l6/xju+Iy570dUBKo7lXXdrX1ba8rDj1iXHdlaqJ70FahrQIBm9U0Y+MGm8fLfSSm5rpPRYlvLK/Xz9I2I/abNebsgMC6h72dbMRYy+uq5zikl/mo/1r+V3AvA1F73vD0/GNPTI2E5zeYfLVXUrw0g6GXom5YWCKvSv708FN89V2XZh2J8blr2beK2+dv2KE2bZfL7NLkaTbRaV654mYDLwjQVUqj0+LA7Tfx1KcfQlcaay2f+I3/xbu+/S8TRI4yWW818K+Qy+GeSzWS4eiywC1PMoZbvQtcp449/gxZemXdB8Bgs8t//j/+b2594+v4zv/z77H/1hsu6Ubsv/UG/uGv/ptL6FFCCG590+v5vf/wK85mX73w58Mtb3gtrekOo16fhQO7r7icEIK4UXPn357TDk/eREmJ8jzCOOI9f/2beeDDH0dPxgjt2Wm+7GtdYOJzDz9BXK/xN3/6H/Hf/78/z+lnj1NkOZ/8rQ9x+5fdQ2d+9pLtWmt5/E8e4J53v/2Kz8KwFnLb++/mD//1rxN5MSDQpuKur30j7/rev3zpBywk/TGmcpNSVjuXtqAeEcQBT9/3CP/rp/7fcxPH1mKU5q1/7T0kyYhRb0B3dYNRb0A6GDHq9ynzgjIvJwZDX7iuGWgMB30iBUpYPGEwleO+SymRShAUPsZ3OoYwCIlrsbOzFY4C5ALy3N8u5lwxGg4pdAVSIITraojJZ6SYOEpJ3yVVYyiKglYtdkBE+bgcC4P2fbKqItcVUgqiwEO0YuRWD10JisgSND1qnsdgbZVAaYa6wJPb3EU3K18ULiujrDTCGERhiSJBqQ2C3uRCqCNEiO8HJMM1ZubaICTjzGCxeMrD9xT1OMBYixCKLPEJAp+yKjAGBy6kxPOcba2QgvF4SFlUBL7TZkRhhDbO5cpTEqTED0KKUlOmQ6JajVGa4ysPmWRsC0+lkshKkyUJMzPzRHEMGDzpoXXFMB1Tb9TwpI9FMkgyJ7AxBmOKiSWrT+T51OOAqtJYBOMkpSwzykpRlSX94Qh/YmHreR6m0tRqMVEYAdaJnAErFZtiGT/w8cMIU5UUWYKUFj/wCIMYi8GrNGmWYwDPk/ieQtuKcX9AHNfwfI+q0FQTypExmjyT6Krk9MllPD+m0oUDIrrixPHTCCyeJ+n1uoxHI+r1GlVRoMIQa0BrByoCAUHg0Wk7OkUc19Bas3I2xFqB58ekec5oPKbf79Hv9UmTMXmeURQllSmcIFy7bpvWuM5XVVE52QtV6TQnVkMYOYvbqiigMni+wg9D5ufmCIMQpRwoLBL3oCl1CdYglYfvR8/HrfRVW9u3UAOcMYYtY7hFKXbm2r5UAIeA8Kkx8uYaZvoLAEAlMPsiRK9y4Ziv0DFeMmgXTjNiVcHG9NELlrvcAH+H8moFQRVTT2bYtXoH0jzPTpu9cHa+9DLycMTppYfYmDmCltXl9/dF10u5vgvX5SbMnGZv0V/Zeff68OgOkGoXJbNJzrQuCDB4WJQ910ERAPYcuallE95XfpZn1B4+592AtNcuZr9kb79EflbPp27bs++V3oWdakQxv/Q9P4h3ObtdA3J03gAbiM5kjPsDWjPnqHx5kvKxX/sd8iSlKssdkHC5Onj7Tdx8z2t45OOf5pt/+G/y1d/9rfzo+/4azz7kQnOefehxlo+c4MBtNyKl031ciY4llWL3DQd54k8/R5Hll1CszERbuV3WWh6+736uf80tlwyad5Yxhl//V7/I0nX7+e6f/BHCOLo8pUwIwvjyJieNTouZXYsMe33ak/NkrSUdJfTXnSh81/UHiBtXB9AzS/Psv+1GDn/u0Svu73bN7lrkzHPHr7qMEILb33IP+24+xLHHnsYay2f/4GO89S+/d2cycHb3ItffeSvf/o9/iH/1vT/CqD/k8U89wPf+ix+97HnQZcXZY6d41we+4YL3z1lpC4SUvPXb3kN3a40nP/EQzUaHXbfu553f8zW05qcuWSdAa+HyrwPc8uV38clf/hD9FdexVb7iTX/tK3j3933jjm20NYaqrKjKinQ0Ih0lJIMRvfXNK673/LrmJ0OSJKSRTy0KQWga9TqyJtwYVTg6VBy77AlrBb4fILc1F+JcwJsQDmj4vs/MVMfN6JYlQRQj5bkTqYT740uJJz3mZ2YYJzme77lkciYtfCEIhcd0u8XZ3mii89BUSUZLBJTKMlg+gcn6ePWIntZov+4G+kqhtQDhWmHK89G2wlqQ1tBogjaGftcShCG+r/A8AaZkOBwShg2isIbBEPgRWZ6hhaE/GGIRBL4PGIQnadRrgMQa7ZyltJmIrC2+76OrimF3k0ZrCj/yUZ6iyrTj5CvwQp8wjomAtTN9pqc7RFGNRr2BkIKyrBiORghhGaWGPC+JBYzHY/IkcUmevkcURUgCRxXD6SWssDjSm2KY5Egp6GYDwsinKnLq7TZxLSBUiiCM2NjoouIA3w/JC5einqUZFZYkS7HGidV9KfHCEGEtvoR2s0aaJNhKUQtjmMzul6XLoajXaySjEdKTGCRaOztYpTx8P8BTHl7gMxr2yfMCYwTYiqJIabYaKK/JcNinFK5DJKSh1xswGI1Ryqc/TFhd2yIIAzxP4StFZUqqwnUJqqoEA4O85wCEttRqdWpxnXa7CcyCOIDWmqosKMqSNMvJ8pQ8z0nTnKIowWjKwml9tDWOqlWWFEXhnKi2aWfCw49j56hmJXleMR6OXacNJwgMfZ84dAnlwhr8wHNp7X9GaptVkgFPaM0uKWkJgQ/EnLt/vCprsvN7pOQMcDWmuAPdgvLuNmQG75nxS6pN2SExWmfrt01vutZB+jZ4CMoacTrFuLbhBNjnrWN7GWU8ZrauZ3brIK3REn4ZnQMZ4hqoPTtAWZBGfZLaFhtTRxg2VsnC4Y525UtXsmwJypillTuoZZd/wFucMcJW1EVHJwi1oVlUNIsK/BTiPpWFBIvE8rB3kPUgYt6u09VTO2dGT5LHLy4lXCfKnPf+q/Vn9GIr8F496cpCCKLg8jbWopqIwc9bdmpdcvzxw9z59jcAoCvN//zZ/8Tpw8dIhiPKrICrMBL9MGD/rTdw8ukjvPUbvppaq8H7v+8D/OKP/KSbgR4nHHnkSQ7cdiNCSow2lHlxSRdle3/23nQ9n/jN32ew2b0AaFxu0qHIch65736+/K98jTP0uUwNu30evu9+/rdf+BdX1IZ8oRJCcNPdd3D2yIkdoJEMRvzEt/4Ahx98lIO338xP/NYvXcuaEDjg0p6dvuqSfhSSjsdXXQacU9edb30Dxx57GoDH/+RzDLd6tGen6W9ssXBgD1GjxoHbbmTx4D4e/PAnuOvL33xF8fz6mRWmFmaJGzVG3T7rZ1ZYPnKC/vom+265gdve/PodUPaNP/y96B+qXJflCkL9a6nGbJuv+cffzkf/7W8ztXuG27/qHm5+x12o86yOhVIEShFE4VUDIa9U1ww0yqJAej5eEGKSoQvcsxa0y8NQXkAgfYQW5GXp1BkTkbeSDlwIBEI6hoHQLlhtujNDpbXjLAsxoVw5e1uF2HFSEhKardAJxD1v4jC17VKlmGp3iOMtMBpR5eSrXZ4eJQSeYi720OWYUWoJPUmjzBiYgCDw3Yx9WaCURBs5sT40RLHACkgzSVFotCmoKp8gEJiyQElJ2Gzh+x5ImGrPs7I+xJMeSvr0+wPCOHKdhKpCV9pZ2iomnR3PzapbSxh6rK+uoU1FvVlHeT4Ci68UpqqcriXwCYKAsirwPQ8hFPVaTKsRE0UxSnmM04wo9Dh7doVOHOPXm2xsrOMHARpLFMYIISiqklKXiCzH9wNn3TtxCJNSYbCMi4JRnlMWKaPCoAQMdEm90UCiWZxfQCqXYj4ej8AalJKYylKVGVZrSiGo0OB5BKGiu7E20ZtUCMIJhU7je4qqKBG2oiwLl6quNdiKOIrI8hTf9yh0iS41xihsWZHnBZWZdLHCOmDwEKgowlMeURQgrSBNU5rtGlpbKl2RZRnlZOBelQVFnuNHAXKrC9rgBQHNZpN6vYGSiqxwYTWOSufSUqVyXZ8wjJFianIszvGiKAuyLGM7F8b3AozWVMbZ5HrCI89T8qIg1xpjQJclunSdGieoB20K8sJ1xaw1VLpEVWon3OfPSm0/vnLg+IQqoIBYCKaFoD35I+FVOVpSBuZOF4ynPGetfYXlhAUbSso3d7AND+/hAUK/eLBhcYGBzdEie86+lm77FFtTxyi9bIf6dC2Ddr+KuPGpd8OgRtLYpL/vGbqdk+fWYSXt4RL7Tt9DZ7D7nIhc2POwxdW24xysjKzY6pxk2Fhlde4pCj85T/fx4rUrr2RNfKOY3bqeA6ff8IWXFxY4t5yxgJ9DMHJc6gkr+c5an33NZYbGZ2h6zAiJHw94PPBY1w3G1pJa15nCSzgYHEdiOFHuIzMRx8YzDKr4S/jMXr4E0Ahf2AD25So70fOpyaz2DggqBKK48BvYK6f58Ic+wR1vuxeA+z/4UR786Cd42zd8Nb//n37NuRJepfIk5anPfJ573v025vfsQghnnXrg1hv51Z/8N3z2Qx/jmQce4Z3f+rVuF7J8R7x8udp96ADpcMzZoyeY37tr53XlqUsGx2ePnmT99FnueMs9VwR6Rx97mumlefbedP1l37/W2n3DQZ68/6Gdmf37f+8jPPbJz2Kt5T3f+U0X5JBcqbJxwtrJZVrTHeIvoEWZWZqnt7Y5obRffRC/eGAPyvcwWrO5vMIf/9rv8u7v+EZOHz7GrusPIKUkqte47s6beeTj9/OD/+6fXtKtsMaycuIUv/Vz/w9rJ87w09/1w0T1mAO33cQNr7ud2998N43p9gXbFUJcErD4QkoIwaE33cb+196AF/oXUsVeorrmEcvW1hb1Wt3x09IxXqUn03QOcDhevSFu1NFJQqU1QRQhrEZObm8Sl5vh+T4oR7/y/RDfcz+mbatQ971u/0hBSOWkCpNEVj1puW9/Qdpa6nHM1N49ZEVJXUpGowFds4HCEoUedSVpeNA3mmbkY8YWgcH3BGla0GjEDEYVuoAgCLBGUubgewIlDMl4hCclZQGmLPFDf0J9cpoSr9ZkbuYAZVkR+jX63Q1nUVs50bu2BiEgDoOJVsXD95xX7TANHG0p8p29rPSojMCdCIvVFuUrlFSs9wcYIQnDiCTNyfKSIBjgSZ8izwh8j421Faan58iLnNmpace5sxblhZMBtkYbTZFXIKAoy8nkrHOQMkAU1dBVDsInTRKCMCTpD1ndGhB4PoWRiInI2ZYJQhvi+ixhXWKok6VOUyKQbOpVWs1psiQl8AOSoqLeDqiqijzLCIMIFXiEOsYwpJy4K1ltCMIQaS15UeApHyuhKgtKq/GigOnWPGury/SHPaanZ9EGamGM0Zo4jknHY+I4pN2oo42l0gVKKtIsJy8ylJVIGRLFzjlMSyfGH/QHTmOkFJ7nEQQBQRCgcU5RZtKZ25k3FgJbTdqLRqO8gLgWY7RBVyVGSnRWISchlb4ShI0GeVmRJikEPn4jngj4zYTWVaJUgPO0cunvUehPXvvSrMuxvi7Hiq+AobUMrUUCc0JwSCm8V6FoXBrL0nMpJ2+tob1zYrud2naI3X4xkpjdITwmJvk/L7y2QcZ0bz83Pfcu/Cpidus6qpNvYmPmCEf2f9LRoq7B1FQgeJZn+NDgM3xV/i7uLb6KbucEZ+efwErDwvotzG0eOqfD2OleXOP3YQVGas4sPsKxfZ++ID/kSxlcXFxB3mTf6XvOvXCVQ7vscZcRlBEK2Bk+JbPUN6/jXO8KkJq9qqRSBaN4C2sFm1NnWV54bGe9c94GAPvTgjODlBMrF1tj/9rzP8BXUbXiGnfuO/BK78YFZa3lpz/4P1nubXH9whI3LO5iz/Qsc1mb3fk0DSIHp4Vg0ZvimY/9FulozOrx0/zqP/1Zvu+n/zFPffZh0uGIUXfA1FVoPstHT7J2cpnv/sm/v5PfIJXk0Gtv4wf/7U/wD97zAY488iRlUeIHPuloTDpKqF/GUhccM0VXmlOHj3Hn2964M+B0g9pzzx1rLZ/6rQ/RmZth6fr9lz8PxvDAh+7jbd/w1S96QOz5Pu3ZaaqipKoqPviL/w2jNXtuvI573/Pl1zQwHmxu0d/cYt/N11+iBbm4gihyGRxaXxVoCCGYWpzjDe99J2/7hvfy6z/zS/zSj/4L7vvvv4tUivd85ze7yd0s59jjz/Dm938ljY4DDNZayrzg6Qce4eP/4/fYXFlDCsn7v/8DHLztJtpzMyhPfVG6dUIIgvjKFL0XW9cMNLpr6ygrqPKUPTUPzwt2Zl/dU9QDqRFSUm80AOu+JOkGuQIwE+2FMQaBQFsmhB3h/IrZTgyfpGUIQWUMQrt8DCusy1bQmrIsyYvCcdkrTV4ZrCdQgaIMPJrtXRyYmUIqyXyjQQA0PB+ZjsnKirYErQusscy0m0RxTFUWeMbQadYIQzfjjnX0sK6vCDyPPM9pdTqM0jFRFDpq06RzM9WapqoqNxudN2k2Ws4GF00UhYAlK0qEKNEjje/5KClJigKjLZ4XkCQjl0UhFFIoqjLH+iGhDDEGTF4y1e7QbLZckJypAEmap9iqpCidQCfLU6IoZDDOiMMapioZDLYABwydpbBFCQ8v8Aj8AOUpfF9hURgNWTGiKjWbG+sEYcBYKSLPzW4UVbmznjxzIn2RpJgoIA5Dtu0/w9CBwyiqE4WRow0h0VoThgFSCsIopKoqSiGpyhJtLEpax9nUFfVGgzCKyLOc4XBEvV7HCySD3hBtLHHcxFKxsb6G1RMvbaMZjkYURY4XBhSVwRrDOBlTaUOt3mBxYRfjQZ+4USP0A/LKGRSkeUlZVFSVy2ap8oIsLwFHdznHl5xwJ4X7DWx3aAyWMIgpynJC0fPACJQXI6wmL5zI3WQlVVUQ+D6+H02EvCAngNpYjTWCrCjJ85zA8xyQKV5aMfEXs3afL2S3lr61FDgXqvPn2c6/tRpgzVoCY7hu4gD2gsHG5LtzedRu3T1rJ5MXV662EASTbfrn79/E1qfRqzjwiR7H5zyoK8zihTdtOdbIlRwEyI0S9VwChZm4g1+4ba8KL6t3sFJTetkFy4VFgz1n72Ju84YdACCQ+FXE4tot5MGIE3s+w7WAgVJl2Due4t03LRKWR3lEHMEvY25+7t341bY2SFxj9+LinXd/JVGXU7sf3AEZX0yAYS8Lc1+6fdjRrsB5VLJLl7o4l8QK677XHeAlGDRWyINzZDwBdNunycLBJausVEkW9s97Zdub6tySe2dD9swE3HtTmz9Ldce+A+ydvjrf/otZ25N6R9ZW+PBjnwfcd+crj8gLmKfNoXCJBX+KfeEcU6pBt3+WJ/70Qf77T/8897z7Hdzxtjfwqd/+Q+dCeRUxuLWWR+67nz03HOTme++6RGA9vTjPN/x/vptf/cmfY7CxxcyuBTdQL65sq715dg2jNSeeOHzV40yGIz73kU+gtSYbJ5fQoqy1HH7ocR771Gf5xr/7PS96sCyEwPM91k8vc/LpIxx59EkA3v2d30Rr9sr6g/Nr/fQKeZIyu2cJ8QUMVaYWZsiSlKqsdoIGr1SduRmKPOctX/8eBls9Pvmbf8Dm2TU2Tp9l6bp9WGv5zB98jBNPHOaH/t1PAk4g/8AffpxP/ubvE9XrvOc7v4lRb0Bnfobb3nz3q4YK+FLVNQONWJeI4YDCV4hg2oEE64yUPOVm7s0kuExKx5m0xjhuuRAIaycUHQcWwL2GtdhJIJ/AUFiLsopSa8CCccJJbc1kkOeoWtpCYQW5tYClEJYk+P+399/xlqVnfSf6fdNKO5xUuaqrc5JaoYUSICGEGECAsfEYG4M9HvOxx8YBxnl8wdd3wPb1vZ47trHNMPYYBtuYYIIxYCGSACGsiEK3pG51ru7Kp07YaYU33T/edU5VdaxuupG6tX/6qKprh7XXXnvttZ/nfX4hIzeGrMgZlCUrq2OyLEcJiRQaXw2ha8mcZ3Vriwc+9xAbq6sMhhVSpnSCvKowRpNnZt8asOs8Wgl89PuidqMUWmqCjxADudaoXFGWBfV8zmA4oCwz8kzQNpJCpzC2GCIh+P3352M6ZoNq2KeJa2LwGKPRWlJUA9qmZTKd4ZpLVKMVQoxMJhNCEFi7QApJlifx+fkzZwDJaKzSVISIj54oLkcZItJUwHnHbO7RWiOFTlQfIci0JjMak2esr6zjm5qiHKABned0bY2Pe01hgOAJSHYmu+xOJEoJQFJVGYu6Zl7PaW3DsKqY7OwQEWijCM7T1jWu7SiKHK0VmdGIGNDGYIwhqOTU1bUphTvPC8qyomkUzWJBrqFaHTFva0yWs3nuPM52eOfp2poYJc45ZtMZXXApebsYIKVmMpkyny9QRUHEo5VhPCgYDpKGwtoOa1Pz1tnkBOZDH9QXktNQOoXTORhiTOGFIdK1DV17hcBTSIgCqSLeeZq2JjMFQhk6F2htQ4wBok+TOyIIlZy2ok/hkc7R2i7ZF79MccuTLvCOVIPO++nF2RCouXoqsKfjuBgCGz2N6gUhRhxwLgTOhMCerHFvH54NfU4mAhhfQecakEJDpY9UDyzI7vMpyMY86YcsREQbUSGjqteAIWW7wurudU959eH8IHk7urpIjWBNw2R0dv/xw/khimaM9v0P4ZX6CJH0Goc2b2dr9VGmwwvP+v7EfgMR6fI5XZ74yVk36Lclnldf8XSIIvLoyQ9in6T9eCmxN/HZo2SV9SrGpybQS8ei3CKKsE952ofY/+N5QUTJkQuvRoUnreDuO2qB0w2LcputtUf2aWM749MEuVcACrzqrnL8uqbX7vf3mY6reBZa38sJrzp+HZ85/ThSCL7x7jc/vfD6DwB7C06NtVyc7vLguTPc+8QpPvHYw3zkocuFegQ67+i8Y8KCB9uzT9nWP//L38OR60/wLX/9zyMQXDp7PhmBtM+cyhy855O/9UHe9e3f/LSCcSEEb/vmd/PzP/jveewzD6RGw1oW02dWk4XeAvXiE2eJISCecmzTQtunf/djHDxxlOA99334E7zl67/qquJ458IlfuhvfB+3fclrWTm48Yyv93yQFTmf/K0P8hs//vO4znLg+BG+4o+++5qL8tMPPIJ3nqM3Pbd5gDYZXdNhm/Y5aVbjjTV2L1yiq1tufcNdnH34FN/74/+KT7zvd/nAz7+Xez/wET7w87/CbHfCI5/6LO/5tz/B7/3G73Db3a/h27/nu7jxrjuopzP+8w/+KN/69d/5imsy4Hk0GkcPrTHKc7JMYOsa19vXCgTImE5K9sTeABHfJwrKPc5XSMmZkKYVMbikqYiC2kXmXUvdNrTW0XWOUZVz/cENpIiJeiIFOsvBpGTqcjhifTRAaoMQmr0vgZAaHxKtyrm0Ql5UOUIkUbvRmkOl4YwMDMucTOuUUSFBawUh7Bdzzjt8sHhvEUJjjGY+n1KUJa1rU6ctNI0P6EYifETnObPpnHoG49URxmjKsqDrUtKz7boUNKgktusQLqKkxLYtRZbjXSA3giI3DHJDrgRb0ym7011uvPXmlF7ZNUShqaoSowx5nrOzs4XznpM3XE+hBPO6I8tyYggoKSjKkrppyE2G05EcQ9t1FCbDR8FeyoEPgUVd42czJrsTZtMJzdkz2M6zfmCNCGRZTp7nWOeQRqOUwjvVU7FSCnbrOlxnmUznXLhwkflggCAQvEB3msJkqCzjwrnzVEWFMAobPYu2Q5KsfQttklWslkmjoBVd1xIB6xyLRdOnprdsbByknk/JymFyqFhd57GHH2FlY5VykDPS4/5zFEiVMZ/uUlYDQkj7bDvXj59Df96C0RmZKYgx9CnxaZrWdQ7bdbT9BAaSrXN/BJOwfL+ojojoUpNdO7z3afoh3T6fV/bltFQKlWUE5xK9rYvY0OK9Q0mF0YbnLou/gPGki+jeBWhVCFZj5JCUnA6BcyFwuexKaIDzIZArhYwRzR7B8ukRYL+ZEMBujJwKgVnf3F+5J891ad8r+SKwGSOXYkQBtynFwf6+g1JyIUYmIUJ7OaM6vVrkurNv4vDFOyibvVVl8dQE7StxlbA6YlxBVa8+6UHPpI9I/13VqxTtmOnw/DW8y6uL1Nhr4LzsgOppOGHPD15aXD+R+YMseUUU3PzYVzCaHb6qMQvSM6su8tD1v8Pk6Cb+ZJkGNjsOeaZFhGubnF3pwnXs/F2cOPv6lNL+NML4J45+nHOHPktd7HClaB+eekxeGW3Bi4//+y/8NX75Ux/j3M4O33j3M+sDXiiaruPMzhYXJ7uc2dni3M42d99wE2+5+Xac9+zWCx7bvMA9jz/Gp049wqefOMXp7UvMmprwHJPRZ8LWuQu8+zv+BG3TkjUN0+1dXGfZubj5jBkhW+cusnNhkze/+5mpQ4PxkFvufjWf+eDHeMNXvw2pFPXs6XM3AC6dOQfAdHsH7/2+6xCAs5bP/d693Pz6V/PbP/1LfOW3fCMPfPxePvLe3+ItX/9V+4/rmpYf+X/+ExbTGX/oL/7pF+3zUVrzCz/0H3jk0ynA813f/s0cvEJH8myIMXL6occQQnD45PHn3Kfh2hipJF3z3NbJ4/VVurZjMZ1x3W038fj9D+Gd4+1/9N284V1v474Pf4JH7rmfR+69nx/869/HV/3Jb+Jv/pt/wtGbTu5b+f/Gj/88r/nyN70omosvRFxzozEqK1bKgiBMojNFz14yuxQRLSVBJJcgKejzKZJ9KDG5GsWYCncRU2ElACHBScnDm9ssgiMvC4o8p1wZUa6N0AePUFQ5VTVA6xwIOOt7QbgHofFB4Lyn6Vqi84muYgxGJ7ei5GQlsLaj6Vq0UlTjFcYHDnDHnbcmUbH3zOYTvIfMGJxLydSqd6fSSKIU5JkhBkeRm164nvjzxIAPIHxABk9ZFcznc1ZEz8cLgao0RKEhDlLRrDRWGybtDhHPol6wtnqAEHzSW+QZbWc5c+4cBw8dgLVVZAQRPZPpnJW1MU1tcdqzvbPNY489xk033IgGus6R55q8KJNFq04TpkFbpImSyIgRjDRoY3rqmiP4dCy9dSAibdfifGR3OqUocrZ3dyiLATEKmrbrP2eXrGrLjBAkWZYavNa2QMT5GmKgaTqk0ti2pm5SE9TZjoDHxkghFUpI2sU0mQB4hRCSbl4TQqKfuZBSwEtVUhUVJjd0bYt3nvliDkGyc2kTIRW7OxM2Dh+mKjIOH0jj9bpJU4pAQEgYDAe91iSNqWM/NQPRNxCJ1COgnzikvBetJaBRRhEjWBdSQGEI+CiREazbc89hf/vWtum814a2bVFSEwG/15x3LdbaK3RPjhDdPs1HSNAvY43Gs0II8hi5SUoq4P6noQ2ci5HNnp645071TLDA5IqmwnE5MPCF/vRdLvvT9s6HwEGVqG5ZjByXkoX3+1OSK8gM7I7O9HqKG/bpMwLoTM3W2iPIYDiwdRMyaKp6rU/+vuKFn3anr+WdvPDGtDMLHr3uQ9zx4Nc8e1P0nBCcO/QZJsNzv49tPANimsQCaUq+/5LptigDjx/7GMfPvQ4RBapeQ0TJbHCRU8c/wvTkLvYta/gb08qlcBHz/m3U/fPnbDb2moyqXuO6M2+4TGF7BvetIxdehYiC00c/SZOnNOdlQ/H8cGR1jT/z9ne9aNuL/TWic5ZZ0/BzH/1v/O//9edorO1NZ+DVx0/ylltu51OPP8qpzQtsz2c9K+FF2ocQ+Y//6F/yCz/0Hzh2yw089pnPEULgx/7Rv+SB37uX173jrVx3200MVsf7RfLnPvop7nzrG1h/BgcjAITgri97I7/107+Id471IwefMeAPoOsnKLub29i2u4o2FEPkv/zgv+Pm197JuUef4K63vYnVwwf4P//WP2C+O2W4OsZ2lvf825/gzIOP8ff/0w9x9KbrX7RGoxwOmE+muM4y3ljjq771m6552zEETj/wCDozV+WAPBOklHjrmE+mHDh+5FkfWwwrpJLsXLzETUfu4MgN1/HZD/4eX/ZNX8NgZcSRG67j/KknOHbz9bzqLXfz7d/zXfv2tzFGzj5yisc++yDv/nPf+oqcZsDzCewLAecjSE9UEuEDUXhcuKKI8g5c2O/Skm1nxBOIcp8eTZASnSnyskIXFSbLedXxE5gsJzeKMk8fnDZmn5rSOsGia2i7ttcXQJ5lCBEp8hzXWdqmRcTIoKooi2JfPS+lxPlI3SyYzWZUeUFRDcmzksLktF1yCxoPRyAkR44cZWtrwmBY4L1nsajZ3toBH5hOZoxXRlSDAbky7M7nLObzlAEiFRpBaDydTZOL2XRC17XM65pBjEiVOvPcGGznMFpQDrK+gA3kpUHJjMW8pnNd0gh0jp2tHYgO7zoGgwGd9XRdaqoW8ylnz57jyJHDKKOomwV1bVldX8X7RAswMgMMRSFp2walk3Vwlinmk3lywZKKTEqcSQV+09W07Qxna5yzjMZHyLIMrVNxLwXM62QtKwUUJkNrjQ0C27VILamDwOiC9Y2DZFkOUhK6js1L5yFIciOpyhWciygJRmuyPGN9dQxCIJROjWrwtLajbToIAiFbfPC0izlIKPKci+fOkeVV8tieTzh+3fVIobl4/glGozGqb4St7WialrIoUELsNwih75wvJ54n7FkuS6EQYs91SiZxuPfEECnyQIw5QsheW5Fog8452q6ls5YYkp4n9tuOMu4vSAcCLvQNTUifq9+bEqIQMSRXKwQ+PDPH9mWPnq+onnTBvfJfe0X85jWuHj6fycW14Gqu/2XKEQIOIpBKcSoE6p6qtUf+2h2fYTI6y6njH3vqNnt+/ukjn0REQVWvMZ4eZWV6nLWd6zCu95i/FvvYfi9Bcmn9YbZXnrjG51yNvWyerbVH2Vx/iIOXbu31Mc9zQ2kwQpCOKALiWedQ17rN/jsqBPFghrt7TJSgHmvABtSpBtFdbiubfMrDJz+AjIpqsY6IkkW5hT0Y6b56g3Ag6x8JUQvsV6zhry8wH5sgNu3lBkaI/c9fBk3mCo5ceBXHzr2GzFbsb+TJB6k/R4wrOHH2bg5s3cwTRz/BxY0HcLp9GprUsv14JuzMZ6wOhk97X2oaIp11aYHwihX5Pf1E03XsLOac3dnisc2LPHDuNI9cPM9jmxfZnO6yPZ/vMy/28OnTp/j06WdOnn8xEEJgd3OL3c2t/dvu+9DHue9DH0dnhvUjh7jjza/n9V/5pbzqrXfz8ff9Lm9+9zt56FOf5cKp05x+8FEOnzzBLXe/mgPHj2DytMh665e8hp/6//2f7G5uEUJk99LW075+jJHZ9g4A9XROM6+pRpePs5CCxz/3MP/8L30Pb/n6r2K8scbt4xHHb72RBz/xaa67/Sb+3f/6z5jt7PJ3fvSfcuD4kRe1cFZGU43T/nzZN/13HL/1xmt+btu0nHv0cYqqfM4MDYBiUFKNh3T1c080TJaR5TmXzpznptfeyRve9TY++F9/nbd+41ezc2GT/+NvfB9v/+Z38+Z3v5Mf///8qxSAvNdohMAv/uv/yNu++WufUwvycsY1NxqdtdQipUAHpWjqlnnXMm8dTQh0viN6y/pgxNp4iBeCTEu0UMS8xFQ5VTHAmCyJoLVKKdA95SnESFSSeeupXU2eZbBIq7/aGEJdU5Ul4+Eo0U5iQGuTVt7pqVkiOVu54Nnd3UEKgdaKwWCMMYZsOCaEwHg8SjatKkNnBSsm44EHHkILQ5Sa2aymyDI21taJhH6l27M9mXNxc4t1LdnZvYTzju3JdipQQ/Irz0xqFIo8T2PKukUITWc7jEz0GiEsUqQ07MxoutaiVUbwUC9mDMdjTG5oFi31osa5jnoRWV1dxfkAMlBWFZe2LhGB2XTGsWNHKKqK6WSCVgZnO2a7E6rhIO1/SBQ2bSRC6TTxGGYM8qznZ0uMEtSNReIxQqNljlaSGAJaJU2NdzbZEwv6ADmNtY563hBHY2JTI1BEKSjziunONFHblML7ACFRs/JiSPSeldEKUgt2drextktaHSnIyzz9MPSTgq5tUSrpFJoapNzLInEYaZAq4INjdXXIqUcucvToMQ5ubBADTHc2KcqCTCs662kWc1BpmhF8SOVU7DUXRLx3eJ9u896nZjf0NTCxT2vda0aSJkMI8MH1nvX9dCRcdlMzqtdamOQiEULatnOuf43L43Elkt1w/wc+hlTfSfn7pq+8XPBcLcQXwiFY2znJWn2AaEdw4pOQz4lAIQQ3SUlHEppvhUC7VzjuNSVPwn5ZKSJBBGaDTWaDTc4evpfB4gAnT7+Rje0bEFFdMeXYOxmeTmAMUTqmgws9XemFHzGnOs4evoeV6dGk2XgBqIsJl9Ye/X3tx5MRhhr3uhH+jgExTz/c/qYKESPuQoc606I/PoEm7B8qL1zSq0RgoLBvv7rJgHRYoxL4myrC8QJ1/xz9ySliuhe0CEU74taHv4rR/BDa5f2057moVmL/4yraMTc/9nauP/2mfeH3bHiB2WATpxrqYvIUQfcSCTuLxX6jEWNk3rac201Nw6efOMW9jz/Go5vn+VNf/k6+9NY7OLN9iYcunOPRi+d56Pw5ntjaZGs2Zd62uPDiTSVeSrjOcuHUaS6cOs1v//QvUVQlEXj/z76HejbHdXb/N6QcDTh28/Xc+ea7uettb+L4LTegjeHe3/koXdNQT585H2K6nVzJuqahay4bTyBEmgTEyPlTp3n7f5+0EcpovuHPfRs/88/+L8499gQ33nU7f/VffD8rB9Zf9NV5pVXKEqvKlDb+9/937vryN3L7m17HeGPtWV9vvjtl9+IWo/VVhqvPEkzSQ/Tp39sXNp/zsdoYxgfWOH/qNAB5VXLP+z/MQ5/4DP/u+/4pb/vmr+Vr/8c/DjFy9MaTNPMFw9UxMUbu/+in2Dpznru+/MWnAH4h4Zobjd3pjJg7gpQ4KXlocxMvBVVVMKgGrBXrDMpBcgQyhkxqTJZhne0XnyJSajwQo8DNF0mArVKOwp7trVAknr+zRNdbi0lBJGUvONdinSMKgZ9NwXm0MVjregtcgcIzHhaMqpJMaaQW6fqvcgbVYR4/fYFyMMIFR5aXVJlh48A621u7GCHY2tlhPl9w8dJFyiLHuo7HnzjNzmSGV57dBy+ASDqOEAUKEnUsgm1BmxzX5ZT5AK0FuckYVAMyZZJ9nE/OUFKKfpV+gSBQVSPm0ykroxFKG7S0LApN07Yc21hjY2Od+WzKxsYGuTY8urOFj4rVlRWCj0nbkRfYZoE2OVEImqamax31fEZEkuUFSIF1lvl8wXg8QKCZTndYG4+pqhIpFERP3nWU3lPPJmxsbFAUmqazOC+I3hOjSwnuSiGkSVSjEBG9PkcgmEy2yfKCECNt26CNYTGbobUGodja3cFoRdM6BsMCZdJ2mqZN4t2uQymN1jolpIuIMTngKUyGDREZI11nMVnJow8/CkIyr2sWTU2mUpNnuw4lcpQE29YcPXaMYZWKpqTDCbi+OUiFf5/yHfqVrd5Fq+0SzS6EVNy1XbvfEHRe9gvbYn9CkhzW+mmJEInzHtIUJBL3c2O89/vlpxQSH9Nx3GtWYgypqUEQ/cvXdeppccUEyZGK88dfRFrCS4WdlSeYDza52KwAC3DJhW8Rr555PF/i0lU6CQGzwUXuu+VXGNQbjCdHWZkeY2Vy7HKOBSk/YXd8hkiagk2G55mMzrIot9IW47PsybP8wO1NNSajc1jdvOBGo8vmzIbnX9Bz93HFBCuMNfbL1/A3lv1+XglBOJwTD2VELcjev8Pee798bCOhUITD+VOfLy7/O+YS99oR4aYS/Ykp6tMzCJHMVqzunkBGeQWt7RoLhb2eEoGxJRvbaWU2XkjXEWtqzhy+h7OH76Uz9ZP2ewmAECOdtZzZ2eLv/PiPcM/jj9HY7iqNxPf/559AipTr9UK1E1+oaBbpvGj7v/cQY2QxmfHgxz/Ngx//NL/4r3+McjTANh3/7Dv/LlIr3vHHvpELj6c8iawoELK3Zw8R27b9dht2N7c5csN1QJrq74XC3fYlr+Hknbfs337khhM89KnP8qavfQd/7h/+nWvKtHghaBcNW+cuct0dN/PXfugfc/HxM9zzgY/wu7/wq3zpN341b/zadzyjFe2lM+eZT6acuO1G8vK5c1eU1qwcWH9WmtkehBRsHD3EuUce54nPPcIv/8hPIpXi//2nvxvbdXzZH/6afZbPyTtvYffiJYarY7qm5Wd/4If5uv/xjz+n3e7LHdfcaDSJHJ5cmYqCVx86RFZkSKHJihyjdRIa5yXOt0zmDap1SJEunkIKtAxIJQlRpMLOezIpQaS0b6M0SoL3Hu89WksyDQQLBGaTHQbVgDxP6eS6MlRmTJFphEyJ24Tk5hSjx7Vzzk92GBRVn0wuaa0n+MC9997L9vYWH/zwRxFa8MTps2xeukRR5PiYkq6LTLG2uoZSkrpNQXlVnmNtR1WVLOwU24DznrqtyXLFaDjEhcB8MePCxUtkZY5BszJbQ0rITZ7C5MqCvMyRUuPcFNu5FLo3S+JrgQYpmM8X5HmGzjIUMVnQKsliMWd7e5cTJ05w5PA6K+MVhJC4GFhMpzRdEk6HoNg4OCAET73oemaKoChKQvBsbm/jbKRt5gyGq2QBovBY5/A+hQVGBIOqZHVlBa0lbWvpOkvbWRbNDB89mdZE50EK2qYmComzjqZZMBiMcY0l0wrvLdF7hNZkRUHsGrq2QWtN8BIlNN5bmqalKIvkulU3FEVBlmXUddsLwm3flAS6WUMxqMi0YbGYs7KyysXz56gGA6IPzBcNm5vblIOKblHjfEz6kihTkytSqKSWgqgUVimcS9qTPZlADEnAXeUGQUYQvQCcCtdPJHwIxCgIfk+f5PHOp4ZCJO5g6HM0VD+F88EmfnDXJVpGjCnAzyf63Z7TlXeJhx4Q1yRQfVkgRgJQA7shsNvb3aZve8KzvdOntyt96UuyK6cPnVnQmf7HKIgrisgnP+eFvlb6M0jPdHCB6eACZ47cg7HF1VsVAaubq8TF/S6CEIS1dD25CjYgJr5vjp9b+Lx/vJ/PVC1CFJE2m75wpcheI6oE4YYSf6LAX18Qh+nn6ym7IvqphBCEG0v8wzXqieaK7iF5TMWRes73sXd3GGrsW1eIRqI/PaW2E7bWHuXA1s38vswZrvwISYsZWTfg+ifewsFLt/H4sY+xuf7QKyAt/cXDj/72r3F2d5tHLpznzM4Wk8X8aT+B1r6CKabXiL3GA8D2Blbv/dGf4v0/9x7WDx/gyI0nufGu27n+zls5fP1xLpw6A4B3ju0Lm1dYuV8+7770G796n+YTY+STv/VB7njT6/gL/+R7yPKXLothtjthtr3DrW+4i2o85MbX3MENd92Ot47f+fn38ov/+sf4xv/p25+22bjw+OneqeroNaVoCyEoBhWbp69NU7Z+5BAf+7X389kPfZw//jf/Aq/9irfwyz/yU3zVn/zDNL34PnjPJ973uzzwe/fyHf/gb/HBX/x1bNvxmre/+RU9zYDn0WgMDh5gUFZ9boDEAy4KgrfYhSc6T+cdpq5RSiYdRU8TsM6lldv+840k2k3osxhsZ/HSM5vuohQURcHKaMxwMKKqchCSIsvI85wYHL5t8LZJjlJ2wcz2q8cEQhQEJD56ms5zbmeLXLUYnXF+e5fN7S1CDLiuI9jA+c0LFHlG29iejtORZxpvFDrLmc6mCCGZzuYc2jjAYDDEx0DdTABPNRiyuzNJq/A20u3OybRiZTBiUA3Jy4JMGVrniNGyvb3D7mROnmVoY1JWhvPUXcNktkvwgfOXLlGVI5qmwflUdEoEu5MJUkhaH9i8cIGiMBzcWMEISbNYEIVAS4m3Hd5F8jwjRIdvFhR5RTke4RH96r+m61pGw8N4C5cuXaAYDBDRIQR0fVq78zCfziirHClUolFJgdeaLEKrJNIYqjLH9wYAosjgisyTokrNmZIFs/mcPMsRUtF2NUZpWh9RRGbNLorkJpUXFSKS3MOkYrFo0pQGgdYaIXLqWU1dzwkChjqJsouiBGLSnMRIXS9YW99AKcViMWdne4u1tXXmTZtoYiIm0X//f61VaopN8jQyWZpIxZAaL2sD1tr9jI1ky5w0HEqm5wqt+vsUwaQQP+9TY2190mE40oVnj0phbaAs8961LTUhzluaNjW9UStCjLgQCeHle1E6/ySB96UQ2OobjicpH67ClU2FiBLlDSroK1b2BYtyu3fz6VfiRcBLSxTPHVb3QiGurhRfMlw95Qh02dOvtF21PzESVgzuTeO08v/kRqMLyC2Lun+O+tziWYXPQXhOnfgItz30ruTadKVI/ZnQP8bphkev+9BVAX3XjBiJlUrNxS0V4bqCqPaE38/y+n2zEUYaf+cAueug8QjX75QC+6aV/W09F0SEqCX2LSvEgSJ+cIvJ8GzfaLyIEJff22Cxzu0PvYsjF+/k0RMf6idWL925/HLBv/2tX/1878LLGsEHZtu7zLZ3OXXfQ3z4Pe8DQGeG4NIkOYTAj3zvP+E3f/IXOHD8CKsHNzh130OM11d509d95X5h7KzlN37i58mLIgUxv4TYubBJs6g5efvNTwoSNLzpa97BP/iTf5lXveUN3HL3q68q3GOMPH7/w8QYOXrjdde8ULd6aIN2UT+j69cehBCsHtrg/o98kte9463c/c4v42f++Q9z4PgRDl9/gs9+6OMAnHvkcV711i/hk7/1Qd7zwz/Jr/77n+Hb/u5fecVPM+B5NBqz2ST5+0uFkAGFRrgUaqaVIobeorVpEXkGKpHaJSShtCBxzX0feGaBGNB5QZnnmEwzPHyQwWBAbnIykyVrNQnENCqtbcBHQYgGFyJN51Masw90Hramc7x3RDw+tJy7eJ7pLGkkcqOIHpquRUiBkhrnHcPRAK01ru0IwGBQopRGSI21LVVVEoNAty1mEAlyh+gFUTmiNKA9qwcMwSvqeUDEjLzoC8YYmU9nNFqzM5tjbcO4LNCZTOI/pbHeMW8WeNuyuXmR1juEkhhtyJUhH6wyby2TeU2eGQieWNfMFgvWN9bYmszZmiyQMo02gw8I4SjNANs1CClp2o7JdJFW7k2GVhndwtN0HcNxlfI7MsN0d5dhWeD28kLIaOo580XNgUMHkoVtkLgQewqbRecmJV+HjBgF9aymqNJqq22TBiVT4G1DI1KQVJHn+ODouuRQZkwSVre+RWU5cWZ7JoK8QsBnybukx9ndnQIBISVd27G6toaUGcFalBQMh0Nmsxl5ljPb3WF1/QAQmVycYIxBKtMncNv0nqLvQyETTUpdUegrJZNoXKteqK7I8yw12v3533UdTd3QNCmbJPSOUVmWXTZKcA7vPHuVkfMe1zcde7kcqSFPDkbJxACkUBiTpUlICETveEkr2pcY9z0DJWqffXIFrrQNlVFRtCusTo4znh5hdXICESXGlvtuSE51eJWW7S4cvJ+t1cfYHZ15qd7K5w3XXGhGiBsGf2NJNPLJbrnEUhFWIyp/dnF2atoim+sPY3XDyvQohzZvp2xWn9WJKojAzurjnDr+Eepi9/ntO6RJRoSwZrDvWr/CWYqnP2GeuuMQwd9a4a8rkFsWcdHuPzcO1f7DnhOX5RX4OyrkozN4VPSTk5dAONWbIogoGU+OsbZ7kt3x6Wt6naeb9H2xNydLPDeeHOR36r4HOXXfg1fdduDYYT76K7/NY595gPWjh1hMptz34U9w6913XSVyfrERY+T0A4/ireO6KxqNPVTjIe/6tj/CL/2b/8hf/Rff/5SpxenPPQLAoZPHrml6IIRgtLbCxSeemnnydDhw/AhSSb7mf/hjZEXO4/c/1OtBxvu0qfs++kne+DVfwV1vexN/74/8OdYObVyVvv5KxjU3GkhFQCJJq4eJo54KqkDPIXdJkNs0KXAt05oQA1lmqKoMrXNCDBRFRVUNKPI8FX5ybxU3OfK0ztP5y6PuECLWeTrb4Vyg6xom0wnnLp7HR3DOE7yntUm8lJmIUZpF3ZGbnGE5xDuLLlLeQ+MsUYBSgnlXI4LEuxYhYbaI+NChVUa9aCBAnhWMhgMMknoxY75oUUWG1IbZbEqwLXmeo3OBazpsLQgyrXhLJdHaMK4qNCWeSBcsnWt6Gpmic12aRMhA11gyZQjBg9RgGwSwvb1FnpWE2CRhsjTszqe0jtSoqRSep7Qiesmiq7FBI2Vy7lIKnA/sTmeMygFCJo3A7nZLnhd4a9mZTemqIW0zx2QlIUpms0n6DPMqJaALsNamKYvQKCFxMRKjoutqAp7WuuTIFCN5UZJlaQpkuxatNZ2zaGUosrQPQhmEkoyGQ4iSiduhbVKTh9Qo35HkGYl2FGOkbVus7ZKNcTGACFs7m4xXVpPLWGdZLOYIlYIPnXVE71g/eACpVX8M9b4IfM/aVgiBFzJ9NiE5drXWgfWE0AAhFS8hYPu8kPRvj3X2CnE3zOZ1smC+aqm+p0fF0PsCJV2H957Q07P2NBt7FVEgnUdCSKRMU5eXL56eZvLkW0WUGJ8znh7Z1yQMFgd6y9cnXZj7f2qfoX3O9upjPHH041fw279IIUA+UqM/NsG9bpSCBK+AbALmAzuoRxaXn/CMm0oTop2VJ9hZeYKzhz7DcLHBxtbNGJ8/7ae6uf4Ql9YeIUi3v43ni5hJ3OtHXJUX+Dw2kxYsBJQKf1zB8eKp9z/ffdIS+/Z15mdmdGcWGFe+uCfZHkNNwHRwnlMnPsL2yqnn3Nu9BkO7nNXJCWRQONUxGZ3FK/vCJkpLLHEFNs+c54f+5venhTCt0FrT1g2nH3yUj/3qb3PithtZPbhBMaj2MzhejEI6OM8Hfv69KKM5dPLYU6YMQgi+9A99Nb/2Yz/Hw/d8lltef3mq4TrLuUcfRyrF4etPADwtJezJGK2t8Pj9D13T/u01WEIKmkXN4/c/xFf891/PYGXEzoVN5rtTgvOsHznI+VOncbbjD/3FP005fPYwwFcKrrnR8AGsD8i+eJZCEIJCKRAxInSadgyqEoInL0rKoiLLc7I8S1MCIfvpRvpQrA843yUhUozpx6S3Dw3RY61jsjth3tQ0TcNitkvrO/CREN2+L76IMdFscCgUHo3rApkpkDJxNb1Pq83epcLPBU9oLVoHnIXQJW1HoEXqSKkDhVbU8wZrLVU1YNEGFl1DcIHp5g6D8ShlMjhHIQegAyJzdHWgmwaMFiAVShqkELgQsMKl5igEdqaXyPQoTQQApQwoiN4zXlnHtRbnLSFaBBopPdErmnZGOUzJ1tt+mxgVWVExqCqqzNC0NVU1wmjIsgwl9b7NnxCCzZ0tyiJDK4X3nrZriRGmszlZlhNEJIpAlhv8Vs2wGmGbhjpErLPUTU2eFxRFQds2RDzWNQgcw9EqUYJrO5rFnKqscMGl9HjfkpnUMDjXoaSi69Lnb20Slnddi3WeulmgjcIYA4ikufEOQbLklWRkeYHr5uRFhq0bOucoBqPUeEbHZHeX0WiF3e1dmmbBcDSkqiratmU0HPbFvcNZT4gq5Yf4JAqnv0/KZGmrda/nIDWBQqZGoWu7JO6OIVEGRY6UPvmv95qVGGM/aUt0KR88PiYR+17xlWhRacIhe/1GIO43IKmp12h17WsDX4gYLNaf8b4oIoPFAUazQ+TdkJXpMbKuukr0jHiSvWrkisIspUC3Zt7rJsQX+UpucorTH5+i718Qn7TYKFyEup8wXcsqX38sI0ncvZXN2Vp9bsvPF/QZ7H2mQ9W7Qv3+P8cX40zYm2rE1Yzzb9/kxGcusbZ1HS/I9vfp0L/vutjliaMf58KB+3G661/76V/gst2uYn3nek6efiPD+SFEFAQZsGbBzvgJumz++5KTLLHEHmKMeOt610/YPH2O7/vj30lelawcWOfQ9ce57rabOHnnLRy98SRHbjjB6qEDlMPB/rTh+TQgpx96lE/99oe48a47+M//6kd56ze8i7d8w1ddldpdjUd81Z/8I/zsD/ww3/Z3/worB9YpBxXzyZSLT5zF5InJ8KH3vI97f+fDjDfWuOm1r+K6225k9dCBfTvgPQxXx+xubj1DQvrVWD24gdKa6fYuv/frH+DiE2d587u/krwseeLBR/jFf/1jvPUb3gUI3vt//yc2jh7mze/+yi+KaQY8j0ZDCIVSEq0kxshkF6vS1ELrFIgn+0IJFDF6gl1gowcfUHmB1poYFYjQU0FSPkKMgUXT4H2gXtTM5gu6psZGi7UpS0JKiRTJIUgiyLQmiwqHxUePDwKlNEoKXLQolRGiILpA4xYYo5JrkLM471PqY9fgpWI4qPBB0jmP9y0Cg3MRITw+Ah6adsG8blhZyTi7s0kXIkoVtN0iUa+0I1dJCC90pBoWCAKLOgXXZUbjfMvCNrS1RShN3YLVC2LwgCRaj1GGGCKL+S4aTecj3nnaboHQgbIc0baGXCmi0tjg6FyLiIqu9cymHS44JrMZOlMMhxlKlAihKItq/xjWdSTLMoJPYmXnPU3bcnHrElVVIIVlOm84c/48J6+7nigiu9NturbDGENnW5QWRALeCbquTQdKNmhTEmIqqGVIjmLT3R1WVtZRmeppQpb5fA4hpOYmRmpnyYtqP4Ml5WKk1dIsU0lkHkErzWIxx2Q59WJBXg65dOEc45U1bNexu72F9Q7ftpSFZ9HW2LZDREnbnU8NQfQMB2OUKRFFEshb7wm9SYG1aTKX6GKe0Hmi8pgsQxtFYfJUxgqRGgfv6FpL2zpaa2nbdo/9kHQb5KRJoEu0OiLeuRRkGZKFbWstruvYOLDBYDDABU89W1C3Dc5arPXYl4kd4zPhDff8iWe9X0R5dWPRI4pIkPapkw9gVl1Krka9WDrpF744LuDPirScn6hGc/8MR0Q870MlrvjzJYcSyAst4WR5lZ7i8/npCtL32t9e8cgbPwkfjKxOjz+pIX7h21+UOzxw02+wMz79lPuelhYVBaP5IU4+8SbWdk+mVPJ+H2SQ5O2QwxfveOE7tMQS14AQAvVsTj2bc+7Rx/nUb30QACEleZEz2ljj0HXHOHHrjRy/9UZO3HYjR288yfqRg5SjYa+LfGoDEmPkg7/46wB89w/+Q4qq5Of+xY/w4Mfv5X/4+3+NrMj3n/e2b/46zj/6OP/8L30POxc2qcYjRusrbJ/fxBQZP/6P/yXjjTXmkxlv/O++gvf/7H/lic89QvCeN7/7nXzLX//zKJ32Y7i2Qj1b4N3VCelPh70m5fxjT/Br/+HnuPXuu5BK84Gf/xXOPnSKD//yb/JHv+s7uHj6LO/7yf/Cn/l//Q3K4UvjzvWFiGsXgw9KVscjtAzJHUokUgd0RHe1mDMVVpLoIy4EgrNga0w+QOgcoqeua6azGdP5lKZrqdtFmkyImJKspST0BbAQIIJEGE2hCwSCTCuIEoMB4bHW4wQEbxECQvD4GIC00tzaGlykd2ClKkc42+FiR103uM4jpCbPcqyzeGmAiBRQ5Tmd9wgf2Z7W2GBYH41SWKHXBCFxTlAYQaVHeN+mojWGNFFpGqQqmSwWeJ8E6zEmehFEnAzE4FOYmwfvI7O6w+hIsB5PwAdBZx2RGc5bFl2LzDTKZBRBMagkVamYzzVNDUQHncfVlmzkWDSR2WxCVQ6pFzVZUVIg6ZoF3nbozOCCZzKZEhE4G7i0fQlpDE3bIr0nBEGW57SLBSbCzPpEixvk5IViOpmwmNYI7dASFrMZBw4fQwiF0AaUwLnk0KSUpByMmM92ubh1ieFgjNEK19vLIiQqy7E+EAnUkwadGSQK2zZILWidZV63yJ1dZvMZTWPZltuM1lZZlesIpalGQ2pryUpJPhzgupRO3nQBVNefx+liCOliCZHgLWWeo7UhxL2Ee0HTJNcr2VmkTNSmvVDIzGRkJqMKHmtTEKTdo1KFpDURUiFlmo4QMqLYsx5NDV/XWogB29RkWYYoSqpqAEKgpGRRP7fd3hcyZHyGS05/8bCm5sKBz112curhpWV79TGCfCr9w+p635XnSnxxTzOuwMvxMPSjA7nZkf3KJcJ1BeFojj9ZEEeamMn9h/2Bo5dkyHMtk3iWT99xigNbN3P83Oso2jHGlokS+hSK3xVajqf0C5eF+FW9yh0PfC1njnySzfWHiSLQ5JN+G5DZCu1yynqV8ewIma04sHUL2mX7j7n8mi/uW19iieeLGALNoqZZ1Fx8/Ayf/t2PAqkxyIqc0foqB44f4cStN3Litps4fuuNHLv5JOtHDjEYj+ialt/+mV/i9e/8Mk7ecQvaaP7i//a93PfhT3Lfhz/BXV/+xv1GoBoN+FN/77v5lrrh1H0P8rFffT8ffe9vsXJwnf/Hf/gBrr/zVt73U7/A677iLZy47SYghROeuv8hfvunf4nf+/UP8Mav+Yp9jUa7qPHOXZNgW0jJAx+7h8c+8zmuu+Nm/v33/zP+6Hd9Bw9+4tMUgwqTGd7/M+9h7fBB3vL17/yimWbA82g0MinI6dK1Mni86FcfRUxuO/sCVo3SGqkNrfPMFjVCeGLrmV7YpGlblBBY16U8DDyZySlyg+1aEnEkghTImEL/lJQpFdkHhAAtBJ3zuNARQwThIaZgN2JEaoWzLgkeAwgliTIiYuzpFRKCp3UOpKSNqZmRItI5h1IpTC96jzE5WYCAoPEBgqQwOY11ICAvK2zXUrcdTdtCVJRlRtO0xN5xC6FwGloije0olKFdtCAFWZ5hQ8BbT1ARJQzO2URN831Csgi01mGKlLxdhwbnXcqxUJFMa6yDunVoIylESpJO4qy0LwKFVJ6u3sa7yHQyZzHTtG2DFIJBNUZrTZ7njIdDvA/Y2rJ2+CAxgnWW4D1NY1EIdne3UVlJEUoEzX5qtZASiaBpFjRdS4yB6WxCUY7oOkeMqVBUvdWtEOkUbNsGIUqIls6GPqBPY11HkZdIqYi9W5gQuh9lpiyTyWSHWb1gbb2kzCsypdne2WYwGtJ1lqZrWFtbo27Se1VIfIDFIjmkJcTUcAi5H6CntKTzyRJYAL6n94UuopTo7WzZp/6JfgWZ3pM8hogUAqEUQQSMTG5oRPYT7/sNpDpDKvJCI0QSmrdtizYao8X+lKfIJPU1pJW+LLCf7ZAuuPPqEvfd8ivMq0tPsWl9Liybilcg9sTfIaIerZGPNeiP7BIOZLgvGRNOJAeqP+hPPgoQXUDdN0dOHF7B+YP3c3HjQYp2TNGOKZsVVnev23/OXrDfFVthOryAUy2DxQHKZuWq5iPvBtzw+Jdy8ok3E5RjNrjQfyciRbNC0Y4AcXl6sYfl12CJlwlijLR1Q3v6HJunz3Hfhz8BpN9Tk2cMV1c4cPwwo/VVHvvsg7zzT3wT7aJGDpP+41Vf+gaeeOAR7nn/h3ntOy6LqoUQFFXJbW94DbfefRdv+tp38NP/9P/i9je+lvs+8klufu2dnLjtpv3HV+Mhd7zpdZy8PTUHN7/uTtaPHKKoKur5gnq+oBg8u5aiGJRUoyGPfTaJ51/ztjfzbX/3r9A1DfPdtEhw7rHTvPff/Se+9W995xfVNAOeR6Ohe+cohEAKhdQGZUzSH+hULNuuo+ks8+kMaxuaZkFrExc/uMTl8zE1AASBECEJxJ3F2whK9lMNkdKR6R15fEBLgw02cfeFTPdLAUpird1fLIoogosYnZFnCucizjtccATvkCLZuwqlUJnZ59cHku96jJK6aRAuolXKVmidw9mO1lpMpggyjUWEyvDeoaTuC80kjF40HbazBBfSfTh2t3YSX19EgtDkWUaI4J1PFr69/kEIUFpS5QXBe6wLKaXaRZpZQ6ELYgDnOzIpGZUVKFK2SAxMJrOUOK41MURmdYMQUJWSvIjoIqd0BW3XUJUVXVdQlgWDwRDnPZubW2xuXqKrF0gtMRF89OTG4JQkNB6kRGqDsxavM8DjrO3NAtJn3DQ9PWwxRxJRMkV/IZI9LMIQo0h0PK2xtqPoqV1aS5TJECKijelX+lu6rmMwHFHkFT5YnA9onbEz2+TAgSMMqiHEyHRnJ00DhEyBgVojSY1rFOB92NcJddYSQkjNMgLnuhT8qBTWdlesOvSNRZrV7TuFRlLi+l6VEPcK55A423uuUzHG/UY5RlJKeEjOWWmi0jvYxEDoKVxSpnO567NhpBRIqajKl+9F6konKacbrG5o8ik7K09wae1hFtVWfwSXFdMSexCXGV4uos61yF/exL1xjLt7nH4H+AOqsfuvt9ixqAcW/X71OisZWJTbLMptAE4f+RR71LXMVshw9c9tm80J0pG3Qw5euhUV0hS9bFYZT49yZedRNOMnPTdlIyBAek1mn3pNuHKi8nwb9yWW+HwhxkjXtGydu8DWuQtAWpj82X/xI/zSv/lxTtx+E3e86XXc/sbXceyWG3j4nvvIq5Lb3/S6p50SfPRXf5s73vw6HvrkZ4k+cPtb7n7ax5WjAe/81j/MPb/zEd7xx76BwcoIY8y+DmVv3yDljEBy+tzL3IgxMNvZ5U9973fzLX/9f0Jpxft+4jd5x7d8I/e8/8P8+D/+l1SjIW/5hnd9UU0z4Hk0GiYvyIfrRCUJQWBdYN52TKebEAOLxYLONj01aI8PnASt0YQU3BckhJQqLRCoPGUVuD4BXCY3w7TCH9NkQUqJ9wEpJEVeoJxHkrygiZFBXjK3nta1aCOTm0+IOOlYy0e0qqWbt0kHIQQuCqpqgPWWtqn3sx+kBISi1DlDXdAFi7OWLrSYXOB8Bwi6ENBBJGqPTIWjEArrPEJK2q5DCYmKApNlCKlomuSG1dqYxMc48qJIhaMA6SREh47p/bbR0do2FapCIIRHGpOsZL1LQXJRM28ss8WC0bAgWEfrLF1rscHj22SNKlWyaV004IIhMx4VW4xOK+fGKIrCsDIaMlnM8cFTGknbRNZWVylKxc7ugq3tLbQ2CAFZlqNkaisGwwIfIm3TJBtgY5LzUpySFzkQ0UWZtDVyb/VaYnsrPdO7k0XnCNGihEELIAR0ltE1HW2XJhFZZghdw8JahE4i7a5tkRKKsiLESAyeRbNgvLpOCJGurSmHK4l2JlPWRtM0KCVRSu9TNDyhbzIUsXdT28PexWXP6SLdn2hTPqZJmRRX6mnF/mOFuPJ5/fagF3jTi8ovb5eYJiF7/94LARf9pGRP0/Fyxadv/6X0HxHafEpd7PT6iz3tybLFWOJZsLdq6SP6oxPElsO+Y42Yyf3h4EuJKJL5iNy0T6E/PVO7kwIL58+4zTaf8cSxj1/eTpRPnVQ8C7QrGC4OXPmK5O2Y9Z2TgGBeXWIyOvO0+3acv3TNr7PEEp8veOfZOpuajrOPnOIjv/ybSCWpRkPGG2u890f/E+/+jm/lyA0n2Dh2mNWDG1TjIcEHPvKe3+TEbTdhW8s3/9U/+4wWvEIIbnrNHWyfv0iMaZHTdh1nHznFYjrj3KOP8/h9D/HYZx/k7MOnqMZD3vR1X8mbv/YdnHn4FIvJjDe86238kb/8Z9BG4zrLgx+/lz/xt7+Tx+9/iJ/9gR/mf/4//hHVS5Sc/oWMa240dhrH9Pwmna0J1hJDIEbf25z2dpwiEnzv/qIURV4igsNGi5IZOje0XY0XJDcfKXHekumUE+Bj4vCLCIGA7Tyt8xip0CJZo2qjcc4zHA4wmUZGwSQGOuewvZOotx3KZjy+OIfRqi9qQ0pcDklM7EOissTWEiIpbFAlW9xgA2Ve4XRGUeS0bdezPCImCvLcEBB0ziaPf60IvsNZC0RaH8ikwFmPQGGkYbUqmc1qQm6QRiKJNE0HHqKU6Fzh6hbrXGpahEIqhRQKLzT0ydNdcFjriUriQkDqZF3btR2TeYP1njIzPY0pFbS2dzySnaJuOsosJzaOPM9oupat2YTt6QJCoK4X5LnBxUi9mCOwGKNYWRkBEtXbvi5mNQLN46dOIYVIieeZoapy1tY3yLRgd9ILwLWgbV2flu2QUSaBtU+TihD33Jk8MUp0lhoI7126jUjb1BSDFYxShBBpupoYGhaLXYajcaLtIZjXNWU1TNMJ2xJi39yKsO9mJkRqlK1r9396Q0xJ9CGkc/nKFYe9icTl5iGJ4Imyb0IuTy8ECikvPwcA0c9BemF4etzlBsQ5l15PyORoRZqoSKlBhH4SBCD6adbLVxB+ae2Rp9wm+GJ3h1rieUMIhI+ohxaINiQq1ZGMuCd+fylPJxfRn5ymgMNr2dXnrbaPBOWe+3E9WjWlzadPuf3s4XuAvh9aTjSWeIUh+MBsZ8JsJ1GT9mx3dWYoqpLh2irD1REP33Mf2xc2ede3fzOTS9sMV8eYzLCXV3UlkjXtw7z53e8kqwqysuCf/sW/SzkcsHpog8MnT3DTa+/kHd/yDYzWVnjfT/4C/+u//QmmWzt0bcc7v/UP79OsTj/4KCduv5nxxhrX3XYTR264jre8+4tLm7GHa280tncZVomek+cGCDiX8gZi8Aip8D6gdPrwjDH7wmyjMlwItF2Lc2k6kQS3YX9lXUVACPIswyhD5x3WO5RJYw6lJEaqpNcg0tmO+WyGE+CDI8SAsKmwHoxXqNuGGBzBglEFQkS8g+hrFvMF0miih/FomOxuXUCaDK0FnW2Y1TuUeUlVjGnrFu/Se0NBUyf9QAgepwRSJAFv8AGpktBZoJFCILUkNC31Yk6R5VgR0EIRYmBjbZ16UaeV99oTvUbJiFGRgS7wUmKbjrZtUKQEdbO6liYayhAiKKDrItFZMpEjcNiFI881mRC03tO5DqVBKIOWkUW7SO5NWU5hSnbm2xD3guNgZzIjRs9ssYO/FCnLiiIzhBAosoyiGDBcGRN9oHMNznta29K0c3Z3LlE3lrbryHNN23QcPHiQwxvrSfdgDCCZ1TOauqZpWjpr0SZDKUNTJ3G+lJpSFTjrUEqkIDzXEaQGkWxedycXcc6CEHjb4WI6L4ajRDPYn1RJASRReXKKkgiRzkPXT7qE0H12hSfEiOqXR/cuCntNQegd0IRIzdJeAwp7rApHCJenIHsIwfZOrAKFSMVECPvbTKOPNN0IBGQMEFODKGJA9aswvndre7li2VAs8aJBiKTfOFUjz7b4W6uUubGqeUnzs69yPvn8n8/P+E7FlX99/vdziSVeasQYsW2HbTum27v7t599+BTf+01/lsHKiI2jhzl28/Vcf+et3Pia2zl6UxKeV+MhW+cu8NhnPtcb1iiKquTb/pe/zJd89dvJq7LPs7r8Xbrl7rvYPH2OH//H/4pf/pGfYvP0uf39+PR/+xiv/8ovRQhBOR5SDiuyonjKPn8x4JobjbrrMDojyzOMKSiLjBAd08kuUhVAxFlL0zmkknhvkUIBycLWW48nIoFRL7ZtjKY0ydUnoDBGomVGURR0NomrQwhEkXIvnPN0oSUEj9ImFW4xYqRBZgIrHbkeoKXGo5AmiZuLLCP6wLxpETJpI5TUeB0JMiKjBB3x0mN96LUnClOUBCFwwVGWBc522K7FC003mWK0IMsrdhczlFJ7ZSyZybCtSyGH/QjOxYAn0HVtsusVkiAknXMIo4jOIVRk0TpEFExCCvSTUlMUQ5xtkFFg6yQyb9qW6CLlsGSy26CEpihAKkUUgqoY0HQLat/ivWc+a4l5xEXXOxjpFCanFFpXeB9og8e7Dt/6NO1wDpTA1zN2pyBFEv9Ltc2gHFAInYpypdjYWKeta9q2pRpKzpw+jRSauql55LGH+dyDDwIRLRVZnjEcDBmUJbnJKPOcRT2jtQKpNAHRO3MpXOehylGmBCHoug6tJUIqrAuUZdU3rtB2XSre93QAUiYxfejbjH5f9xycQkz7E/cKfvaCI9OP956+AthP90y0KGCP7kSvPemzMoSUxF6JIPeE4ollheipUQEgCuIVHOrQNxj4AFISvCCQtBlJ2C9JFY7qG6Illlhin0plA/ozM9SjNe6uIf7WAXEgifoyTeL3W2pHEmXLfGQXMbn2icMSSyzx+YdtO3YuXGLnwiUe+uRneD/vuSw8X1vhwPEjFFXJ1rmL/Ldf+FVueu2dDFdXCCFQjYdPu00pJQdPHOXojSe57Y2vJfb1Qj1bUM/mHL7hxB/kW/yCxTU3GrvbE1bHB7juuhOsjEdUZYnWMJ/vMtnd5dLWJdqmQQSH0AbRZ2wQQesMUcVEQeqLNaM1ZaYZFBVBpKyIGAJ5nmNMtr/yHIKHEJk1Nc52KKMRQqOkRGuFlgopBIu2QciU9u29AwVBpMJbkLIjnPUMqjGD3LDoFkyaQOs6BBJvA94vMEVOCJ5hUeGs4/HNJ7DWMxQVSEHrJCFYgoBFYzGZoF605EonSo6KxJAUv4XOyLKMMi+oqorpYk5wjsW8ZdbO2dqZoEwSA0cf0YHeXjc5aCmZqEquayFEqsGA2XSavOUBgWK6M0eFiMhgGly/Wg5n5+coBoa8UBR5hRA6pWMHhVYClRka26CcwTtHSyBEiw8R27ZkmQYiwkIXPbYLGKlRmcB5n2hn0lCUOU3dsLu7m4T2tsGoDKklo9Vx0m1IlTQ1UmG9o13UbG1vMZkqtDbUiwahIqBQKu2Pc566a8BkRBQ+JrG5j73kPIbUzJpBcnASgrbtKMocH8LlEMeYQiGThkLgbIdUMukdQtyTaKBEEvj3RKj9hco0dZBJPyIzEH1YX5B9oxCAPce1y7zqRI1KF52U5dW7UfWJ4IG9KUbSXCSRuker5Nq1Z80shCRGmfa1p2yFp/HSfznj6ZguLzX7ZYlXGPrvklh49Ecm6I9PCQcM8VCGv74kHDCQyX3h+BVusteESHK/0p+aou6dJdrUF8A0Y4kllnjh2Been72wrwEB+Iff/l0MVka4zpJXBaO1FY7fcgOrBzfSVKKvJ/a28dhnH+BPf+93c+q+Bwkh8Oi993P7G1+bFjaBvCz2LfS/GHHNjYbSgiLXLBYNWuX4IMiMIi9WODZa59Dho0wnO+xubVM3DcQk+A6AFAadKaRY0PYr+kKAjILadnTOQoDMGHansyR47QXjNjqM0hRFjixKpIhkWU7XWaRK4u+2TVa5AknbdGilETG5CM2mMxqtiAKGg4qV0YCzFy8wbZNwPZmkOgggVUgJzc4ycw60RhcZUnvqpkYEMJkGKekQZEVJphVVphlVw/TYKBBGU+Y5uUlTF2NyRoMSFRMx5sjGARZtzWwxZ3s6AQn1dIaT0NWOrCyJPtL6lryQVFmOJa3m161lMKrIdE7TpSC3PMuw0aGNgSjQSiFVMgoOLiJkKrb3MiIIinrSEEKi5uR5RdMukkBaSiIpo0IbAz41Pz4LBOf3NQYxRGZdjQ2OtrM0rWVldZSOX4RM5YyGK9R1TVmWFFmeQuecxWpFWVbUbUtPPqJtG6pBzmw2w9qWthaomNEFoEmuUE4IlI6I2Pah2wpUToyWrmlxPiCE7tPi96zuUtBjeu9JBwKq11QIYn+/cy65QYWA0hofIsSkP4okG1vney1F3LObpReF72kyfE+Fivsj1j1alI97bkvxitFrTOe6EIAj9T/9dwYI7DUY9I5qiXIl5CurwPE+8MiZXXamDSC47vCIlWFOmb+8U9CX+APG3nQDwAbU2ZZ4tkN9akYcKeKBDH88J1xXEAeKaJ592nFlOy+6gLp/ngToYdkGL7HEKxkxBGY99eoD//m9/Ldf+FWq0ZCDJ45yw6tv5443v44bX3MnR286SV7k7F7c4oZX38YDH7+XdtHwxOce5u1/7Ov3t7dyYH0/CPCLEdfeaEiNswu2L9XMphNWV9cYjYe0ncb09Jf1A8c5cOAEXbdgNtmmns+w1iV3IG+TXadP9p0+hCRSbsBZR66TnajWBu/sPi1qz87HyJQwbZ2nw6GUQWvdh/11KBRFZjDlACnTqvtsUWNDop1orciznOmipe5anLcYadLPRYy03oKN5CFgfWDWLsjKDBkERZamJvkgS5QYJGVVkWuD0ZqV8ZBc5zjvUUJSZDlCJrcnIcBbh207gvPYEOmaOUILBvmATGq2JrtYleFx5KUhWEuWZYwHFUVZYJRmezKh6zpGw4o8M0lEnysQBowk61PRM53Tug6pJSlfxOBbn6YgUuB9pHWWGCPORpquIQTJaDDExZQb0rkFMur0dCnJTIUTc4IQxJDcv+bNnAjMF3OENimQx7VkSuHjFCM029tbydHJe7bcNpnRiBCYT2ZEpTi/u0mZFThnUToFPFZ5idYZ0TmGgxLvAq3ztF2HNIFusaCJEecDzlvq2QRR5nTNDKNzlDB7xlYIBNKn5mZv2iBEsv2NeKRQxOgIgqR76DXXMSRaRHKBAkiWznsTiRDiFc5UYl+PIZRMbmo9oSqQHhMRSJHc0JIgPSL6/A4lIPqegiX67RF6RyqJlGkyIkV63p6u45WEi9sLfvNjj7M7T+J8oyWjKuNL7jzC7devo/fsS4XYb+yWWOJZcWXTEWOiOk0c8pEFaElcN4RDGf6GkrBhoLg87RAxQgDRBOSWRT2yQF60iAtdum9/w0ssscQXA64Unj9y7/287yf/C0prhqtj1o8cZPPMef63P/+32Tp7nrZuOHj8CMSI7bov6gZjDyI+WbX6DHjt3W/k2NHDiJgckYRU6CynLIcMhwOqqkpTByHwvu3pLYLOLujqBXXb9g5DoQ+kS1auxEjXdSAEmTEopdPKMBC8J89ztNKomFad512D6AMC6Ys2a1tECPt2sF3XpzsTsC5NRISIBBdZBE+IjspkrK2vE0JgMpuxeWkHJRRKwmQxRUuF0gItM/AeaRRlWdLYFJYmhURLSV6ULJqaMsup8pJhVWG9xyjdi4QFWicakpSStmlQQhLwTOYNkYg2mi5Y6sWCGB3RC7xP+gOlNFLAdDpFSIG1Ntmy4lF9VkYIkaLs08uJEDyFzhNlKC9QIiKUgRBTboQLlOWAzAjqukWolC2iVZboT84hSR7RXecYj0e03RzrHVrmaJFsfMu8YD6f0voOs2cLaxRN5xgPhxTlgOlkixiTjkFIidCQCYM0mkvTCZXOU16IdYyGI4zSjFdXaesFw2GJswHnI9Z2dF2TNBVEJtMpPqTU9zzLsdb2KfIGk2cUeZ4MCZylGI76CZhJNL3OYrQi+GRTFmI6X/a0HpenCKmxSPSlVOgKCSFxtXo6VNz/dwwhaXV6nuaecDw1Dz45T/VuVZ7UMMSQAiQDLjUgQqRzmfSYROPqaSGw/5hf+LmffTGvA39g+Inv/9tPe/tnH73EBz55msm822/cpBSsDnOUSvqY1WHBicMjylxx3eExUgiqQr/iGq8lXmJcYd6AEMShIq4b/IlEiZCXLPJCBzYgpv5yc9E//pWEb/17/9/P9y48byy/70t8oaMaj1g7tMFwdcx4Yw2TZzz22Qf4qz/w/Ry96SSj9VWyIu+NZV7e5/PXlbc852OuudWSPaVDkVZYow90ncfZhvl0l7KoGK2uMCjztJYbI9YGpFJk5ZDV9Q1s1zLb3e1tS1NqdogRF2Jvd+rQUuKdR4pImef43klI9k4+eZanBHBCorcg0MpgbYuNgRA8IQp0psmNItSexnYMyopcS2y9wEdFUQ7Y3Z2yO5nSeUemFFomoe/GygqZMhSZYTqb0wQwWYYNfeknkutQ6zpCI1jMF0QPMUQ625Jpg8xytia75CYlbRuTJV1FtMznHVErJrNdmqZjPBym7aYeAQi98Lujo0VISdd15HmB1vl+Y6ZNCnia7W7TLGqQirwwGKNIC/QaH2N6baUwwqBYIJTE5DkSxXhYUNcLWt/RuKZfPdd0wVPIkrwwdF2HiIbxYIwNqUkc5RkKkHKEahuUjCkUUQjaaYdYMczm26BS0a6USaJ7AToz1G2D0QKtkhBf6QwnO4K3XDjfYrRJmhupU6I4kawoic7T1DURUh6K662HlSB4T5lVdF1DM59SVSnjYzKdJK2O0MgsNSFRRESWI7zA1jVIcCSqlCDZNO99/4VMCd9RxP1pRhSgEJcpWWGPKuX3wxuVSgoQH3wvGd9rQFzadkhaEY/f/47FuCcl37N9pc/OCEjE/n2vNNx5wwbr44Jf+sDDbE+afQra1qTZf8zmTs2DT2wjhcAYhdGS4weH3HhshdtOrmH0tWcPLPFFjCdPO6YOpg71WP2sj19iiSWWuBYsJlMWk6daTn/PH/qzlKMBqwc3OHTyGNfdfjPX33kLR288yaGTx1k5sE5elftBgK8UPA/qVG/LqTXBeYJzeE/KuFCRxdxTN3OKomJjY713J1J99kVGXpQMR2NWV1Zom5rJdMpiMadtWwqj0bpAilTkxjzff12ZKm9in8a9Z+bnfd/w9BQWmWli1yGiYGU0xIeA954qH1AVULcdnbe44JFKMp8v8CGwMhqhVXIg8gRUn24dSE5bVgqkFoTo8UpgO4eWBbJQWAux7ciLghADs6YmKkGmDGqxQMRIkB1iOkMrjcPRtjVSaHKlKbICISQde+nkEVNkBOfoQsC5lkJrdJoHEGKkbuYomXH82GGc6+jmDWE8Tk5H0TOfNhSFwYwS7claR+gssjBkuN7WVSCCp25rlEo2rUgBPq3IZ8rg0WijiMGhtaJuW9pZh9QpWJEISkpWxmtUgxV2drYYjMYkDbNGBAGto+tlvdUgw0dHpgxN02FDSy4kxhhmbYOMEdelSZHvPCEOGcmMGMF6i/UtmckRMdB0C0bDEePxGKUVs+kcHwWTuINSUGYl3UIThUDqdI62XYvtGrqF66cOnnIwxAiD956yqogout45Sol0nmmtEVJgtCHGdE45F6BvqiKJ8iVFBBEvh/vhU7CeEMTgkVoTfMAH3085ZN9UpPNYSUH0ce/Lxp6cBmJ/e7LilVK8rO1tnw2H1ipuO7nOB+85k5y/nuFCG4HOetrOcf9jWzzw+DaffniTt7/+BEcPDJdC8iWuHa+gH/MllljiCxveOWbbu8y2d3nicw/ze7/2OwBIJSmqivGBNQ4cO8yJW2/ixO03cfyWGzh640nWjhykHFTozLwsG5BrJ49JlYSzfWHv+pyBEFPxJGRABsliMcO5jsGgJEYoywFKGXZnC3JtKMucYrhGXo1YLPoshcWcrk0C3852KTPBw6AaIKW8bDGqJME6NBJlBLHn2ye3oxlGSnSVk2mDJzJbzDHKkClNZjJiiDRdm1axpNinb9k+gNB5z6JryDJJgGTHqyU+CLoQ0UGQ5TlSgO1aAhEhVQof1JJ2scA1gd1mh0ob8rKkEILWOBZt2z/WgJTs1jOC8AgVCELRNC1BiaQtEPR6Bp+mN1KR5xkIKPIhbdvy6COn0gRIaVZWVrj+uuMIpXj4sUfxeJyXOGdpnccpjZaSqKBUhqyfCkmtcD4wbepUWEtBWVUYlUTszrdEkYLtEnUn4lpHURRILdFodqY7BBFobc3u1pxhVZFnWcp+MAWxrZFGUjcdMToyk8IZlVAU2qB1hmyh8x4bI9IDMVJKz+7sEj5GfHDEkPItonN0bUNXVGxub0MICKCoSkwm8VhEVChNuk9GvA9o3dvOymTDHIlYVzNb7GCdY+SHKJnR1EmAL6VGSYPRGXmekw9KBoMx1jqcdTTtDOccRJHoWHtJ30JcQf2R+41w6Kl9zjmMMdA7Uglkatz6BHApBcEHRAz7/vix/58UAkKyiH4l4vzWgk987vw11357F9wQIo+fn/LrHznF13/5TayPk1f5nh7mmduOdExfjhfuJZZYYoklXhkIPrCYzlIC+SOPc+8HPgqk37isyBmtr7J+5CDHbr6e6+64hWM3X8/xm69n4+hhBisjTJHvhwl/IeJ5TDQCzWKByQ1KpgI95TykbAVIBZEQnrZbYF2LlBLvPXW9YDRaIRQl1qYVcm0U1gtaK1l0AoKirEpEiDhvU3hcv4LsvYPgyfIKkalU3MVI8GkS4GNK+5aZwvm+qCdiRLLRVVpjjKFtW1wXiN5TmgIICBxKCkAxNIodl8Tl89ajRESSGpJZXZMZg5ARFwMEyPMivb61iKgBQTWsKPsPXSuZUrttm5Zhg8dHjxSBEAJNoHc6atHa4FxLu4hkVUbE44LHaJP2LwgKozClZI4l6JyD1Tqz2HHgyEGqsqDrLGurK1za3SLEkLQRUuKiB6tQQhFVGl60Lvk9FXlBFSp816FzTZ4ViXrkHJ1tk9OSlESfhLgqU2RFjm87Js02xhTsTHex3pNlGfNulnQjjUNLzXA4oqwGTKYTvFf4qLC+pgsO5z2r2YCmDcSQtBJCQpkXtG2HbyNCRKIIaKOxrkN48BGarkWqZOVrhGJruk2IDuccZZGyVGzXJRcwJelsoushNEIFTGbQQoIJFFmOFiY5VfmG2cwh0IQQ00ROSy6cixhTIKQiyzLyqiA3OSbLqcohyiiEiNi6pXWO4D1Nu4DQ52/0TY9S6nKzQa/eEKHXeUDwe1kdSaexn9lB6k2ebaX/5QznAr/7qdN09vlnhOw1dxe25rzndx/mzhs3UFIwry2PnNndo+Q/LU4eGXH0wJAbjo4xWi2nIUt83uF8oLOeMl/qj5ZY4osZMUbauqE9fY7N0+f43MfuSXeIROevRkPWDh/g6E0nOXbzDZy8/WaO33oDB08cY7g2TiGDXwA6kGtuNFZWxgghcc7hey669IEsz9KK/p49eUiWosRI9CEFy4UFk8kORZHoLlmesbOzw8WLF1E6p6gMhdZEpTEChEj2rJnSxJhhbbefUxCjxDnfp4UrPOn1Sp2cmOpFhxKSsiiQWYbOMrz3LOYzdqYz5k1D13UYozFaMxyNk+sRga4NBA9t45INqdRYb9EqYDKJyg226VPBY7J7zbVmWBhcF4lZhSkKYkhFbXAO7SDLNDZC0zYYbdCqYFYvKI3BBYcXiuBSUJ9Qkq6xaKWTFgSQREye0rArrTl2cEAXPLtthxEVRV6kDAghksVrFHjX4UKfzZDlSUxdL1hog5GKQZkjhSQC49GYxWKO95bZfI7tReVCykRR63wKDjQ5Siq2t7ZxwVEYjWsbvHPUdYu3jvkisjIoMIWmazw7kwmz2RylNOPxSqKh+Zxm0RJIlLpAoFnMQUmGq6uMB2POnzuHKZNNsRQSRCrYu7ZNzkMqIhQQRJ+tAT4IIoq2aUAVWG+RUSKROG+RUvcNTSrqXfRkeYb3js61iJ7K1S0cyqSVhBAD1gUIgrqdEYNLYvxLkTzPUDpRr4bDMYNqBAjKQYk2GdIuyMuCQDpXvI09pc8RvANpAN8Xysm+YO+CIIQk+Mvp5qGnWQW/53b1ykKIkd15ty+6f77YazbOXZpz7tL8mp93fmuOlIITh0Z87VtvYDzIn/tJSyzxEmHRWH7lQ49y7tKc1992iLe8+ujnvUhYYoklvsAQUwDhbrvF7uYWj376c/t3Ka0phxXjA+scuf4Ex2+5gRO33cR1t9/EwRNHWTm4QTkc/IHqQK650chMQVEUeG/xzrGoF9g20ZzKqkJrjez9QaNIgXJp6uBSsSwVztZcujRHCM2lrU0W8w6hNdlcUeYDFnWLFprhsGBQ5SAFIga0zgAQkTRNQNB2STAshep560kcrrOcLFMgFNKoXs8ROL91CWc9Lkac90ilKIxiXk9pmoDwAucbxuMhQgja1mF9SDSi4NIER2hMIbG2TVoOD1JLqtEB6npGrpL4uguBKNLkxYXI0BS08wWDfIj1Dus7Ip7OefJsgLQBUeXUs3ly8zUaGQI2AkriYiA4gQ2WDsFOZ4FAjIqNlSHaGKRUtIsFs/mMECPD4Rrbky0UoERAKPDoNHUi0lpLpnPmi5phGbHeM6tbkn45oqUhKyuazjIeDRiXA7amu+xe2gUJPjgMaXpTZh4ZNE5HdGexraUNaSqQZ4rG1rgmMplPKUrD6mDE4bXDaJMlSpcPqCxHyGQt3DaWznl0SM5fMYIOEESiqokIWitiiDjvyFWOjWn6oY2k0AW+DckJyieLOSkgSI8UgAsE6VAmo3UWEcC7rk/kduRFjlIpsZ4QyYwGZfDTGqQnMwajRR+u6Kmt5cLWBfT2JSIC7y3Rw3AwZHV1LZ2TJgM8WinKcpwaCJ3RtQuCc2kKFJPWw4fYN+t7jUey24x44DKVcImr8UIvmiFETl+csWjcstFY4vOKtvM8fn5Ka9Pfb3n10c/3Li2xxBIvI3jn9q14zzz4KL/365d1IFlRMN5Y5dB1xzh+y40cu+UGrr/zFo7dfANK70UACA4cP4LJsxdtn67Z3naJJZZYYoklllhiiSWWWOJa8crjYCyxxBJLLLHEEkssscQSn3csG40lllhiiSWWWGKJJZZY4kXHstFYYoklllhiiSWWWGKJJV50LBuNJZZYYoklllhiiSWWWOJFx7LRWGKJJZZYYoklllhiiSVedCwbjSWWWGKJJZZYYokllljiRcey0VhiiSWWWGKJJZZYYoklXnQsG40lllhiiSWWWGKJJZZY4kXHstFYYoklllhiiSWWWGKJJV50/P8BzaDPBzK9dYwAAAAASUVORK5CYII=", "text/plain": [ "
" ] @@ -272,43 +118,15 @@ "\n", "from perceptionmetrics.utils import conversion as uc\n", "\n", - "image_fname = dataset.dataset[\"image\"].iloc[1]\n", - "if dataset.dataset_dir is not None:\n", - " image_fname = Path(dataset.dataset_dir) / image_fname\n", + "image_fname = dataset.dataset[\"image\"].iloc[0]\n", "image = Image.open(image_fname)\n", "\n", - "label_fname = dataset.dataset[\"label\"].iloc[1]\n", - "if dataset.dataset_dir is not None:\n", - " label_fname = Path(dataset.dataset_dir) / label_fname\n", + "label_fname = dataset.dataset[\"label\"].iloc[0]\n", "label = Image.open(label_fname)\n", "label = uc.label_to_rgb(label, dataset.ontology)\n", "\n", - "raw_pred = pm_model.predict(image)\n", - "raw_pred_np = np.array(raw_pred)\n", - "raw_label_np = np.array(Image.open(label_fname))\n", - "\n", - "model_idx_to_name = {\n", - " class_meta[\"idx\"]: class_name for class_name, class_meta in pm_model.ontology.items()\n", - "}\n", - "dataset_idx_to_name = {\n", - " class_meta[\"idx\"]: class_name for class_name, class_meta in dataset.ontology.items()\n", - "}\n", - "\n", - "pred_ids = sorted(np.unique(raw_pred_np).tolist())\n", - "label_ids = sorted(np.unique(raw_label_np).tolist())\n", - "\n", - "print(\"Predicted class ids:\", pred_ids)\n", - "print(\n", - " \"Predicted class names:\",\n", - " [model_idx_to_name.get(class_id, f\"unknown:{class_id}\") for class_id in pred_ids],\n", - ")\n", - "print(\"Dataset label ids in sample:\", label_ids)\n", - "print(\n", - " \"Dataset label names in sample:\",\n", - " [dataset_idx_to_name.get(class_id, f\"unknown:{class_id}\") for class_id in label_ids],\n", - ")\n", - "\n", - "pred = uc.label_to_rgb(raw_pred, pm_model.ontology)\n", + "pred = model.predict(image)\n", + "pred = uc.label_to_rgb(pred, model.ontology)\n", "pred = pred.resize(label.size)\n", "\n", "plt.figure(figsize=(10, 10))\n", @@ -318,1474 +136,45 @@ "plt.show()" ] }, - { - "cell_type": "markdown", - "id": "bcb40b82", - "metadata": {}, - "source": [ - "## Translate Dataset Labels and Run Evaluation\n", - "\n", - "The evaluation call maps ground-truth labels from the nuImages ontology into the model ontology with `translation_direction=\"dataset_to_model\"`.\n" - ] - }, { "cell_type": "code", - "execution_count": 6, + "execution_count": null, "id": "a1f6ca6b", "metadata": {}, - "outputs": [ - { - "name": "stderr", - "output_type": "stream", - "text": [ - "100%|██████████| 50/50 [00:01<00:00, 35.81it/s]\n", - "/home/tejass/Downloads/TUDELFT_ROBOTICS/GSOC/gsoc2026-Tejas_Stanley/PerceptionMetrics/perceptionmetrics/utils/segmentation_metrics.py:147: RuntimeWarning: invalid value encountered in divide\n", - " return np.where(denominator > 0, tp / denominator, np.nan)\n", - "/home/tejass/Downloads/TUDELFT_ROBOTICS/GSOC/gsoc2026-Tejas_Stanley/PerceptionMetrics/perceptionmetrics/utils/segmentation_metrics.py:164: RuntimeWarning: invalid value encountered in divide\n", - " return np.where(denominator > 0, tp / denominator, np.nan)\n", - "/home/tejass/Downloads/TUDELFT_ROBOTICS/GSOC/gsoc2026-Tejas_Stanley/PerceptionMetrics/perceptionmetrics/utils/segmentation_metrics.py:203: RuntimeWarning: invalid value encountered in divide\n", - " denominator > 0, 2 * (precision * recall) / denominator, np.nan\n", - "/home/tejass/Downloads/TUDELFT_ROBOTICS/GSOC/gsoc2026-Tejas_Stanley/PerceptionMetrics/perceptionmetrics/utils/segmentation_metrics.py:222: RuntimeWarning: invalid value encountered in divide\n", - " return np.where(union > 0, tp / union, np.nan)\n", - "/home/tejass/Downloads/TUDELFT_ROBOTICS/GSOC/gsoc2026-Tejas_Stanley/PerceptionMetrics/perceptionmetrics/utils/segmentation_metrics.py:240: RuntimeWarning: invalid value encountered in divide\n", - " return np.where(denominator > 0, 2 * tp / denominator, np.nan)\n" - ] - }, - { - "data": { - "text/html": [ - "
\n", - "\n", - "\n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - "
animalhuman.pedestrian.adulthuman.pedestrian.childhuman.pedestrian.construction_workerhuman.pedestrian.personal_mobilityhuman.pedestrian.police_officerhuman.pedestrian.strollerhuman.pedestrian.wheelchairmovable_object.barriermovable_object.debris...vehicle.carvehicle.constructionvehicle.emergency.ambulancevehicle.emergency.policevehicle.motorcyclevehicle.trailervehicle.truckflat.driveable_surfacemacromicro
tp0.04.578900e+040.02.280000e+020.000000e+000.00.00.07.052100e+040.000000e+00...3.026940e+051.243870e+050.00.02.299100e+046.436000e+037.031500e+042.559201e+06NaNNaN
fp0.06.403000e+030.02.690000e+020.000000e+000.00.00.03.562000e+031.000000e+00...1.643800e+043.742000e+030.00.03.728000e+035.490000e+022.202200e+048.568000e+03NaNNaN
fn0.05.353000e+030.01.494000e+031.254000e+030.00.00.03.064000e+031.146000e+03...1.231400e+043.680000e+030.00.01.579000e+031.489800e+049.345000e+039.131000e+03NaNNaN
tn3326172.03.268627e+063326172.03.324181e+063.324918e+063326172.03326172.03326172.03.249025e+063.325025e+06...2.994726e+063.194363e+063326172.03326172.03.297874e+063.304289e+063.224490e+067.492720e+05NaNNaN
precisionNaN8.773184e-01NaN4.587525e-01NaNNaNNaNNaN9.519188e-010.000000e+00...9.484915e-019.707951e-01NaNNaN8.604738e-019.214030e-017.615041e-019.966633e-010.7892790.978781
recallNaN8.953306e-01NaN1.324042e-010.000000e+00NaNNaNNaN9.583611e-010.000000e+00...9.609089e-019.712650e-01NaNNaN9.357346e-013.016781e-018.826889e-019.964448e-010.6760580.978781
accuracy1.09.964656e-011.09.994700e-019.996230e-011.01.01.09.980079e-019.996552e-01...9.913558e-019.977686e-011.01.09.984045e-019.953559e-019.905696e-019.946789e-010.9982320.998232
f1_scoreNaN8.862330e-01NaN2.054980e-01NaNNaNNaNNaN9.551291e-01NaN...9.546599e-019.710300e-01NaNNaN8.965275e-014.545358e-018.176305e-019.965540e-010.7891150.978781
iouNaN7.957077e-01NaN1.145153e-010.000000e+00NaNNaNNaN9.141120e-010.000000e+00...9.132528e-019.436913e-01NaNNaN8.124602e-012.941096e-016.915187e-019.931317e-010.6177490.958443
dice_scoreNaN8.862330e-01NaN2.054980e-010.000000e+00NaNNaNNaN9.551291e-010.000000e+00...9.546599e-019.710300e-01NaNNaN8.965275e-014.545358e-018.176305e-019.965540e-010.6904760.978781
animal0.00.000000e+000.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...0.000000e+000.000000e+000.00.00.000000e+000.000000e+000.000000e+000.000000e+00NaNNaN
human.pedestrian.adult0.04.578900e+040.09.000000e+014.470000e+020.00.00.05.200000e+019.140000e+02...1.235000e+032.140000e+020.00.06.370000e+020.000000e+006.580000e+021.288000e+03NaNNaN
human.pedestrian.child0.00.000000e+000.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...0.000000e+000.000000e+000.00.00.000000e+000.000000e+000.000000e+000.000000e+00NaNNaN
human.pedestrian.construction_worker0.02.390000e+020.02.280000e+020.000000e+000.00.00.02.000000e+000.000000e+00...0.000000e+002.000000e+010.00.00.000000e+000.000000e+008.000000e+000.000000e+00NaNNaN
human.pedestrian.personal_mobility0.00.000000e+000.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...0.000000e+000.000000e+000.00.00.000000e+000.000000e+000.000000e+000.000000e+00NaNNaN
human.pedestrian.police_officer0.00.000000e+000.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...0.000000e+000.000000e+000.00.00.000000e+000.000000e+000.000000e+000.000000e+00NaNNaN
human.pedestrian.stroller0.00.000000e+000.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...0.000000e+000.000000e+000.00.00.000000e+000.000000e+000.000000e+000.000000e+00NaNNaN
human.pedestrian.wheelchair0.00.000000e+000.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...0.000000e+000.000000e+000.00.00.000000e+000.000000e+000.000000e+000.000000e+00NaNNaN
movable_object.barrier0.03.130000e+020.01.500000e+010.000000e+000.00.00.07.052100e+046.000000e+00...3.000000e+007.490000e+020.00.00.000000e+000.000000e+003.480000e+021.260000e+03NaNNaN
movable_object.debris0.00.000000e+000.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...0.000000e+000.000000e+000.00.00.000000e+000.000000e+000.000000e+000.000000e+00NaNNaN
movable_object.pushable_pullable0.01.100000e+020.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...0.000000e+000.000000e+000.00.00.000000e+000.000000e+000.000000e+000.000000e+00NaNNaN
movable_object.trafficcone0.09.600000e+010.00.000000e+000.000000e+000.00.00.02.320000e+020.000000e+00...2.000000e+019.700000e+010.00.00.000000e+004.600000e+011.360000e+024.700000e+02NaNNaN
static_object.bicycle_rack0.00.000000e+000.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...0.000000e+000.000000e+000.00.04.000000e+010.000000e+000.000000e+000.000000e+00NaNNaN
vehicle.bicycle0.02.070000e+020.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...0.000000e+000.000000e+000.00.00.000000e+000.000000e+000.000000e+000.000000e+00NaNNaN
vehicle.bus.bendy0.00.000000e+000.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...0.000000e+000.000000e+000.00.00.000000e+000.000000e+000.000000e+000.000000e+00NaNNaN
vehicle.bus.rigid0.00.000000e+000.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...1.400000e+028.200000e+010.00.00.000000e+000.000000e+001.290000e+021.900000e+02NaNNaN
vehicle.car0.01.746000e+030.05.300000e+010.000000e+000.00.00.02.890000e+020.000000e+00...3.026940e+055.070000e+020.00.07.370000e+021.821000e+036.235000e+034.619000e+03NaNNaN
vehicle.construction0.03.890000e+020.06.820000e+020.000000e+000.00.00.08.730000e+020.000000e+00...8.930000e+021.243870e+050.00.00.000000e+000.000000e+008.140000e+029.100000e+01NaNNaN
vehicle.emergency.ambulance0.00.000000e+000.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...0.000000e+000.000000e+000.00.00.000000e+000.000000e+000.000000e+000.000000e+00NaNNaN
vehicle.emergency.police0.00.000000e+000.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...0.000000e+000.000000e+000.00.00.000000e+000.000000e+000.000000e+000.000000e+00NaNNaN
vehicle.motorcycle0.06.670000e+020.02.220000e+027.190000e+020.00.00.01.760000e+020.000000e+00...1.369000e+031.490000e+020.00.02.299100e+040.000000e+000.000000e+004.260000e+02NaNNaN
vehicle.trailer0.00.000000e+000.00.000000e+000.000000e+000.00.00.06.000000e+010.000000e+00...0.000000e+002.900000e+010.00.00.000000e+006.436000e+034.100000e+020.000000e+00NaNNaN
vehicle.truck0.05.660000e+020.03.350000e+026.600000e+010.00.00.04.200000e+010.000000e+00...5.173000e+031.542000e+030.00.04.000000e+001.282800e+047.031500e+047.870000e+02NaNNaN
flat.driveable_surface0.01.020000e+030.09.700000e+012.200000e+010.00.00.01.338000e+032.260000e+02...3.481000e+032.910000e+020.00.01.610000e+022.030000e+026.070000e+022.559201e+06NaNNaN
\n", - "

34 rows × 26 columns

\n", - "
" - ], - "text/plain": [ - " animal human.pedestrian.adult \\\n", - "tp 0.0 4.578900e+04 \n", - "fp 0.0 6.403000e+03 \n", - "fn 0.0 5.353000e+03 \n", - "tn 3326172.0 3.268627e+06 \n", - "precision NaN 8.773184e-01 \n", - "recall NaN 8.953306e-01 \n", - "accuracy 1.0 9.964656e-01 \n", - "f1_score NaN 8.862330e-01 \n", - "iou NaN 7.957077e-01 \n", - "dice_score NaN 8.862330e-01 \n", - "animal 0.0 0.000000e+00 \n", - "human.pedestrian.adult 0.0 4.578900e+04 \n", - "human.pedestrian.child 0.0 0.000000e+00 \n", - "human.pedestrian.construction_worker 0.0 2.390000e+02 \n", - "human.pedestrian.personal_mobility 0.0 0.000000e+00 \n", - "human.pedestrian.police_officer 0.0 0.000000e+00 \n", - "human.pedestrian.stroller 0.0 0.000000e+00 \n", - "human.pedestrian.wheelchair 0.0 0.000000e+00 \n", - "movable_object.barrier 0.0 3.130000e+02 \n", - "movable_object.debris 0.0 0.000000e+00 \n", - "movable_object.pushable_pullable 0.0 1.100000e+02 \n", - "movable_object.trafficcone 0.0 9.600000e+01 \n", - "static_object.bicycle_rack 0.0 0.000000e+00 \n", - "vehicle.bicycle 0.0 2.070000e+02 \n", - "vehicle.bus.bendy 0.0 0.000000e+00 \n", - "vehicle.bus.rigid 0.0 0.000000e+00 \n", - "vehicle.car 0.0 1.746000e+03 \n", - "vehicle.construction 0.0 3.890000e+02 \n", - "vehicle.emergency.ambulance 0.0 0.000000e+00 \n", - "vehicle.emergency.police 0.0 0.000000e+00 \n", - "vehicle.motorcycle 0.0 6.670000e+02 \n", - "vehicle.trailer 0.0 0.000000e+00 \n", - "vehicle.truck 0.0 5.660000e+02 \n", - "flat.driveable_surface 0.0 1.020000e+03 \n", - "\n", - " human.pedestrian.child \\\n", - "tp 0.0 \n", - "fp 0.0 \n", - "fn 0.0 \n", - "tn 3326172.0 \n", - "precision NaN \n", - "recall NaN \n", - "accuracy 1.0 \n", - "f1_score NaN \n", - "iou NaN \n", - "dice_score NaN \n", - "animal 0.0 \n", - "human.pedestrian.adult 0.0 \n", - "human.pedestrian.child 0.0 \n", - "human.pedestrian.construction_worker 0.0 \n", - "human.pedestrian.personal_mobility 0.0 \n", - "human.pedestrian.police_officer 0.0 \n", - "human.pedestrian.stroller 0.0 \n", - "human.pedestrian.wheelchair 0.0 \n", - "movable_object.barrier 0.0 \n", - "movable_object.debris 0.0 \n", - "movable_object.pushable_pullable 0.0 \n", - "movable_object.trafficcone 0.0 \n", - "static_object.bicycle_rack 0.0 \n", - "vehicle.bicycle 0.0 \n", - "vehicle.bus.bendy 0.0 \n", - "vehicle.bus.rigid 0.0 \n", - "vehicle.car 0.0 \n", - "vehicle.construction 0.0 \n", - "vehicle.emergency.ambulance 0.0 \n", - "vehicle.emergency.police 0.0 \n", - "vehicle.motorcycle 0.0 \n", - "vehicle.trailer 0.0 \n", - "vehicle.truck 0.0 \n", - "flat.driveable_surface 0.0 \n", - "\n", - " human.pedestrian.construction_worker \\\n", - "tp 2.280000e+02 \n", - "fp 2.690000e+02 \n", - "fn 1.494000e+03 \n", - "tn 3.324181e+06 \n", - "precision 4.587525e-01 \n", - "recall 1.324042e-01 \n", - "accuracy 9.994700e-01 \n", - "f1_score 2.054980e-01 \n", - "iou 1.145153e-01 \n", - "dice_score 2.054980e-01 \n", - "animal 0.000000e+00 \n", - "human.pedestrian.adult 9.000000e+01 \n", - "human.pedestrian.child 0.000000e+00 \n", - "human.pedestrian.construction_worker 2.280000e+02 \n", - "human.pedestrian.personal_mobility 0.000000e+00 \n", - "human.pedestrian.police_officer 0.000000e+00 \n", - "human.pedestrian.stroller 0.000000e+00 \n", - "human.pedestrian.wheelchair 0.000000e+00 \n", - "movable_object.barrier 1.500000e+01 \n", - "movable_object.debris 0.000000e+00 \n", - "movable_object.pushable_pullable 0.000000e+00 \n", - "movable_object.trafficcone 0.000000e+00 \n", - "static_object.bicycle_rack 0.000000e+00 \n", - "vehicle.bicycle 0.000000e+00 \n", - "vehicle.bus.bendy 0.000000e+00 \n", - "vehicle.bus.rigid 0.000000e+00 \n", - "vehicle.car 5.300000e+01 \n", - "vehicle.construction 6.820000e+02 \n", - "vehicle.emergency.ambulance 0.000000e+00 \n", - "vehicle.emergency.police 0.000000e+00 \n", - "vehicle.motorcycle 2.220000e+02 \n", - "vehicle.trailer 0.000000e+00 \n", - "vehicle.truck 3.350000e+02 \n", - "flat.driveable_surface 9.700000e+01 \n", - "\n", - " human.pedestrian.personal_mobility \\\n", - "tp 0.000000e+00 \n", - "fp 0.000000e+00 \n", - "fn 1.254000e+03 \n", - "tn 3.324918e+06 \n", - "precision NaN \n", - "recall 0.000000e+00 \n", - "accuracy 9.996230e-01 \n", - "f1_score NaN \n", - "iou 0.000000e+00 \n", - "dice_score 0.000000e+00 \n", - "animal 0.000000e+00 \n", - "human.pedestrian.adult 4.470000e+02 \n", - "human.pedestrian.child 0.000000e+00 \n", - "human.pedestrian.construction_worker 0.000000e+00 \n", - "human.pedestrian.personal_mobility 0.000000e+00 \n", - "human.pedestrian.police_officer 0.000000e+00 \n", - "human.pedestrian.stroller 0.000000e+00 \n", - "human.pedestrian.wheelchair 0.000000e+00 \n", - "movable_object.barrier 0.000000e+00 \n", - "movable_object.debris 0.000000e+00 \n", - "movable_object.pushable_pullable 0.000000e+00 \n", - "movable_object.trafficcone 0.000000e+00 \n", - "static_object.bicycle_rack 0.000000e+00 \n", - "vehicle.bicycle 0.000000e+00 \n", - "vehicle.bus.bendy 0.000000e+00 \n", - "vehicle.bus.rigid 0.000000e+00 \n", - "vehicle.car 0.000000e+00 \n", - "vehicle.construction 0.000000e+00 \n", - "vehicle.emergency.ambulance 0.000000e+00 \n", - "vehicle.emergency.police 0.000000e+00 \n", - "vehicle.motorcycle 7.190000e+02 \n", - "vehicle.trailer 0.000000e+00 \n", - "vehicle.truck 6.600000e+01 \n", - "flat.driveable_surface 2.200000e+01 \n", - "\n", - " human.pedestrian.police_officer \\\n", - "tp 0.0 \n", - "fp 0.0 \n", - "fn 0.0 \n", - "tn 3326172.0 \n", - "precision NaN \n", - "recall NaN \n", - "accuracy 1.0 \n", - "f1_score NaN \n", - "iou NaN \n", - "dice_score NaN \n", - "animal 0.0 \n", - "human.pedestrian.adult 0.0 \n", - "human.pedestrian.child 0.0 \n", - "human.pedestrian.construction_worker 0.0 \n", - "human.pedestrian.personal_mobility 0.0 \n", - "human.pedestrian.police_officer 0.0 \n", - "human.pedestrian.stroller 0.0 \n", - "human.pedestrian.wheelchair 0.0 \n", - "movable_object.barrier 0.0 \n", - "movable_object.debris 0.0 \n", - "movable_object.pushable_pullable 0.0 \n", - "movable_object.trafficcone 0.0 \n", - "static_object.bicycle_rack 0.0 \n", - "vehicle.bicycle 0.0 \n", - "vehicle.bus.bendy 0.0 \n", - "vehicle.bus.rigid 0.0 \n", - "vehicle.car 0.0 \n", - "vehicle.construction 0.0 \n", - "vehicle.emergency.ambulance 0.0 \n", - "vehicle.emergency.police 0.0 \n", - "vehicle.motorcycle 0.0 \n", - "vehicle.trailer 0.0 \n", - "vehicle.truck 0.0 \n", - "flat.driveable_surface 0.0 \n", - "\n", - " human.pedestrian.stroller \\\n", - "tp 0.0 \n", - "fp 0.0 \n", - "fn 0.0 \n", - "tn 3326172.0 \n", - "precision NaN \n", - "recall NaN \n", - "accuracy 1.0 \n", - "f1_score NaN \n", - "iou NaN \n", - "dice_score NaN \n", - "animal 0.0 \n", - "human.pedestrian.adult 0.0 \n", - "human.pedestrian.child 0.0 \n", - "human.pedestrian.construction_worker 0.0 \n", - "human.pedestrian.personal_mobility 0.0 \n", - "human.pedestrian.police_officer 0.0 \n", - "human.pedestrian.stroller 0.0 \n", - "human.pedestrian.wheelchair 0.0 \n", - "movable_object.barrier 0.0 \n", - "movable_object.debris 0.0 \n", - "movable_object.pushable_pullable 0.0 \n", - "movable_object.trafficcone 0.0 \n", - "static_object.bicycle_rack 0.0 \n", - "vehicle.bicycle 0.0 \n", - "vehicle.bus.bendy 0.0 \n", - "vehicle.bus.rigid 0.0 \n", - "vehicle.car 0.0 \n", - "vehicle.construction 0.0 \n", - "vehicle.emergency.ambulance 0.0 \n", - "vehicle.emergency.police 0.0 \n", - "vehicle.motorcycle 0.0 \n", - "vehicle.trailer 0.0 \n", - "vehicle.truck 0.0 \n", - "flat.driveable_surface 0.0 \n", - "\n", - " human.pedestrian.wheelchair \\\n", - "tp 0.0 \n", - "fp 0.0 \n", - "fn 0.0 \n", - "tn 3326172.0 \n", - "precision NaN \n", - "recall NaN \n", - "accuracy 1.0 \n", - "f1_score NaN \n", - "iou NaN \n", - "dice_score NaN \n", - "animal 0.0 \n", - "human.pedestrian.adult 0.0 \n", - "human.pedestrian.child 0.0 \n", - "human.pedestrian.construction_worker 0.0 \n", - "human.pedestrian.personal_mobility 0.0 \n", - "human.pedestrian.police_officer 0.0 \n", - "human.pedestrian.stroller 0.0 \n", - "human.pedestrian.wheelchair 0.0 \n", - "movable_object.barrier 0.0 \n", - "movable_object.debris 0.0 \n", - "movable_object.pushable_pullable 0.0 \n", - "movable_object.trafficcone 0.0 \n", - "static_object.bicycle_rack 0.0 \n", - "vehicle.bicycle 0.0 \n", - "vehicle.bus.bendy 0.0 \n", - "vehicle.bus.rigid 0.0 \n", - "vehicle.car 0.0 \n", - "vehicle.construction 0.0 \n", - "vehicle.emergency.ambulance 0.0 \n", - "vehicle.emergency.police 0.0 \n", - "vehicle.motorcycle 0.0 \n", - "vehicle.trailer 0.0 \n", - "vehicle.truck 0.0 \n", - "flat.driveable_surface 0.0 \n", - "\n", - " movable_object.barrier \\\n", - "tp 7.052100e+04 \n", - "fp 3.562000e+03 \n", - "fn 3.064000e+03 \n", - "tn 3.249025e+06 \n", - "precision 9.519188e-01 \n", - "recall 9.583611e-01 \n", - "accuracy 9.980079e-01 \n", - "f1_score 9.551291e-01 \n", - "iou 9.141120e-01 \n", - "dice_score 9.551291e-01 \n", - "animal 0.000000e+00 \n", - "human.pedestrian.adult 5.200000e+01 \n", - "human.pedestrian.child 0.000000e+00 \n", - "human.pedestrian.construction_worker 2.000000e+00 \n", - "human.pedestrian.personal_mobility 0.000000e+00 \n", - "human.pedestrian.police_officer 0.000000e+00 \n", - "human.pedestrian.stroller 0.000000e+00 \n", - "human.pedestrian.wheelchair 0.000000e+00 \n", - "movable_object.barrier 7.052100e+04 \n", - "movable_object.debris 0.000000e+00 \n", - "movable_object.pushable_pullable 0.000000e+00 \n", - "movable_object.trafficcone 2.320000e+02 \n", - "static_object.bicycle_rack 0.000000e+00 \n", - "vehicle.bicycle 0.000000e+00 \n", - "vehicle.bus.bendy 0.000000e+00 \n", - "vehicle.bus.rigid 0.000000e+00 \n", - "vehicle.car 2.890000e+02 \n", - "vehicle.construction 8.730000e+02 \n", - "vehicle.emergency.ambulance 0.000000e+00 \n", - "vehicle.emergency.police 0.000000e+00 \n", - "vehicle.motorcycle 1.760000e+02 \n", - "vehicle.trailer 6.000000e+01 \n", - "vehicle.truck 4.200000e+01 \n", - "flat.driveable_surface 1.338000e+03 \n", - "\n", - " movable_object.debris ... \\\n", - "tp 0.000000e+00 ... \n", - "fp 1.000000e+00 ... \n", - "fn 1.146000e+03 ... \n", - "tn 3.325025e+06 ... \n", - "precision 0.000000e+00 ... \n", - "recall 0.000000e+00 ... \n", - "accuracy 9.996552e-01 ... \n", - "f1_score NaN ... \n", - "iou 0.000000e+00 ... \n", - "dice_score 0.000000e+00 ... \n", - "animal 0.000000e+00 ... \n", - "human.pedestrian.adult 9.140000e+02 ... \n", - "human.pedestrian.child 0.000000e+00 ... \n", - "human.pedestrian.construction_worker 0.000000e+00 ... \n", - "human.pedestrian.personal_mobility 0.000000e+00 ... \n", - "human.pedestrian.police_officer 0.000000e+00 ... \n", - "human.pedestrian.stroller 0.000000e+00 ... \n", - "human.pedestrian.wheelchair 0.000000e+00 ... \n", - "movable_object.barrier 6.000000e+00 ... \n", - "movable_object.debris 0.000000e+00 ... \n", - "movable_object.pushable_pullable 0.000000e+00 ... \n", - "movable_object.trafficcone 0.000000e+00 ... \n", - "static_object.bicycle_rack 0.000000e+00 ... \n", - "vehicle.bicycle 0.000000e+00 ... \n", - "vehicle.bus.bendy 0.000000e+00 ... \n", - "vehicle.bus.rigid 0.000000e+00 ... \n", - "vehicle.car 0.000000e+00 ... \n", - "vehicle.construction 0.000000e+00 ... \n", - "vehicle.emergency.ambulance 0.000000e+00 ... \n", - "vehicle.emergency.police 0.000000e+00 ... \n", - "vehicle.motorcycle 0.000000e+00 ... \n", - "vehicle.trailer 0.000000e+00 ... \n", - "vehicle.truck 0.000000e+00 ... \n", - "flat.driveable_surface 2.260000e+02 ... \n", - "\n", - " vehicle.car vehicle.construction \\\n", - "tp 3.026940e+05 1.243870e+05 \n", - "fp 1.643800e+04 3.742000e+03 \n", - "fn 1.231400e+04 3.680000e+03 \n", - "tn 2.994726e+06 3.194363e+06 \n", - "precision 9.484915e-01 9.707951e-01 \n", - "recall 9.609089e-01 9.712650e-01 \n", - "accuracy 9.913558e-01 9.977686e-01 \n", - "f1_score 9.546599e-01 9.710300e-01 \n", - "iou 9.132528e-01 9.436913e-01 \n", - "dice_score 9.546599e-01 9.710300e-01 \n", - "animal 0.000000e+00 0.000000e+00 \n", - "human.pedestrian.adult 1.235000e+03 2.140000e+02 \n", - "human.pedestrian.child 0.000000e+00 0.000000e+00 \n", - "human.pedestrian.construction_worker 0.000000e+00 2.000000e+01 \n", - "human.pedestrian.personal_mobility 0.000000e+00 0.000000e+00 \n", - "human.pedestrian.police_officer 0.000000e+00 0.000000e+00 \n", - "human.pedestrian.stroller 0.000000e+00 0.000000e+00 \n", - "human.pedestrian.wheelchair 0.000000e+00 0.000000e+00 \n", - "movable_object.barrier 3.000000e+00 7.490000e+02 \n", - "movable_object.debris 0.000000e+00 0.000000e+00 \n", - "movable_object.pushable_pullable 0.000000e+00 0.000000e+00 \n", - "movable_object.trafficcone 2.000000e+01 9.700000e+01 \n", - "static_object.bicycle_rack 0.000000e+00 0.000000e+00 \n", - "vehicle.bicycle 0.000000e+00 0.000000e+00 \n", - "vehicle.bus.bendy 0.000000e+00 0.000000e+00 \n", - "vehicle.bus.rigid 1.400000e+02 8.200000e+01 \n", - "vehicle.car 3.026940e+05 5.070000e+02 \n", - "vehicle.construction 8.930000e+02 1.243870e+05 \n", - "vehicle.emergency.ambulance 0.000000e+00 0.000000e+00 \n", - "vehicle.emergency.police 0.000000e+00 0.000000e+00 \n", - "vehicle.motorcycle 1.369000e+03 1.490000e+02 \n", - "vehicle.trailer 0.000000e+00 2.900000e+01 \n", - "vehicle.truck 5.173000e+03 1.542000e+03 \n", - "flat.driveable_surface 3.481000e+03 2.910000e+02 \n", - "\n", - " vehicle.emergency.ambulance \\\n", - "tp 0.0 \n", - "fp 0.0 \n", - "fn 0.0 \n", - "tn 3326172.0 \n", - "precision NaN \n", - "recall NaN \n", - "accuracy 1.0 \n", - "f1_score NaN \n", - "iou NaN \n", - "dice_score NaN \n", - "animal 0.0 \n", - "human.pedestrian.adult 0.0 \n", - "human.pedestrian.child 0.0 \n", - "human.pedestrian.construction_worker 0.0 \n", - "human.pedestrian.personal_mobility 0.0 \n", - "human.pedestrian.police_officer 0.0 \n", - "human.pedestrian.stroller 0.0 \n", - "human.pedestrian.wheelchair 0.0 \n", - "movable_object.barrier 0.0 \n", - "movable_object.debris 0.0 \n", - "movable_object.pushable_pullable 0.0 \n", - "movable_object.trafficcone 0.0 \n", - "static_object.bicycle_rack 0.0 \n", - "vehicle.bicycle 0.0 \n", - "vehicle.bus.bendy 0.0 \n", - "vehicle.bus.rigid 0.0 \n", - "vehicle.car 0.0 \n", - "vehicle.construction 0.0 \n", - "vehicle.emergency.ambulance 0.0 \n", - "vehicle.emergency.police 0.0 \n", - "vehicle.motorcycle 0.0 \n", - "vehicle.trailer 0.0 \n", - "vehicle.truck 0.0 \n", - "flat.driveable_surface 0.0 \n", - "\n", - " vehicle.emergency.police \\\n", - "tp 0.0 \n", - "fp 0.0 \n", - "fn 0.0 \n", - "tn 3326172.0 \n", - "precision NaN \n", - "recall NaN \n", - "accuracy 1.0 \n", - "f1_score NaN \n", - "iou NaN \n", - "dice_score NaN \n", - "animal 0.0 \n", - "human.pedestrian.adult 0.0 \n", - "human.pedestrian.child 0.0 \n", - "human.pedestrian.construction_worker 0.0 \n", - "human.pedestrian.personal_mobility 0.0 \n", - "human.pedestrian.police_officer 0.0 \n", - "human.pedestrian.stroller 0.0 \n", - "human.pedestrian.wheelchair 0.0 \n", - "movable_object.barrier 0.0 \n", - "movable_object.debris 0.0 \n", - "movable_object.pushable_pullable 0.0 \n", - "movable_object.trafficcone 0.0 \n", - "static_object.bicycle_rack 0.0 \n", - "vehicle.bicycle 0.0 \n", - "vehicle.bus.bendy 0.0 \n", - "vehicle.bus.rigid 0.0 \n", - "vehicle.car 0.0 \n", - "vehicle.construction 0.0 \n", - "vehicle.emergency.ambulance 0.0 \n", - "vehicle.emergency.police 0.0 \n", - "vehicle.motorcycle 0.0 \n", - "vehicle.trailer 0.0 \n", - "vehicle.truck 0.0 \n", - "flat.driveable_surface 0.0 \n", - "\n", - " vehicle.motorcycle vehicle.trailer \\\n", - "tp 2.299100e+04 6.436000e+03 \n", - "fp 3.728000e+03 5.490000e+02 \n", - "fn 1.579000e+03 1.489800e+04 \n", - "tn 3.297874e+06 3.304289e+06 \n", - "precision 8.604738e-01 9.214030e-01 \n", - "recall 9.357346e-01 3.016781e-01 \n", - "accuracy 9.984045e-01 9.953559e-01 \n", - "f1_score 8.965275e-01 4.545358e-01 \n", - "iou 8.124602e-01 2.941096e-01 \n", - "dice_score 8.965275e-01 4.545358e-01 \n", - "animal 0.000000e+00 0.000000e+00 \n", - "human.pedestrian.adult 6.370000e+02 0.000000e+00 \n", - "human.pedestrian.child 0.000000e+00 0.000000e+00 \n", - "human.pedestrian.construction_worker 0.000000e+00 0.000000e+00 \n", - "human.pedestrian.personal_mobility 0.000000e+00 0.000000e+00 \n", - "human.pedestrian.police_officer 0.000000e+00 0.000000e+00 \n", - "human.pedestrian.stroller 0.000000e+00 0.000000e+00 \n", - "human.pedestrian.wheelchair 0.000000e+00 0.000000e+00 \n", - "movable_object.barrier 0.000000e+00 0.000000e+00 \n", - "movable_object.debris 0.000000e+00 0.000000e+00 \n", - "movable_object.pushable_pullable 0.000000e+00 0.000000e+00 \n", - "movable_object.trafficcone 0.000000e+00 4.600000e+01 \n", - "static_object.bicycle_rack 4.000000e+01 0.000000e+00 \n", - "vehicle.bicycle 0.000000e+00 0.000000e+00 \n", - "vehicle.bus.bendy 0.000000e+00 0.000000e+00 \n", - "vehicle.bus.rigid 0.000000e+00 0.000000e+00 \n", - "vehicle.car 7.370000e+02 1.821000e+03 \n", - "vehicle.construction 0.000000e+00 0.000000e+00 \n", - "vehicle.emergency.ambulance 0.000000e+00 0.000000e+00 \n", - "vehicle.emergency.police 0.000000e+00 0.000000e+00 \n", - "vehicle.motorcycle 2.299100e+04 0.000000e+00 \n", - "vehicle.trailer 0.000000e+00 6.436000e+03 \n", - "vehicle.truck 4.000000e+00 1.282800e+04 \n", - "flat.driveable_surface 1.610000e+02 2.030000e+02 \n", - "\n", - " vehicle.truck flat.driveable_surface \\\n", - "tp 7.031500e+04 2.559201e+06 \n", - "fp 2.202200e+04 8.568000e+03 \n", - "fn 9.345000e+03 9.131000e+03 \n", - "tn 3.224490e+06 7.492720e+05 \n", - "precision 7.615041e-01 9.966633e-01 \n", - "recall 8.826889e-01 9.964448e-01 \n", - "accuracy 9.905696e-01 9.946789e-01 \n", - "f1_score 8.176305e-01 9.965540e-01 \n", - "iou 6.915187e-01 9.931317e-01 \n", - "dice_score 8.176305e-01 9.965540e-01 \n", - "animal 0.000000e+00 0.000000e+00 \n", - "human.pedestrian.adult 6.580000e+02 1.288000e+03 \n", - "human.pedestrian.child 0.000000e+00 0.000000e+00 \n", - "human.pedestrian.construction_worker 8.000000e+00 0.000000e+00 \n", - "human.pedestrian.personal_mobility 0.000000e+00 0.000000e+00 \n", - "human.pedestrian.police_officer 0.000000e+00 0.000000e+00 \n", - "human.pedestrian.stroller 0.000000e+00 0.000000e+00 \n", - "human.pedestrian.wheelchair 0.000000e+00 0.000000e+00 \n", - "movable_object.barrier 3.480000e+02 1.260000e+03 \n", - "movable_object.debris 0.000000e+00 0.000000e+00 \n", - "movable_object.pushable_pullable 0.000000e+00 0.000000e+00 \n", - "movable_object.trafficcone 1.360000e+02 4.700000e+02 \n", - "static_object.bicycle_rack 0.000000e+00 0.000000e+00 \n", - "vehicle.bicycle 0.000000e+00 0.000000e+00 \n", - "vehicle.bus.bendy 0.000000e+00 0.000000e+00 \n", - "vehicle.bus.rigid 1.290000e+02 1.900000e+02 \n", - "vehicle.car 6.235000e+03 4.619000e+03 \n", - "vehicle.construction 8.140000e+02 9.100000e+01 \n", - "vehicle.emergency.ambulance 0.000000e+00 0.000000e+00 \n", - "vehicle.emergency.police 0.000000e+00 0.000000e+00 \n", - "vehicle.motorcycle 0.000000e+00 4.260000e+02 \n", - "vehicle.trailer 4.100000e+02 0.000000e+00 \n", - "vehicle.truck 7.031500e+04 7.870000e+02 \n", - "flat.driveable_surface 6.070000e+02 2.559201e+06 \n", - "\n", - " macro micro \n", - "tp NaN NaN \n", - "fp NaN NaN \n", - "fn NaN NaN \n", - "tn NaN NaN \n", - "precision 0.789279 0.978781 \n", - "recall 0.676058 0.978781 \n", - "accuracy 0.998232 0.998232 \n", - "f1_score 0.789115 0.978781 \n", - "iou 0.617749 0.958443 \n", - "dice_score 0.690476 0.978781 \n", - "animal NaN NaN \n", - "human.pedestrian.adult NaN NaN \n", - "human.pedestrian.child NaN NaN \n", - "human.pedestrian.construction_worker NaN NaN \n", - "human.pedestrian.personal_mobility NaN NaN \n", - "human.pedestrian.police_officer NaN NaN \n", - "human.pedestrian.stroller NaN NaN \n", - "human.pedestrian.wheelchair NaN NaN \n", - "movable_object.barrier NaN NaN \n", - "movable_object.debris NaN NaN \n", - "movable_object.pushable_pullable NaN NaN \n", - "movable_object.trafficcone NaN NaN \n", - "static_object.bicycle_rack NaN NaN \n", - "vehicle.bicycle NaN NaN \n", - "vehicle.bus.bendy NaN NaN \n", - "vehicle.bus.rigid NaN NaN \n", - "vehicle.car NaN NaN \n", - "vehicle.construction NaN NaN \n", - "vehicle.emergency.ambulance NaN NaN \n", - "vehicle.emergency.police NaN NaN \n", - "vehicle.motorcycle NaN NaN \n", - "vehicle.trailer NaN NaN \n", - "vehicle.truck NaN NaN \n", - "flat.driveable_surface NaN NaN \n", - "\n", - "[34 rows x 26 columns]" - ] - }, - "metadata": {}, - "output_type": "display_data" - } - ], + "outputs": [], "source": [ - "# Translate original nuImages dataset labels into the model ontology.\n", - "# Classes ignored by the model config are masked before this translation is applied.\n", - "fallback_class = next(iter(pm_model.ontology.keys()))\n", - "dataset_to_model_translation = {\n", - " class_name: (class_name if class_name in pm_model.ontology else fallback_class)\n", + "# Inspect overlap first\n", + "model_classes = set(model.ontology.keys())\n", + "dataset_classes = set(dataset.ontology.keys())\n", + "print(\"Model classes:\", len(model_classes))\n", + "print(\"Dataset classes:\", len(dataset_classes))\n", + "print(\n", + " \"Overlap:\",\n", + " len(model_classes & dataset_classes),\n", + " sorted(model_classes & dataset_classes)[:20],\n", + ")\n", + "\n", + "# Build a safe translation map\n", + "# old_ontology = dataset.ontology, new_ontology = model.ontology\n", + "# dataset ontology -> model ontology\n", + "fallback = list(model.ontology.keys())[0]\n", + "ontology_translation = {\n", + " class_name: (class_name if class_name in model.ontology else fallback)\n", " for class_name in dataset.ontology.keys()\n", "}\n", "\n", + "# TorchImageSegmentationModel.eval expects a JSON filename, not a Python dict.\n", "translation_path = Path(\"local/data/nuimages_to_model_ontology_translation.json\")\n", "translation_path.parent.mkdir(parents=True, exist_ok=True)\n", "with open(translation_path, \"w\", encoding=\"utf-8\") as f:\n", - " json.dump(dataset_to_model_translation, f, indent=2)\n", + " json.dump(ontology_translation, f, indent=2)\n", "\n", - "results = pm_model.eval(\n", + "results = model.eval(\n", " dataset,\n", - " split=\"val\",\n", + " split=\"train\",\n", " ontology_translation=str(translation_path),\n", - " translation_direction=\"dataset_to_model\",\n", ")\n", - "display(results)\n" + "display(results)" ] } ], From 2eeef6df3cf550778320fc98b5d495ac1527d43b Mon Sep 17 00:00:00 2001 From: tejass Date: Wed, 1 Jul 2026 20:02:03 +0200 Subject: [PATCH 08/10] Rebasing tutorials --- examples/tutorial_nuimages_segmentation.ipynb | 1757 ++++++++++++++++- 1 file changed, 1684 insertions(+), 73 deletions(-) diff --git a/examples/tutorial_nuimages_segmentation.ipynb b/examples/tutorial_nuimages_segmentation.ipynb index 68086277..f2b435ec 100644 --- a/examples/tutorial_nuimages_segmentation.ipynb +++ b/examples/tutorial_nuimages_segmentation.ipynb @@ -1,47 +1,112 @@ { "cells": [ + { + "cell_type": "markdown", + "id": "eeb1b612", + "metadata": {}, + "source": [ + "# NuImages Segmentation Evaluation Tutorial\n" + ] + }, { "cell_type": "code", - "execution_count": null, + "execution_count": 1, "id": "478f413c", "metadata": {}, - "outputs": [], + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Requirement already satisfied: transformers in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (4.30.2)\n", + "Requirement already satisfied: filelock in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from transformers) (3.14.0)\n", + "Requirement already satisfied: huggingface-hub<1.0,>=0.14.1 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from transformers) (0.36.2)\n", + "Requirement already satisfied: numpy>=1.17 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from transformers) (1.26.4)\n", + "Requirement already satisfied: packaging>=20.0 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from transformers) (24.2)\n", + "Requirement already satisfied: pyyaml>=5.1 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from transformers) (6.0.3)\n", + "Requirement already satisfied: regex!=2019.12.17 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from transformers) (2026.5.9)\n", + "Requirement already satisfied: requests in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from transformers) (2.28.2)\n", + "Requirement already satisfied: tokenizers!=0.11.3,<0.14,>=0.11.1 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from transformers) (0.13.3)\n", + "Requirement already satisfied: safetensors>=0.3.1 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from transformers) (0.7.0)\n", + "Requirement already satisfied: tqdm>=4.27 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from transformers) (4.65.2)\n", + "Requirement already satisfied: fsspec>=2023.5.0 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from huggingface-hub<1.0,>=0.14.1->transformers) (2025.12.0)\n", + "Requirement already satisfied: hf-xet<2.0.0,>=1.1.3 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from huggingface-hub<1.0,>=0.14.1->transformers) (1.5.0)\n", + "Requirement already satisfied: typing-extensions>=3.7.4.3 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from huggingface-hub<1.0,>=0.14.1->transformers) (4.15.0)\n", + "Requirement already satisfied: charset-normalizer<4,>=2 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from requests->transformers) (3.4.7)\n", + "Requirement already satisfied: idna<4,>=2.5 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from requests->transformers) (3.16)\n", + "Requirement already satisfied: urllib3<1.27,>=1.21.1 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from requests->transformers) (1.26.20)\n", + "Requirement already satisfied: certifi>=2017.4.17 in /home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages (from requests->transformers) (2026.5.20)\n" + ] + } + ], "source": [ - "!pip install gdown" + "!pip install transformers" ] }, { - "cell_type": "code", - "execution_count": null, - "id": "a37f27c1", + "cell_type": "markdown", + "id": "13273f08", "metadata": {}, - "outputs": [], "source": [ - "# Download\n", - "!gdown 15Py3AKyGzWNX4M3OJKlQQsnH6IehkoEd -O local/data/image_segmentation_model.pth # model\n", - "!gdown 17eJ6aei6yBAbK1aA4mhTeMdXYycorqcT -O local/data/image_segmentation_model_cfg.json # configuration\n", - "!gdown 1spXt5_ISG1ZaHHO2DTtCSqgkohnhwuzt -O local/data/image_segmentation_model_ontology.json # ontology" + "## Download model from Hugging Face" ] }, { "cell_type": "code", - "execution_count": null, + "execution_count": 2, "id": "2a47f7a9", "metadata": {}, - "outputs": [], + "outputs": [ + { + "name": "stderr", + "output_type": "stream", + "text": [ + "/home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages/huggingface_hub/file_download.py:949: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.\n", + " warnings.warn(\n", + "/home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages/transformers/models/segformer/image_processing_segformer.py:99: FutureWarning: The `reduce_labels` parameter is deprecated and will be removed in a future version. Please use `do_reduce_labels` instead.\n", + " warnings.warn(\n" + ] + } + ], "source": [ - "from perceptionmetrics.models.torch_segmentation import TorchImageSegmentationModel\n", + "import json\n", + "from transformers import AutoImageProcessor\n", "\n", - "model = TorchImageSegmentationModel(\n", - " model=\"local/data/image_segmentation_model.pth\", \n", - " model_cfg=\"local/data/image_segmentation_model_cfg.json\",\n", - " ontology_fname=\"local/data/image_segmentation_model_ontology.json\",\n", - ")" + "model_name = \"tejasstanley/segformer-nuimages-b0-full-bs8\"\n", + "\n", + "processor = AutoImageProcessor.from_pretrained(model_name)\n", + "cfg = processor.to_dict()\n", + "\n", + "pm_cfg = {\n", + " \"resize\": {\n", + " \"height\": cfg[\"size\"][\"height\"],\n", + " \"width\": cfg[\"size\"][\"width\"],\n", + " },\n", + " \"normalization\": {\n", + " \"mean\": list(cfg[\"image_mean\"]),\n", + " \"std\": list(cfg[\"image_std\"]),\n", + " },\n", + " \"ignored_classes\": [\"background\", \"vehicle.ego\"],\n", + "}\n", + "\n", + "model_cfg_path = \"local/data/segformer_nuimages_b0_full_bs8_cfg.json\"\n", + "\n", + "with open(model_cfg_path, \"w\", encoding=\"utf-8\") as f:\n", + " json.dump(pm_cfg, f, indent=2)" + ] + }, + { + "cell_type": "markdown", + "id": "f7f0b0e0", + "metadata": {}, + "source": [ + "## Build Dataset and Model Ontology\n", + "\n" ] }, { "cell_type": "code", - "execution_count": 5, + "execution_count": 3, "id": "444ef811", "metadata": {}, "outputs": [ @@ -49,60 +114,149 @@ "name": "stdout", "output_type": "stream", "text": [ + "Jupyter environment detected. Enabling Open3D WebVisualizer.\n", + "[Open3D INFO] WebRTC GUI backend enabled.\n", + "[Open3D INFO] WebRTCWindowSystem: HTTP handshake server disabled.\n", "Built nuImages segmentation dataset with 50 samples and 26 classes.\n", - "Saved compact ontology to local/data/nuimages_segmentation_ontology_compact.json with 26 classes (max idx=25)\n" + "Saved model ontology to local/data/nuimages_model_ontology.json\n", + "Dataset labels will be translated in memory during evaluation.\n" ] } ], "source": [ - "## NuImages Segmentation Dataset and Model\n", - "\n", "import json\n", "from pathlib import Path\n", "\n", - "from perceptionmetrics.models.torch_segmentation import TorchImageSegmentationModel\n", "from perceptionmetrics.datasets.nuimages import NuImagesSegmentationDataset\n", "\n", "dataset = NuImagesSegmentationDataset(\n", - " dataset_dir=\"local/data/nuimages-v1.0-mini\", version=\"v1.0-mini\", split=\"train\"\n", + " dataset_dir=\"local/data/nuimages-v1.0-mini\", version=\"v1.0-mini\", split=\"val\"\n", ")\n", "\n", - "# Build compact ontology (contiguous idx: 0..N-1) to avoid sparse-index issues (e.g., idx 31).\n", - "sorted_items = sorted(dataset.ontology.items(), key=lambda x: x[1][\"idx\"] )\n", - "compact_ontology = {\n", + "model_labels = {\n", + " 0: \"animal\",\n", + " 1: \"human.pedestrian.adult\",\n", + " 2: \"human.pedestrian.child\",\n", + " 3: \"human.pedestrian.construction_worker\",\n", + " 4: \"human.pedestrian.personal_mobility\",\n", + " 5: \"human.pedestrian.police_officer\",\n", + " 6: \"human.pedestrian.stroller\",\n", + " 7: \"human.pedestrian.wheelchair\",\n", + " 8: \"movable_object.barrier\",\n", + " 9: \"movable_object.debris\",\n", + " 10: \"movable_object.pushable_pullable\",\n", + " 11: \"movable_object.trafficcone\",\n", + " 12: \"static_object.bicycle_rack\",\n", + " 13: \"vehicle.bicycle\",\n", + " 14: \"vehicle.bus.bendy\",\n", + " 15: \"vehicle.bus.rigid\",\n", + " 16: \"vehicle.car\",\n", + " 17: \"vehicle.construction\",\n", + " 18: \"vehicle.emergency.ambulance\",\n", + " 19: \"vehicle.emergency.police\",\n", + " 20: \"vehicle.motorcycle\",\n", + " 21: \"vehicle.trailer\",\n", + " 22: \"vehicle.truck\",\n", + " 23: \"flat.driveable_surface\",\n", + "}\n", + "\n", + "model_ontology = {\n", " class_name: {\n", - " \"idx\": new_idx,\n", - " \"rgb\": class_meta.get(\"rgb\", [0, 0, 0]),\n", + " \"idx\": class_idx,\n", + " \"rgb\": dataset.ontology[class_name][\"rgb\"],\n", " }\n", - " for new_idx, (class_name, class_meta) in enumerate(sorted_items)\n", + " for class_idx, class_name in model_labels.items()\n", "}\n", "\n", - "ontology_path = Path(\"local/data/nuimages_segmentation_ontology_compact.json\")\n", - "ontology_path.parent.mkdir(parents=True, exist_ok=True)\n", - "with open(ontology_path, \"w\") as f:\n", - " json.dump(compact_ontology, f, indent=2)\n", + "model_ontology_path = Path(\"local/data/nuimages_model_ontology.json\")\n", + "model_ontology_path.parent.mkdir(parents=True, exist_ok=True)\n", + "with open(model_ontology_path, \"w\", encoding=\"utf-8\") as f:\n", + " json.dump(model_ontology, f, indent=2)\n", "\n", - "max_idx = max(v[\"idx\"] for v in compact_ontology.values())\n", - "print(\n", - " f\"Saved compact ontology to {ontology_path} with {len(compact_ontology)} classes (max idx={max_idx})\"\n", - " )\n", + "print(f\"Saved model ontology to {model_ontology_path}\")\n", + "print(\"Dataset labels will be translated in memory during evaluation.\")\n" + ] + }, + { + "cell_type": "markdown", + "id": "4e68ee53", + "metadata": {}, + "source": [ + "## Initialize the PerceptionMetrics Model Wrapper\n", + "\n" + ] + }, + { + "cell_type": "code", + "execution_count": 4, + "id": "e34db9f2", + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Torch detection not available\n", + "Tensorflow not available\n" + ] + }, + { + "name": "stderr", + "output_type": "stream", + "text": [ + "/home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages/torchvision/datapoints/__init__.py:12: UserWarning: The torchvision.datapoints and torchvision.transforms.v2 namespaces are still Beta. While we do not expect major breaking changes, some APIs may still change according to user feedback. Please submit any feedback you may have in this issue: https://github.com/pytorch/vision/issues/6753, and you can also check out https://github.com/pytorch/vision/issues/7319 to learn more about the APIs that we suspect might involve future changes. You can silence this warning by calling torchvision.disable_beta_transforms_warning().\n", + " warnings.warn(_BETA_TRANSFORMS_WARNING)\n", + "/home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages/torchvision/transforms/v2/__init__.py:54: UserWarning: The torchvision.datapoints and torchvision.transforms.v2 namespaces are still Beta. While we do not expect major breaking changes, some APIs may still change according to user feedback. Please submit any feedback you may have in this issue: https://github.com/pytorch/vision/issues/6753, and you can also check out https://github.com/pytorch/vision/issues/7319 to learn more about the APIs that we suspect might involve future changes. You can silence this warning by calling torchvision.disable_beta_transforms_warning().\n", + " warnings.warn(_BETA_TRANSFORMS_WARNING)\n", + "/home/tejass/miniconda3/envs/pm/lib/python3.11/site-packages/huggingface_hub/file_download.py:949: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.\n", + " warnings.warn(\n" + ] + } + ], + "source": [ + "from transformers import SegformerForSemanticSegmentation\n", + "from perceptionmetrics.models.torch_segmentation import TorchImageSegmentationModel\n", "\n", - "model = TorchImageSegmentationModel(\n", - " model=\"local/data/image_segmentation_model.pth\",\n", - " model_cfg=\"local/data/image_segmentation_model_cfg.json\",\n", - " ontology_fname=str(ontology_path),\n", + "model_name = \"tejasstanley/segformer-nuimages-b0-full-bs8\"\n", + "model_cfg_path = \"local/data/segformer_nuimages_b0_full_bs8_cfg.json\"\n", + "\n", + "\n", + "hf_model = SegformerForSemanticSegmentation.from_pretrained(model_name)\n", + "\n", + "pm_model = TorchImageSegmentationModel(\n", + " model=hf_model,\n", + " model_cfg=model_cfg_path,\n", + " ontology_fname=str(model_ontology_path),\n", ")" ] }, + { + "cell_type": "markdown", + "id": "16b24b8a", + "metadata": {}, + "source": [ + "## Inspect a Sample Prediction\n" + ] + }, { "cell_type": "code", - "execution_count": 6, + "execution_count": 5, "id": "47f1727b", "metadata": {}, "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Predicted class ids: [1, 3, 8, 11, 13, 16, 21, 22, 23]\n", + "Predicted class names: ['human.pedestrian.adult', 'human.pedestrian.construction_worker', 'movable_object.barrier', 'movable_object.trafficcone', 'vehicle.bicycle', 'vehicle.car', 'vehicle.trailer', 'vehicle.truck', 'flat.driveable_surface']\n", + "Dataset label ids in sample: [0, 4, 9, 17, 23, 24]\n", + "Dataset label names in sample: ['background', 'human.pedestrian.construction_worker', 'movable_object.barrier', 'vehicle.car', 'vehicle.truck', 'flat.driveable_surface']\n" + ] + }, { "data": { - "image/png": "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", + "image/png": "iVBORw0KGgoAAAANSUhEUgAAAxoAAACqCAYAAAAqYjm6AAAAOnRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjEwLjksIGh0dHBzOi8vbWF0cGxvdGxpYi5vcmcvJkbTWQAAAAlwSFlzAAAPYQAAD2EBqD+naQABAABJREFUeJzs/XmQJdlZ341/zpLL3Wuvrq7qbfbRaAMJDUhCG0JCC5tsswcgAlu2kVgcBIHlsLFBAWEMDgmzKYwt3tdIIX7ya4ORMUhYI5BkAxLaZp/e96717ltmnnN+f5y8t6q6e2Z6NKNlZvKjGHVV3sy8mefeuvd8z/N8n0c45xwFBQUFBQUFBQUFBQVPIfKrfQEFBQUFBQUFBQUFBc88CqFRUFBQUFBQUFBQUPCUUwiNgoKCgoKCgoKCgoKnnEJoFBQUFBQUFBQUFBQ85RRCo6CgoKCgoKCgoKDgKacQGgUFBQUFBQUFBQUFTzmF0CgoKCgoKCgoKCgoeMophEZBQUFBQUFBQUFBwVNOITQKCgoKCgoKCgoKCp5yCqFRUFBQULCPo0eP8qM/+qPT3z/+8Y8jhODjH//4U/YcQgj+9b/+10/Z+QoKCgoKvvYohMbXGL//+7+PEILPfOYzX+1LKSgo+Cox+RyY/BfHMbfddhtvf/vbWV9f/2pf3g3zp3/6p4WYKCh4lnL69Gne/va3c9ttt1EulymXyzznOc/hJ37iJ/jiF7/Iq171qn2fc4/2X/EZ8vRGf7UvoKCgoKDg+vziL/4ix44dYzQa8clPfpLf+Z3f4U//9E+57777KJfLX7HreMUrXsFwOCQMwyd03J/+6Z/yW7/1W9edKAyHQ7QuvoIKCp6JfPjDH+Z7v/d70Vrzgz/4g7zgBS9ASslDDz3Ef/tv/43f+Z3f4X3vex8//uM/Pj3m05/+NL/xG7/BO9/5Tu68887p9uc///lfjVsoeIooPuULCgoKvkZ5wxvewItf/GIAfvzHf5z5+Xn+/b//9/zxH/8x3//933/N/v1+n0ql8pRfh5SSOI6f0nM+1ecrKCj42uDkyZN83/d9H0eOHOF//+//zcrKyr7H/+2//bf89m//Nq95zWs4dOjQdHscx/zGb/wG3/qt38qrXvWqr/BVF3y5KFKnvsb50R/9UarVKufOnePNb34z1WqV1dVVfuu3fguAe++9l9e85jVUKhWOHDnCBz7wgX3H7+zs8LM/+7M873nPo1qtUq/XecMb3sAXvvCFa57r7NmzfMd3fAeVSoWlpSV+5md+hj//8z+/bm723/zN3/Bt3/ZtNBoNyuUyr3zlK/nUpz71ZRuHgoICeM1rXgP4lITJZ8PJkyd54xvfSK1W4wd/8AcBsNby7ne/m7vuuos4jlleXuZtb3sbzWZz3/mcc7zrXe9ibW2NcrnMq1/9au6///5rnvfRPBp/8zd/wxvf+EZmZ2epVCo8//nP5z3veQ/gP7smn1N70yAmXC8l4nOf+xxveMMbqNfrVKtVvuVbvoW//uu/3rfPJK3sU5/6FP/sn/0zFhcXqVQqfPd3fzebm5tPfFALCgqeUn71V3+Vfr/P+973vmtEBoDWmp/8yZ/cJzIKnrkUEY2nAcYY3vCGN/CKV7yCX/3VX+X9738/b3/726lUKvyLf/Ev+MEf/EHe8pa38Lu/+7v88A//MN/0Td/EsWPHADh16hR/9Ed/xD/4B/+AY8eOsb6+znvf+15e+cpX8sADD3Dw4EHAr4S+5jWv4fLly/zUT/0UBw4c4AMf+AD33HPPNdfzsY99jDe84Q286EUv4hd+4ReQUvK+972P17zmNXziE5/gJS95yVd0fAoKni2cPHkSgPn5eQCyLOP1r389L3/5y/m1X/u1aTrV2972Nn7/93+ft771rfzkT/4kp0+f5jd/8zf53Oc+x6c+9SmCIADgX/2rf8W73vUu3vjGN/LGN76Rz372s7zuda8jSZLHvZaPfvSjvPnNb2ZlZWX6mfHggw/y4Q9/mJ/6qZ/ibW97G5cuXeKjH/0o/+W//JfHPd/999/PN3/zN1Ov1/m5n/s5giDgve99L6961av4y7/8S+6+++59+7/jHe9gdnaWX/iFX+DMmTO8+93v5u1vfzt/+Id/+ITGtKCg4Knlwx/+MLfccss1f7MFz1JcwdcU73vf+xzgPv3pTzvnnPuRH/kRB7hf/uVfnu7TbDZdqVRyQgj3wQ9+cLr9oYcecoD7hV/4hem20WjkjDH7nuP06dMuiiL3i7/4i9Ntv/7rv+4A90d/9EfTbcPh0N1xxx0OcPfcc49zzjlrrbv11lvd61//emetne47GAzcsWPH3Ld+67c+JeNQUPBsZvI58Bd/8Rduc3PTnT9/3n3wgx908/PzrlQquQsXLkw/G37+539+37Gf+MQnHODe//7379v+Z3/2Z/u2b2xsuDAM3Zve9KZ9f8vvfOc7HeB+5Ed+ZLrtnnvu2fc5kGWZO3bsmDty5IhrNpv7nmfvuX7iJ37CPdrXzNWfVd/1Xd/lwjB0J0+enG67dOmSq9Vq7hWveMU1Y/Pa175233P9zM/8jFNKuVardd3nKygo+PLTbrcd4L7ru77rmseazabb3Nyc/jcYDPY9/qEPfWjf50zBM4Mideppwl7D1MzMDLfffjuVSoXv+Z7vmW6//fbbmZmZ4dSpU9NtURQhpX+ZjTFsb29TrVa5/fbb+exnPzvd78/+7M9YXV3lO77jO6bb4jjmH/7Df7jvOj7/+c9z/PhxfuAHfoDt7W22trbY2tqi3+/zLd/yLfzVX/0V1tqn/P4LCp6NvPa1r2VxcZFDhw7xfd/3fVSrVf77f//vrK6uTvf5J//kn+w75kMf+hCNRoNv/dZvnf59bm1t8aIXvYhqtTqNUv7FX/wFSZLwjne8Y19K00//9E8/7nV97nOf4/Tp0/z0T/80MzMz+x7be64bxRjDRz7yEb7ru76Lm266abp9ZWWFH/iBH+CTn/wknU5n3zH/6B/9o33P9c3f/M0YYzh79uwTfv6CgoKnhsnfabVaveaxV73qVSwuLk7/m6RWFjyzKVKnngbEcczi4uK+bY1Gg7W1tWu+1BuNxr48bGst73nPe/jt3/5tTp8+jTFm+tgk/QK8P+Pmm2++5ny33HLLvt+PHz8OwI/8yI886vW2221mZ2dv8O4KCgoejd/6rd/itttuQ2vN8vIyt99++3ThAHyu89ra2r5jjh8/TrvdZmlp6brn3NjYAJhOyG+99dZ9jy8uLj7u3+8kheu5z33uE7uhR2Fzc5PBYMDtt99+zWN33nkn1lrOnz/PXXfdNd1++PDhfftNrvlqH0pBQcFXjlqtBkCv17vmsfe+9710u13W19f5oR/6oa/0pRV8lSiExtMApdQT2u6cm/78y7/8y/zLf/kv+bEf+zF+6Zd+ibm5OaSU/PRP//SXFHmYHPPv/t2/44UvfOF197neSkZBQcET5yUvecm06tT12BuxnGCtZWlpife///3XPebqRYunKzfy+VdQUPCVpdFosLKywn333XfNYxPPxpkzZ77CV1Xw1aQQGs9w/ut//a+8+tWv5j/9p/+0b3ur1WJhYWH6+5EjR3jggQdwzu2Lapw4cWLfcTfffDMA9Xqd1772tV/GKy8oKPhSuPnmm/mLv/gLXvayl1EqlR51vyNHjgA+ArI3XWlzc/NxowKTz4H77rvvMT8HbjSNanFxkXK5zMMPP3zNYw899BBSyqJCTUHB04Q3velN/N7v/R5/+7d/WxSHKSjK2z7TUUpds8L3oQ99iIsXL+7b9vrXv56LFy/yP/7H/5huG41G/Mf/+B/37feiF72Im2++mV/7tV+7bmi0KC9ZUPDV5Xu+53swxvBLv/RL1zyWZRmtVgvw/o8gCPgP/+E/7PuMePe73/24z/H1X//1HDt2jHe/+93T803Ye65JT4+r97kapRSve93r+OM//uN9q53r6+t84AMf4OUvfzn1ev1xr6ugoOCrz8/93M9RLpf5sR/7MdbX1695vIg6PrsoIhrPcN785jfzi7/4i7z1rW/lpS99Kffeey/vf//7961ggi+H+Zu/+Zt8//d/Pz/1Uz/FysoK73//+6dNtSYrk1JKfu/3fo83vOEN3HXXXbz1rW9ldXWVixcvcs8991Cv1/mTP/mTr/h9FhQUeF75ylfytre9jV/5lV/h85//PK973esIgoDjx4/zoQ99iPe85z38/b//91lcXORnf/Zn+ZVf+RXe/OY388Y3vpHPfe5z/K//9b/2RTuvh5SS3/md3+Hbv/3beeELX8hb3/pWVlZWeOihh7j//vv58z//c8AvTAD85E/+JK9//etRSvF93/d91z3nu971Lj760Y/y8pe/nH/6T/8pWmve+973Mh6P+dVf/dWndpAKCgq+bNx666184AMf4Pu///u5/fbbp53BnXOcPn2aD3zgA0gpr/GXFTwzKYTGM5x3vvOd9Pt9PvCBD/CHf/iHfP3Xfz3/83/+T37+539+337VapWPfexjvOMd7+A973kP1WqVH/7hH+alL30pf+/v/b19XXxf9apX8X//7//ll37pl/jN3/xNer0eBw4c4O677+Ztb3vbV/oWCwoKruJ3f/d3edGLXsR73/te3vnOd6K15ujRo/zQD/0QL3vZy6b7vetd7yKOY373d3+Xe+65h7vvvpuPfOQjvOlNb3rc53j961/PPffcw7/5N/+GX//1X8day80337yvUt1b3vIW3vGOd/DBD36QP/iDP8A596hC46677uITn/gE//yf/3N+5Vd+BWstd999N3/wB39Q1OMvKHia8Z3f+Z3ce++9/Pqv/zof+chH+M//+T8jhODIkSO86U1v4h//43/MC17wgq/2ZRZ8BRCuiGEVPAbvfve7+Zmf+RkuXLiwr6RmQUFBQUFBQUFBwWNRCI2CKcPhcJ95dDQa8XVf93UYY3jkkUe+ildWUFBQUFBQUFDwdKNInSqY8pa3vIXDhw/zwhe+kHa7zR/8wR/w0EMPPWqZzIKCgoKCgoKCgoJHoxAaBVNe//rX83u/93u8//3vxxjDc57zHD74wQ/yvd/7vV/tSysoKCgoKCgoKHiaUaROFRQUFBQUFBQUFBQ85RR9NAoKCgoKCgoKCgoKnnIKoVFQUFBQUFBQUFBQ8JRTCI2CgoKCgoKCgoKCgqecGzaDv/pVL6NSqSAlSBEgpUKpAKUj4kqdUrXCaDRiefkW6rMVynEZJSAIFULAlfXTPPDFR7j77peyvHyASqVKuVQiDEOsczhnMcaSZQYhfOdZZy3WWjJjcA4QIAQ4B5kxjJKMwSglMxZhHQ78Dr6Jtf/HwV4TipMgbb5FiPwxf6xwbt++0yebnDDfdy/O+Uf3bhXTPfOzX2WDEQgcNj9W5ns6nDVsXjrF+sZF1g4eJh332dlpMje/yLGjRzh75gQb29uU4hrPvesOgiDgyuULXLx8kXazx+HDx2jM1FlaXqLf7yMk9PsD0iQlS8ak1jIe9RkOeqTWoVLDyKSYLMOmGQYHzmKtw1qLFAKEAgEOgXQglAAhEULgnMVZh5D5PQiHSy1OOLCO8XhIGAbUazWSJKXV6SOV8K+tAykV/syTsZQI4fa8fs6/B5zDOotWAUJqcAaHIAxCpJJYa+l02rSa22RphhMSmyUsLC6DCrBSI5F0O10yATZJMCbBmAznUhr1GvVagziKCYMSOgiw1iGlREpJfzCgudOi1+9hrMGlY8JyRL1aIy5XGA4NvV4LByilUIFk7cAKSngdb6yj0+ly7uwpKtWII4ePoVRAZjKsc/S6Ce32JpVKmUq1TrefkiUDDh5apVZdwuDIsgyXZWTpmDQZ8gf/7+891p/r1yxf+P1/TZzOEYh5ZGaRmZk+5pTg4QN/wceOP8jHfr8GTuw71jlHoxrhHHQHCaFS/MjLXsxLjh1CCEGoFKXgBj7ShPDv7fycmbX0xwmVKESJ3ed0QG+cYK9jY9vs9JBSsFCt8KWY3La6ffpJAsBsuUSjFD/ueQRMr1Hk/+1lcrxw8MDhP+dK6TQHe7dzx8VXk6kxmRoTpTXGQY9PP+f/od5f4XknvxNpNb3yJp+54/1YmaFsADiMzMAJnP9YwFlQAf5vFEHBk8fhWO5WuevKEnPDEmGmrnldv5zon3j6VRT8So5PQUHBY3MjNu8bFhpCSpSQCCERUiKVRGmFCgN0ECGEI5ARcVkzHvZxJqVeq9Fs7qBkyNnTJwmkIh2P2NrcQEmFlBJjDQiQQmCMw1qDcc5P0pzD2AxjDM4JQHgdISQm250QA9eJzTjcZH/8l/BEePiJ8mSPXbGAEHsmvQ6cu/bL/KoxFXtFTX6cI/8wdHuPc/snB07kx7rJ5eJMRrfbp1KuE2pJa7OFFLC6usL6xgXarRYusxw+dIg4jNje3mBra5NOZ8jy0goLi7MsLi4y6PVQOJrb27S7bVrdNoNeDzPOsGRY63BOoKRGSi91MifQWgEKJYUfc+EQUmEd4AypMQgngcyPr5M4Z5BSADIfX4vDC8UoKhGXIsrVGnI0opxZrAXhDMjJCyby12YiOBROmHy8JMaJfBIp8lEW6CBEK0WaZexs7zAej2l32qRpitYBQRiAkCgdIIQmwSKkINBe4AilMTZDK41zAf3ekOEgoV6rUq9kBFFIGJbAQZZllEsx4fIiUTum2WqSSkGSGprNFjUzRsoIJaUXzBZs5kjTlLAUY61DYAm0plwuoaUkGY8pVQJMkiK0IipFhKMKvX4fIQXlco2xLCNlhJAKiSXQAUZIL0jd7uT86cZc56j/QaT7H3BgxIg+XY7IW5itttnpjKZ/L845tJa85sWHKUWa//7x4wzHGZ88fppX3HaMchh+SdcjhCBQiply6bqPN0rxNduc8+Igv+wnPOV27B7/VLL3Oiq9ZarHVwjFHEQObSK0ifwiAI56f4VD6y9GWk2qh5xb/jROGLQJufXctzCIdzh34G+nizv7P1+Lid5TgcOhrOTOjUVWujW/sRjagoKCZxhPoLytwAmBEgohFUpH6CAmCKvE5ZBea4tyZRmTJYzHQ7rdFp3WNjrUhGHEoJ+wMHsQpRXD0ZCd7W2EVIxGgjAMCIIA6yzGWpzzggNnyazFGjdVTUIIv4rtXD7p9/EBtzeukP8onJsqAR+cmIRFJnv6HfeKib3q7HorJ3snPnvZe317971aYFx9fufcdB+lQ5aWDxKolI31iwzGY2655Q7SdERzZ4fhKOPI0VtYWJil22txaf0S2802tWqN+cVFFhYWGI1GWJty4vgJdppNNra26Y/GZFmKM45aOUBJjcWggwCtA5TQ4BypVSiXj4+zWGeRYqIJ/JJmZvwkNwwDLzqRe0I6AlD+RydAOnQYE0YxxjqkHOCc9ZNnJRG5EPQvk9wVa0JNZCJSWNI0xRiHDhRaKbIsY2tri+ZOkzRLKZfKpKnBZgYrJFoJcAFhFFGOK6AFSgWYxgxpakjGY1rdDoP+AGsMUsYYY2m2Ogz7QxozNUqljDCKCYIQYwxSShbmZomjgO2dHUbjETbLaLV7NOoSpSVpZhCAxfkoXP5eE9JHcRaXl4iiGCG8SK/P1Oj1RigJ9ZkZWi1HvzsAIahWGoRRtBulU/59rlxwQysIX6tkaoyyISKPVjjhcMLSj7dpdS3pg7cymwm+9e4u/+OvTjAcZ9O/nTjULM9X6PTGJKlfYHh4fYN7L13i7qNHv2L3sPdj4UuZF34l5pKHmy/gYGgYRx12SmeZax8FYKd+mkHc5Pkn3gKAE5ZzBz7N+tyDlMdz3HruNdT7B3j4yEf9QkMx8/2ycqQ5w+ygRCYt2haZzAUFBc88bjyigUAK6VOnpCKMKkRRmTAqobVmPMwoVzO6nTZBEJCmCZsbV6hWZghDzXAwQC0ILl4+TxRXqddmphOmJE1Js2zyRAgfTMBhsdZgLVi7KyKk8yvT1ocO/ESd3Z995MN/RbrJv/sEhcvlSX5nuQh5IiLjRsO3YnpN155j37Y8rFKtV9heP0u/N6Bem6NSrXHpwik67R6zc0usrCyTjBIuXrxIt90jDiscPHiQlQOLpGlKliUcP3WcQTJmfbNJZgVKlwiDMuBAGtIsY3tnmygIkSpASkc5iqg2ZtBhiNbaX7d1GAPWZX5FWSg0EmsdWgcImY9zPh7W5YLPGYSQ4CDUAYGOSJVFCIUU1r8aTuSvgvWTGZdPOh3T6JVPbfNRFa0gHadstjcZ5iv/aTpCSIkQ1keppEQHilq9TqVUo96oUS5V0aFECk2Wp+ZZa2jUa3S6HZrNFuPR2ItoJMNxwnB9i0Z9QLVWI47LRFEZIQSZSalUSoThAXaaTbq9Ppmx9PsJlXJEgMQ4n8pljJ2udgshCKLAp53k701nDUJppBIIJ3AYGo06ze2MXrdLoDRZlqCD8m4qmRRIrdAuuqH33tcin37Of+HOM9/GTHcNo1JOHfwk3fIG3XgT9Zm7IasgBCzPlSnFAcNxNo0i3nSwQTnSdAdjhHQIB8e+bsTKnUMYUKwG78Hh2Jk/QXk0R6O3Ot0+DvpkaoRwAisN55b/lguLnwPg6KVvYqa3xsOHP8rG7CNfrUt/xuPTdCG0isv1Luu1HrODEi89e4jAqOJ9XFBQ8IzihoWGksKnyAiBEBYhLGGkSbMxjA2Jycgyh0wNNhtjsaRJSjdrEUUByThjnI6IVEytXCXQGqX8yrWflO2JWOSTXJ+b79Ojpiv/CIyw+YRd5RM52BedmEzmJ4pFiOuuzPm0KTedyOyurvtHrxYU09jJHvGwN2KxV6xcL7Kx/7nFNCIzdXM4y3DQp9VqkhlYWVtjZ/MKzWYbi2JtbQ1cyuUrF+l0+4zGGcduOszy8iI4SJOEUyeP0+0N2dpsIsMAkY6RVmCFRCEQWiIcZJkj0JCNBmQmo9vt0B+NppPjcrlMuVIhjiKk8ilVNjN5dMmLA4l/nVx+r1KAQaBEkIu8POVOK5T2/wqZRyqU9KMgvOgQIvd5OJfnhfsUD5AMhwOG/T69TpfeqIeWmrhU8hN46X0RtXJMqTzHzMwspXKZUqlEGEZeEE0iSs4ihEMpSRQHzOoG5UqZXrtLs9tmPE5QQQBW0O8OGQ5T6o2UcjkjjkvT6IbSioXFeeJSmZ1WMxdkjtQkIDVCSMZJgnUGjQQnsMaQjFN6vTZhFDI7M8/YGjJj0DpEWotWgpnZOVo7TdrtLmG5RRQ3kE5Oo3NWSoR6+q58GpnghI9GbDVOcnHx81hpEEaj9vh1QBAGkqgMOjTc8bIRz1kdI3qC+bWMI89P2L4o+ebvzqj0cuE1EavCIJzyUZNn26Qt/5g5f+DTnF75vxzYfg53nH0do7DL5sxxyuNZ6oNlAFq185xZ+RucMDR6q8z01kiCPjv1Mzhhi2jGlw1HNYn45lNHCI3mUr3D+Zk2RjqCp29WZEFBQcF1eWJCQ/h0F+scg14L5yTV2iwmTcmyEa3WFmtrR2k21+m0WyTJEK1DxqMI5wzJaEwpKnNwZY3G7AwWhzV+Mulz+x1Siqn51+ZmcOf2eDGcREqBkMIflk9Md0WGT12ByeqxFw4+IjLxRXiuEQhuEv2YuAH2pFT5TCD/857jp4/vSYG6HtcaNycekIm4EVgLze1NhoMxqwePYrMxWzs79Idjjt10K5VKzNb2FZo7OzSbbY4euYmlhXm01oxGI86fPcV4mNBu9jhwcJWLVy6jjCNQEid1/jo6ssyihU/nSYUkc94PkyQpQkCSjOkPRoTtNkoppFRUyiXiUokwDFBKEoR6auree//KOaxLcVZOx0kphRSCTqfDcDggCAJA5F4Rb7o2xvp0LWtxTvh7SoYorRkPh9gsxQkfmTAGLBKEplSpUq1UqFQrxKUyQRCjlEApRZqOvTDI78Fai8l8dCbN/42jiGBeU6qU6XS6NHdaZABSk5mM7a0dxtUB9foscVwhiCKkc4ChUaugtfaib9QjcwLlLNIKTGZ8ZE46kKCVojcac/nKOrMzNeqNBl76eb/RYNAnTUdUKnUasw3aO012tjao1OqUywu77xO8P//pyjc88MOEWRkAbSKk82l84HCBN0cLQCvBd37LEZIX/C2yNKZUt8j+Bu6La4hGh1tePGK03eDOh15HdX6GTuUKAIN4m0sL93Lkyt3Mt4/xpbkonr444eiVNrk8fx9WZn7pHDi//BkuLn6eFz38A1SGi2zMPsyp1U9iZUZ1uMhdp99MnCoeOPYRkqD/Vb6LZz5zgxL1UYxyglu257h5e+5Z9C4tKCh4NnHDQsMhfbUWHCYzZDhazSsMBj3CSoVRMiDJMprNClmWTyBFxHg8ZmhHgKJcqXDzrbeyuLyIdY5xkjAaDVBCYozBYrwJF+EnnBOzt8+jwlkJk/SafDVd4HaL00yN2WI6id/1SeSCZGLTEPsjFhORMlkVnRRW2VvJxU5/uz7Xioxc3OSRkL3Rjr1pXP4fR7e9RbfdIi7VmZ2tcunSWdrdPGXqwALt7g7rV9ZpbrdZXjrA/OIcpXKZ0bDPpUvnabbabG/vsHxglbhcZnlunjAMkVLhrMUJQzJOGI5iglCTJgm9bpcRI7IsQ8mAJMuwTiKcxRiJMSlZNmbQ7RLGZYxLqNca1GccIk9nGydpPgO2KATjcYYOAqQUlOIYpRQg2d5q0e5s0WjMM/XKOEcQeN9BNkmfcxBGIWmS4SwgJNYxGUWEEpRLMTONOuVymSiKCIIArTXOGdLUj3UUBURRCYSvfKasyaMvPjqQpilJMkYqRblUIQ5jSnGJnWaTfr+PUwohJf1+wnC4TqNRp1yrEIcVgiAgy1LiOCAIGmxsJNjMYITAAmnqqxUJ56tzIQUqjNAyZDTOyKyjFCpAkCQGJSOa3RY4QbVap96Yodfv0O92KJdmp1GjyXvp6UqUVac/z3WOsrRzO5cWvwjKYFfPo0YlXBbgVMbZi5tU2paDB/wig6t2yZ7/WTYvOz7+/ioqE2yX7+XKLedJ4z7ZWCAkKO04tfoJ4qROZTjP5PV+WjH5WMiFwg1FZxxkesQDax8lOXkYPVMjLY9oVy/RrVxB2RBpNeOwy/FD9zAOewAsNm8lTkrE6hTj+ErhzfgKcLnWpROPmB2W2BdvL4a9oKDgGcYTqjrlcKR5hMFYH4EYDAdkCJABaTJmc2uDer2GkAonEpyzJOOUMHKsb1ygVi1Tq1ZpzM6RJAM218/RbjVJrSMZJ9x62/MRIiMKywgh90Qb8B/CUmCFACcxWJzM00r2IPOytfsn/nLfh/je1Kd8i79PIX2Bld356HS/vSlQewXKXsGy9/GJd+Tqa9kf+RDTClipsSACDh1eZbu5TrvTIwxKHDt6mGQ84sqldVqdDuVGjaXlRWZnGgxHQzbWL7G902Kn1WZ2fpFSOWZ2do6V5QM+7Wg4xtgEYwxBEBCGIZVyhTRJyObnScdj+sMBaZbQ6fRAWLC5sdlZcBnWCTLnSMaWceijU9ZalJJkWZan1ZGLAkeSJERRiJSTKmUaqXQeAclwaO/FEAJjrS/yO01jyysrWXDWIpEICYEOmYkjKpUKcRwSBD66IqXCGkdGhtaaMPKPGWsZJylKOcJAo6RCyQxr8vAAoIMIkxlSm6ADTb0+Q6lUot1p02w2SZIEm1eU2ml1GYxGzNYz4rL3bmAMEkG5WmE86iOcREhBZn3UJIgihJNYLEpLwjBmPO5j0hSjNaPRAIdEBgEyiOl2+16gVWpU5QzlyhxOThU0wl7fP/R0xApDYxRwxQmccNgDl7Dzm2A0SMMRm6KCqdQHHK7aZeao4LmvgHs/HvOe96+zsCaJylV6WyGveP5hDh2sMFhY576jH+amyy9joX3TE5s4O39tw7hFnNRQJnzsCf800nlt2uSNPNfuuf19OuFIgyHbjVM0a+eIkzoHN59PlFZ91bfHKC+73TjNyI2QRrNmj7ARP8jnbv8g8XiG5578dpQJOLX6SZKgj3CCtY0Xsbb5XHR0P/evfoZ2PLrxay+4Lg43LXYAXEe4CRJteHB5k7vPHsoTgAsKCgqemdyw0AjCiLhUJRn3GScJSWpweLOvdJJslGJdhkIxTsZoFSARjEcjrPPlPrudDsdPPMxolPB1X/+NbO5cYHN7m53tK9jMUa7Ms7F+nixzLMwvUW80di9A5JEK4SMrE+PwfhO455qqPHlkQtrd4khXiwZ31WRhctz0x72Rjz3PMz2fZRpZ2SsiJjkvk/SofZe177oFM415YiWwJqXZ7DBODLffdhuBFly6vE6r1UKgWVleYXlxkdFoyM7WFts722xub1Gvz1OuVJmbn2flwAF6vQ5ZphgMDGnihYav4OQTgXUQILUijALiUkxqMur1BqMkIR2PSUcJg9GQsRuRGYNw3sgNxgtN60vb+jr7fgImtMwz4byAMMYghff3KCkRQoGTWLzXA/zkTUgBNjeV44WkEwInpM/Vj6rEcYkoClF7PArGGiAgCDRBqBHWkaUZIFBaEWhvOLcmwWQpWZYCEiG9ByRNEoT0+2bGgs3QoWR+fo5SOabV7NDudgGDU5L+KGE83mRmpkG1aoiiCK0CtBBeXON8PxbnGA2HxFGYFybw91+tlahSIh0nREEIwveLCTTMNBo0W9Dp9EFISuWa97VMQ2uTVLunr0cjkwnCCYxOeOjIR7wfAAdpiDp/BACzeg4XJoTTCfX+kKUOHN/09/rc+pIx5x4IuPBASG9bc/dNx1irz0NPoPoVxsJxYvUvme0eQtkQJyzS+ryziU8EBJkakeoRpXEDEPRL25xd+Ru2GydZaN3C2sbX061cIUjLLLZunS56WJFhpWEQ79ArbXD6ZMZiNMPNlUMIK31amMgYhR2itMokGVPaAMiPVxnnlz+DA45dfik4LxZOHLqHUdjJo6uCSwv3Mtc5wtHL30h5NMejRWmqw0VkeUz2nC9w0UVkaoQVltXNF9Dor/KFW/4/WrXzOOGQVrG8cwflDM6u3cuZuSYT1TSp+lbwxHA4okxz69Y8M8OYTjzmSq1HszQkVbsGDOkEvTDBSoeyxTgXFBQ8c7lhoZFmqV99dfjV7SzzFXM0mCxjYWGF/niAEILxaIiIygxHPZw1hDqiVq/mpVEFw2HCKBnQH3a4eOGsNzIPE6KowoXzZ7j55udRr8/mKUZ70o2QGDepWLTLXq/FhH1VnSbeDAHsOXLvxF9KOfWB7E9r2j/ZuV61KMdU8VxjAp9MDqfHM5lu2OlX+uS5VCApl0IunD9Pvzfg4MohZmbKbG9u0NrZYpQkrK4d4eDBFZJkRKfdYn39AptbTSrlGRq1KgeWFjmwskK326XZbKGkJtAaS4Q2Xhw653yqmvEVmKzzvUyUUiilfMO6OMZUU+qmTpJkjJOxb6bY6/nyuNb7a4TwlcEMuScB60vGIrGpxWbGm7+lpNGooZX1KU6T8ZIShG+MNxk5KcgbqgmcsDgLOi8eICBv4OjQgSYMIpRUWJuRpr5fhRQwHg0YDHoMRiPGwwQhHJnBR08cBIGgFJeISrnJW0iCwPetyNIMISSVcpk4iqhUyrRaLYbjEYQBZBmtZofxOKFerxPFJZTWvgJW3hfG2mmBXpwQSOnvYW5uLq/q5VBKEUnJeJzggDhSzM/UaLeg2+khBJSTOsRV/z5y3ng/+Zt4OvLp5/y/aBMibUCncilf7QWxM488fQs4uPTFGp2VB7jlG8a5ILx6IuZF7eKRlMUjKV/3uiHOCoLTXcS5hfzPXCAvHYLFFgCdymUuLdzLbedfjZEZJ9Y+jpEp1eES63MPkOohN116OTqLOXHoHsZBHxCszz3IxuzDOGHRJva+B2C7cYp+vE2mEpKg70XDmuCik+yYKtpElEezjIM+vfIGQeZTQpUJWWzdzGLzds6u/DXtymXGYRfpJL3yBlZmdEsbZHqcp9T4e02CAVfmH6Bb3uCmSy+jMlygNJ7ZH2ERjvJojtvPvZYzB/8PvdIWwkkWWjcz376JXmmDXnkDJxxxGnPnlcMsplfYWjjDpYYXGdpI5gdlOvGYYZAWYuMGmUQxaknE8y4vc7Q5M41q3HVlie3ygLOzLRLtxcbh5gyL/TK6EBkFBQXPcG68j4aDNPPN3tIkJcsyTGqJIp8zPr+wxq2Lswz6HXq9NqNhnyTpMbZjdCBAOpLBiOUDB7npppvRgWR7c51WcxslA2YaMwyHPd9fQfg+DYjd+IIU3iCcZJYk3RN3EOArYflrtNZN02Kuvv5pVGTv5uukRDlnpyJh3+GPEkGZ6IjrnXv/BUzEyeQoPxUVeWaYsRmbm5dpd7pUq7McXD1Ar9tmc2uTdmfI8soaR9bWsMaXQL144TxbO32CqEytXmdxeZEDKwcY9HpooahXq94HIxxSCNI8wuC7fgNKoZTEOp/+lKUWZ0xuxHcIofIeJyGlOMZaS61Sw1qD0tKfQyik0N4PI62/L+krUkkBpXIZIX2322q9ShjpfBzknpHJ/TR7hdqenidpmuAcu13jlcyjGpIsM6D9hGw8HrKz3WfQHzJKRiSpQeoQrRWBjkD41D+JYDzMaHW3CbWmXq9RLpdw1iGlRufdpZMkQWtNo1GnVCrR6bbZabUxTuCUYzxO2N7aplYre7GhBMbl1biAzGS56T9/D0uJCtS08IGx/r0upUJJjXGGMAqpz9RpNlO63S5xqUOlNoOUAQIF2Kd1H41R1Nr3u8C/1INxQkNnXFkf86efuEzYKLNyW0p19vplePaKd6kAZXGr57HOR8NIAygNoNRju3GaC0ufYxBvc+zSSzEqoVteZxDvsDl7fHrORw59DGB/xSWxG/1I9ZAHj/7ZdB//sJj+qxSAZaQ7OBzd8sb0Lo1KcTgObj6fw1deQr+0Tbe8zihq4yvpZWw3Tk/3v/qzZPJp0S9tcd9Nf0KYVnjB8b9Hdbiwby9pFfPtY6zPPUSvtEWUVLn93LfSLa9zcvWvyPQYgLX1u1nYvI1B/ZN8cfUMY50xMyzx0jOHqI8j7l/e4L4DG4+WoVVwHSpJwGsfuZlSpvcJQOkEi/0Ki/3K/hQ8KMa3oKDgGc8NC41SpYZxY5LeCJNZytUaxjhq9Vn63T46cpw7cz/D4QiEodtpkwxHOBsgGNFpDUlTw4Xzp5BIMneYnc1tRoMRxg5I05R6fY6VAwfzXh2T1W4/qVdSoJUijCJ6o4ThyHcWlkL6ntRutyrP9RaJvG7YNdJe7a2YbLv2XzHJWkEKkaf8uMk3f76P3J38ietNEdgjXsTul82ePYWDQa9Lt9VEOM1NR4+QZWMur1+h2erQmF3g8NpBtIJ+v8eFS+dotttIpWk0Zjh4YIm1AwcZdHtIrRiNRozGY0yWgDGIyWRe5Ok3VuQ9SmyefuQjGkJKjLVkJsNkKWlqAOmPk4IwVjgrCaM4jzrsigaHI3dt5zjvz5D+vH4cJtW5JqLLp2NZ/PndRCfmxn0BGGMRArQKEBKM8REUJSXGZrTbbbrdLqPRkDRNwVnfrTsKCbTy5XmF8J4QCyiJUCHaCcZZxtbmNpVSQKVep1SuYGyMVHkkyDqcywhDzfzcHKW4zE5zh96w570X1tDsdKhmhiiIkdJ3UEdIkjTxHhPpTd/GGob9Ia2dHYKSZmFxGS2Ur7SGT68bJSPCKKYxM0un1WJ7a4NyrUattrhn5frpOzu5/y/3d8SeWTYcvC0lXmnh1vvMP6fNt33dNrUFQ2XGPspZro+r9DC3P5D/IsAJMml54NifAg5pNe3aBWY7h/m6R76Xz932/6Nf2tojKq79u9yLjy489j779r0Oo6jDqdVPcmX+flI9YvKi3kjkYLKnE5blnTsoJY39OzgwMuXEyifYGfcQmWaucxRpFadWP0m/tI1wAuk0iR7w4JGPcWX2NJm0BEZx2+Y8M6MY4byHoDCF3xgOh3SCchoQmklqnruumNgrYAsKCgqeDdyw0Oi0mwRhBAjGSYoVkjDUdDobpIlhe2sTay3JeAwSRsNh/kHrU4SsERjjSMZDmjtXQKRsN7fIrCUMSmgdYV1GakZcunCacrnKzMJSXkpUYPClVIO8U5/Ic5SnKVFub6LTbprT/v4Wu5/u1/S12Oe9mG6d/r8Q+CpCfg8mAmQiHiYBkN3+D1fJjVyQ7I2SAHmJXoEZj9neuER/mHL46E1EccCli+doNdvEpTqHDx+mVIoZjnpcuHiBze0djJHMzc2xcuAAqysH6fZ6KCVYX79Ms7XD1s4W3U6f4TDDCIm1GVqkSDRRFBKGATqMUFqj8F6BzFiMM5D3ttDa35+1hiyPaAFo7SfT1lqkktNGe36svEJweYTJ90bxpZF9MEeg8siVfwv6wZO5/0ZgsbnpVeAgzUWRSadGmH5vSK/XpT/wAtbYBIEvfyu1Qirfz0IKcEimDRuFwzuqHUiLRuJEyCCF/naT8nBIOY4olWvYMECrEKUCssx3qK5UykRRSLvTo93pME7HKKlJEoMUCQKHJMMSkKW+lHAY+smHUgrrHO12m7KJkAtLaAEoSWIc6WhEq7VDfbZGOa4iZubodTsMBn2qtUXAV1jbazR9uvGx/6c2/Xn19oTX/aOuD4IFBnvwAtJIVg4n+R6TP+4bQVz7az5RnkQfjEp54Oj/ojRucGD7LoxMrj3NlxXBTv0MO/UzlEezOOHI1Pjaa38UJqmWcVJnoXULygTXHLo5e5z1/gB19gUsLxsWoxpnVv6aQbQDwEz3EKubL+Chox/Je5p4T8HLTx9hqVfZ7dj+lN3zs4OVTo071xfJpOV8vcsgSCmlAQuDMtVxiCwiGAUFBc9SblhoxKUac7OLSC1Zv3yB/nBAbzDCpoYgiGlublGuRqSjEZVGA2clSiqMAykjjhw5zObWRYaDEd1Om3GaMOyPCXRMY6bOYNij1eqys7PF8vIxFhZazC8eyFfNHcbBeGwxmSOzjkBpwkD7Mrj46kip2eN6uMp8vZuSs9t0b8LV4uDqilJ7cYhplEWISbRj0ml8dzV/NxUrV0B7DOd7zeIuX8I3zpGkjoX5A8zPz7GzfZnNnS2sCzm0tsbsTJ3RqM/FCxfY2dwhSQSzszMcXFlidfUAvUEfKSUXLp5mc2OL7WaHrVYbKRRBEPnu7WMYGUuv1cJZb5iWyhuhoyAgikLiKCaIYkIdgMjLDDuLdRYhNUr5hoo6iNBS5aVbyYWeA2txk3K0zuIDGT7lSQmJkgLnJM6pqUlaCrtPglnhTeNKa8ChyNACtPApNqP+AJMJ4qhOGFaRYtLdHR8VkV4kSaly8ZKf11qE9BWxpAPrcs+PtD79y8Gon9Dv9qlUxtSqFUqlCiqwaO2jImmaoJVidqZBuRTTbncZDHq+0oDUjEddrDVIYREKhMw9PIj8tQiJohiTWbI8jcsiwBpUEOKsoNNsI+uSsFSmJuvUajNMCk/5lfqn7zRwb2Zce0OhtJtuswcuIVqziG4DV+18iZOyq/9eQZuQpZ07cMJyeeE+BnGTU6uf3LeTdNr7i6T3kflPHYd0iiipMgq7T7qJ3TQigcuN3rsRm6tLFu/r4ZPn/5eSOqXxDDddfDn1/sqeO/T7Wmm40ngY8dAagdCsDW7iyuLfcH7573bPJSwbcw/nAsdjpMUKi8yrfw2ClG60+3jB43O53uVKrcdSr8JLzxwG4JPHzvJ3hy5yqNngULtBbRztio7HehvtMyB+WS+7oKCg4MvODQuNamWWtaO3MjdfQ+qUjStN2t1tRBAgtaQ3bNIbCZSDcTJCq5BqtUyz2SJJh7Q624yGI8CRmYys10VpjQw0aZahdEC310MTUCtXuemWm6lUYp/rLyWZMXQHIxyOSikiDgJmqlWUlqTOsNPq0e0PSNIMB0h2Ldd7S89OJre7uDydSOybvz2W2PDbJ4d7cSEsICcCJO9y7fdEOIeFqYl3mnk1Sd8XjigMWVpZpVEK6HebbGysMxgY1g6vsXxgiSQZsn7lAls7O/TGY2Zm5lk9eJCDqwcZDAdIBGfPn2BzY4dut0eSmrwrtkJKjUQhncDkfSmE8mlBxkI6FAwHQ6CPQqAEqEATxd6bUSpXieMYgcJJhxaSOArRCNJ8tKVQ5MEPhBIIoRHSEQYB7WYTi2B+YRHIc8qFb9Tnx9J3CXfOoYSYdkaUwnfUBu9jQAjcRJyZFKkC7CSaBJjM+I7jzk6jS16IWqyFNDX0hz69KjMpo8GYVq/HcDhAWIsQEqcEqRB0eiP6/QG16pBKrUqpVEHrEKV9VMI4QxjHLIQhg35Mq9NmPBrT7fQwJkNKyYyayYWUwtjMG8C1QIUh43GPLDVoZX1FMOeQOiIq1ej2WzTbLWaEIy5Vp/fuG0q6p3Xq1KMjQGW4+c39254Q+V9caw4xLGEXNiBIOXrppRzaeBG90iZbMydI9Wg36oljqXkHRy6/hFQPOb36KdqVy1MREKQlXvjIP+Di4hc4v/yZG7qkqZ/LSV/M4KoUJIHAyf0eD4AwrVAez9CqXrzmnJXhAs899e2Mwi790g79eIfacInSaHY6UsIpFrs30Vk+h+2N6NYzNma8B0VZn0a11TiZp27uPnMmLVuVAZUkpFka8fnVywyCtJjjPgGsf8FZr/U4Pdfk1q15vunsIT5z6BKn5pucnm+irGS1Xeeu9SUaw9iP79WD7CBVhiu1Hsvdqk/FKl6IgoKCpzE3LDS0jkjTPl/4/H1cunSRYX+McJpavcQoGXkzqxkzTgxKBTRmqjhn0do3TUuSFCssg9GAUlCmUq4TlTQ7OzuoSKGUJtQB87Nr3HTTrcw2aigVTLuASwN1UcobsUXEUTjtv1AKFMuzdaJA0mwPGSWpL39L/qV+Ve8Lz17x4Ce8Mi/LOs2DyvfbPezanhlTU+q+iqN7JxbOr7CTG6X3PLcQeVJPvhI/O1MjHXdY37hMuz1kYXmZw2srmDRlfeMKm9tNev0RMzNzrK2ucujwGsNRH5zl9OkzNNs79Hp9jIVKucIoSfApSb4SksBXmdJBiBSKUPnUKKciSpEm1gqtfSpVFGiCSKKCmFBHKAlxGIDywk8ILyrRco8oU74alFKT4cqjIgatJlWl/EAZa320ATfNezfGG9CFtL7CUj7GWMdkSX/SZ8OYEDWZgDPZZvd1KyfvND7x31vriOKQ8XiMdZBUEmr1Kr1en06nw2g08r1GhMNphXXQ6g7oDnrUa3VqtTphFKN1iJYSmyUICbVaBR0GrF/ZZDw2CGEJw4gszRgOBuha6EWMNX5lOo6p1cpkaYINAiY3poRjtlEDBP1+m067iRBQqtZ2084mY/I0xgf4HLfdPSaqPMXLt0mIevg5iEEF9w3/BxdkhFnZj/u4QZzUyVSyL5qw1ThJaTzDsUsv5bknv5Nm7RynVj/BMGqTBAM2Zh9he+bkE/AsOKKkzu3nXssw6nB++dMMc9P3Y5HqAW09vO5jw7jF52/9/0iCHlYapFPoLCbIYgCUDakOF1ho3cIt8ibO3PrXPFRqM/msKY1mmW/dzFbj5HWv4/7lDR5c3vRFKeQkclPMcG8E4QSLvQoC6EUJXzx4hYuNDs9ZX+JFFw5Sn494cHmTVBnOzDXZrgz4uosrrLbr10Q3HHCh0eGvj5zn5q15nn9lmShTTyqSVlBQUPDV5IaFRlix3PfFz5COLUFYIhEJpVKVaq1C2hr5buHG4HCUyiWqlQbtjvdtlCslkmxMNs4QTmGsJQgU5bjCMO6DSqlVF9FBxNrBQ8Sl2FfoUc53tXaWQGqi2DdOs9YyHA68idYqTObTcCqlEIlkq9khNWZiMUbkOfp7cddZGHaTpUgeLZJxdaO9vYLEn8A/nEc59nxZT3ofyPwxO02k8HEXnMWalI3Ll2g3W9Qa89x0+DAIS3Nrna31TbqdIXNzSxxcPsDhwwcZDvvgHGdOn6LVbLG50yQIYuqNKqPhiEjFhHFIHMfoQCOcQ2tNpRyjdUCgFUqC1BqpAv+zlL76Ul6Ryo+T92IoubfpYV79SEiEMOC8uLDG5ib+yTA5TIavEmV3PS7+1fGj4Owk0pMLA2dQbuLx8M8j5G6amnOQGONTt/LQkESQGeubNQrHxJ4+8YhMXk+tFOP8MSll3lVcUy6XGA1HdDs9Or1OXvJXI7XAGEOz2WHQ69Oo1yjXaoSBH0OBL/UchpqZmTobVyTWpmglCcIApTXWubxkr0TpgNn5WS+8pH9vSqmwyQjwYzzTKAOG/qBNp9kkKJWoVueQQuOcfPoHNBzMrRnufNnIB58ET12kRme4RhO7chFXGqBNSGk8498TTvH8E2+hW77COOhz7sDfMozaWJkyzKthhWmZ5Z07CNMKDwZfZHzoJKkekDyKAHg0xmGP44fuQTiZl8p9bCZlbPcvUuw+ZmXGOPLVrIKsxHNOvwFpA44f+hj90hZOODqVy3QqlzmwfRdZ1PepXk6w2LqVfrzFibWPX1cs+QgLmN0Y7BO612c7vu6A4+WnjyIdtEoj7j2wzl/edJpyGlAfxygryJQf+2405v8cOcdd60vcsjVPZPYICeE4O9vCCseJhW0GYcIrTh0tXpGCgoKnLTcsNM6fOU21PEe9FqJ1iUMHbyWqRDSbVzBZhssmOfkaZw1aK5I0IU1H7GxdwgIzMwuMkiEmTQBLFGmszWi2RmjR5dhNt7C8dIi11YO71Yf2rv4rn1pjrURaQZr5hnHGGD+ZlIJASQIt854dfsK7t8SoEG7fefcbTvf4KqaiYuIemDyW7+nYd9xudGPv9r3spmoIJ3Y7YU/O4Bxb2xtsbzdRYZUjRw4ThQHbzXWubFyh2xuysLDE0uIChw+vMBoOcM5w6tRJWs0mzWaHcrVGGPiu2PNzC4CvdKWUmk62ozAgDkPIfQzOWj/J2xMdcMZPeK2Y+F0sIu8xYqzDOgfWgMn8RNtmZPiUImEsUk66x/tKT6lJkVIinY/ggCSzlmmoIR8fh82jO74kpMjFjK9Ipabj5K/D5L03dl8QY3z/i0nlL5sLldxZg3MC65w3vJvMV68yBqUUpVJMGAaUSiVKFd8ZfDAYYpzxvhKpSY3lynaT6nBIo1GjXKqhgwCtNNZYtAoolcogYnQY4AAlVd7U0AtMKQWBDvDJdIDwURylAu9pst43MzPTQAjBqNehublJrTZPpTKfe1GeWDWmr0WalxV/9OsNqnOWV/1Qj6Wj6VNwVgHScmp8ktn5jLoyZBhOrX6CICszjJpUhosc3Hw+8+2bqQ4X6Je2USZgrnN0tySpg0ZvlXjuYTIhqQ9WSNoD1ucffALX4hiF3elv0mqceGoqORmZcHHxiyw1b+fmi6/kgWP/k1QPEQh6pS1Orv7V9M9K2ZADW3fx0NE/m/bmKHiqcWxVBpyZbXLH5gLL3Sr1UcRnVy9zfqZNf8/7ALyQy5TliyvrXGx0ecm5NWZG0fT9Z4WvYnXb5gKHm41Hf8UKL0dBQcHTgBsWGv1Bm9nZJWbnllg+sEqaDImikEuXThGXqoz6fbSKsSZhNB5y8dLZvEOzzRu+GaKwzWA4IEkdg0Gf0cjns1tj6HS3abbnuePOF2JcyijpE7iIWER7UpX8tfjJs0DnaTvOOrIsQwmFlJIw0owzg7VMJ3TT8rbXWTi9WiDsbQco8h3sPnO5u0aeTJOjpkbxiQ9jrxncTX7cY1b3wqfX2WF7/TKIkGNHjjDXqNBut7hy+TI77S7zCwdYWpjl0KGDjMdjjLOcPnGc7VaTTqdHpTGbd6mWHDt6lNnZWTqdNsPh0PdzsH6VX2vlc7StBXwpVt9jbpyXu4XMOrRwGCzWCawFazOUUvnycz6IeSRB7EmB8qOtUFKhlPdUKOsbAcIk0iDQ1qKkyoWfH2/rvBHXl/jyz+Oswzi/L3n6mXMOZ71A8BIwP976/hQSNRWXNi/dOynBK5xhOILUSIwxJGlKmmQ4l6GUIIxCZvUM5UqZfq9Pu9VmMB56caU1CsdwlDEabVGtjmg06sRRGaV9Z/JGo46xmb9Gk/qf8+vzqWCW8XDIdmubMIxYXJj3wmri9naWJBkSBBG1Rh0pBeNkyLDfoVqdx3F1bO7pxeRv2BpBv6U48tyU+uL1e2V8qaw9Z4xQu783a+enPwsnCYyf1DV6qzR6q7s7ir3/OuR8C6PGPHD0f3mvxXWiDdP7uupVObTxIg5s37Vv24Wlz3Jp4d4ncWeT6IZha+YE241TSKumTQQnj0+uP0wrOGF5+MhH9pm/C26cxzLp793mhOPk/A63bs+jjaCUBtx9fo1jO7N8du0SnavM9ZNjtssDPnbrSW7emuemnVmiVGOkZW5Q4nmXlx/Xo9GJx95gXiiNgoKCr1FuWGgMBwN2di4SlQKStMbZs8cRIiSOayitGHY7CC3zBnD+GCE0FkNmDUprxqMRWguCqEQ5rjMatUmSEZlxOBLOnD1Dr/dhbj52J/OLDZTUVMpV1lYPo/TEOCz2pNPkQiOPKPh8fEGlVEJJTX84ZpSm+6pKTasTTeriOrvrzp5Og/eICOfzlWVuFp9UrPLf5XtSDfaIlWnn5l37AdOmdNMyh5MHfPZ9t91CyZAjN6+xtDRHv9fl0uWLbG/3qNQWmJ2d5dChNZLxmDRLOXnyEXrNPt3umMbsElEYUa6UOLS6yuLCHIN+D2NSrEnJRmMSY1FCME7G+Yq/nE6+pSRPgVJI6VASlNJoMX0hMdagpb9vY/PxNt5Ij7U4a0isxRpLAuAMwjqcMxjjoyDOOrI8VOUsOJshnPUFkJ3AYlEWLJnv8yHkNLXGD7WZipDJGE/W9i2+HK40PsdcOl8aV7j8PSJ8ZEMKSZa63Lzp05mCIMI3/xuRZT6aE0U+MlQul+n1+97snfgO3joIMJmg3e4zGIyYaTQoVysEOvBRolxcWufIshRrrBdFzpva08yws9OkVikxPzeHEBYhvShNkoTm9g6VSoVytUatWiMYhlRqC16sPs2FxrEXjrl8PCBLBDPLhhe/uU9c2etferIIVLD7835XlmOhdTOV4cLjPp1AcueZb+PUwU9xZf7+G6o4VRsso7OYZv0s3fI6h6+8BOEkSdCnXbnETu3sk7mxPde2+xliVbZ/22QfJzi49Tw2Z05Me4UU0Yz9TKLLwHV7G4Efx9V2nW40pl0aXfccExJtONdoo/Jzqty78drjN/PIwhYPLm+Syd330eR7Y6QzHlje4PjiFnGq6UUJYabJlJ325Xg0tssDKknApOx3QUFBwdcaNy40RhmtTo/B4CEunjtNrz9CBQEOR6gko3FKhETIvCOyk0gZEMmAZJwglCTQMf1hh8bMDOVKxGisMVb41W8ko1GPnW3JwYNtHnjwDL12l4W5JRYXD1BS0T4xAbtVhaz1+fZZ5qv9BFoTaAXCkfYypjpD7hq8YRKtYCoInPDpPGLPTG6ScpTHHvxCe37cXoMu4Ce3uWCZNvUTPl1ncs1yzxFuEt4QguWDh5ipVViabzAaDTl/4TzNdpv5pXnmGnMcPbKGSVLSdMyJE4/Q7vfo9wfMLS4QakWtWuXokTVqtRqtdotut4u1jjCMQCiCPJogASdUnrNuSU3mhaHJsGSMUkdiLSJLkTYlM46xwXdcNikC4ztjO+dTfYwDZ/JBsQih9335CgRIX0EK67DOj7HMhZ7D+p59wvsrrFA4J3AyF4ba+8D9cPqJtpQKoRyYSbqVw0qFBKyyaANGgrBeRFkJzkhvMM9TE9IsQzrHyBgcfaIwIgxjgkCSpinOGKRUxHFEGIZUK2V6vR6tTpdxmiCUQIsAi2V7Z5v+oEe1WssrROX5X9ZHTbxBffImdCitCFVIkmRkziDxaWtZZtA6xDpNq93GYqmWa1QqET5otKfR4dOUN729Q3tTkY6hNm+Jyv51/krhxI3JNIEgSqrcdu411AZLnDh0D/ZxxMYwbEPUBKBVvcCp1U/SLfkO5FZmX/EGeNuN0wzinS/78zwZ6fvVED8TgaGd5I71RWpJyLmZNhcbnevu3y6NGOnsuo8JJzjU8ulNFxpt/vrI+X2PL/UqfNPZw9y5sURgFA8sbzAKsul9OxzaKBb7ZVqlEb0owQrH3KBEYBSZsqTSEF/VbXzCcrfKNBP46fuxUFBQ8Azmxs3gOkAjSZIEZwVBWEJKy6DXwWmNtQLjBDb1qVAqEBgy0myEs46oEqGjCDHyWTvD0Zgwiun3OmQmIwpLKJfinOH4iQcZj4ZIocgyw9bGBoePHgLYLVGbT7aM8ek+1vn8/iwzREojhKAchlB2DMYp4zQjCgLGaYqxe/Ljp+Wi3HTVT+z5PBdCTNNanNu/6jV1eUwnf5OoBVOjtHWTa849GXkEZfK94PLzhjagPDtDlqVcvHCOnY1t6o0ZlheXWTt4kDRLGKUjThx/hFa3xXiUsbR8AKE08zMNjqwdIowCtreaRHFMtVql3W4x6A/oDweM0hSMX+lPjY8sKGf9yr715WWF8ClhUitv0M7FlBACLQUW7aMe03FReWqSnprgpQxw1uRRJm8ul0qhhPfSmFwoTibOfvE/HztAOUXm/PFSCB8VkRLjDBJfulZKP342ED51RHj/B0LgjMVog0YircNJ35dDGovLz2ddgrUGgUQ5gbGW4XDEaDgkCELiUhkpAjKTkCQpSmniOCKKQiqVKt1ej1a74/txACLwoqHdalOtVtEq8r4hIMsyppEYryzReVnnLBvn0SD/vk3TMTqIiUsR/YGh1e5gnaNWa+Tld8Xue+tpipC+G/gT56lJSN9qnOLw+ouRNiAJ+rtlbLMYZULynMbp0ygbsLRzBxuzj9CuXtxXrWrv9QgERo+nvznhuDzv06R2BYaYTsp3kzQfPzXnS8EJR7e8/pSec9/591x3nGnm+2WiTDM7LF332TJh2SkPp0LPCEcnHjPWGUb6XkhXi8Cn4rqn4+0EUaZQThKnmqPNGZZ6VWZGMRbH+UYH6QT2ajEofCWpx7qeS/UOToCZlOV2Yno/G9U+H7vlJMpKBmFKqsy+6wKYH5T45tNHGCtDP0r4zNpF1ms9PnHsDCOdkWjDq08cozGKr3nuB5Y3ONipcbBTf9JjVVBQUPDl4IaFhsWSZmNEPukOdMDEbJwkGVqHfpVaBgTK54w6B6ura1y6cAUlNXEUY6t1klEfazLKlQqlUoOZmTpSSNJ0hFYxaZaQZgaTpZRKdaSWWGf8wvkkwiAENjcmZ85gjfWTNufIMonWGq0ktXKJUGv64zFaKaSwZKn3U5A305ukYKnpuXfFw8RwLGFPCITdZHPyyIrLk1qmE0r2PT45xgmNsxaT+UiMEyCkIFQRo+GIi5cucGVzm6g6w9LSMqsHD5LZhNFowMlHHmGn3cJawdKBg2ipWJifY/XgQayz7DSblMslNrc32dzcYjxOEFL4CbHw3bmRAoVFsXt/0/vJ71sp5dPTctEkEQgpyQxMmu/hfIqRwX9pKuG/pKWSmMwhJyVppe8CLoWvYGXytCLhACtzz4r/cjfOwZ5VY4fv4K3y5/MXaHFOYZVAZgKnDNpaH8FwPu3LTlb+lENm3hyeCYu0vq+BUhopcj+78eLWv86QJGNG4xFhUCIuxQRBgDGWJEl82d9yiSCOqJQrtDttur0uxvq0NCUkaZYwGPQZDobU6lXSLMVal/tJHM75ldFKuYwUFQajIdVSSJqOsMYgA5ipN5AI+kNBr9NDOkulUqNUbgC7IvvZg4MkRPRruNntJ3WmVA9x+En4/Tf9Can26TDV4QJHLn8jc50j+1eOBYRZieee/A66lcvs1M8yitpUhguMwi7DqEk/3p6mMO295mtFifdNACRBH3BIJykP57HSMIpa+QT1yTUG9Jf95X+PrHRqrLbrLPeqNEaRf97HSOHZKyQcvn/HIExJpaFZHmGFY6wz1qs92vGIVF3bg+RG2CsuSlnAgW6VtVaD+TxKIJ1A29165ELAN507xIVGm4cWt+jE4z0lfh8HAUbt3leYKV58wft+Pn3oIqkydOOrO9ALlBXctjlPYBXnG23+z5HzHOzUWOnWuG1zgc+tXuZy3ZvIlZVTEXM1a+0G1SS47mMFBQUFXwvcsNCIQkUQKNLU4JyjWqvS7Tcx1mEwhML7HWYai0RxyNb6ZYQQDAZDDiwfZmGpyn1f/CJRoJBaE8dVwiikZlMGvRZBEJFZQ63eIG0PCGRIqRIShJqN9Q1mZ2f8BFFKbyzOowXWmj19E3xKT5JlGOcIcrERR4GvRJVlSAJcIKYej0kKj8/X373fiZnYWwp8tATrKyBNPCFusiJv8hV2a/NT+XKsMj+/n6jn/gbnownWTSpVeYGTZAkbV85z+coVyuUqB5dXWFlZxriMwXDIiZPH6bRaKB2wsDiPVpqlpWWWlhYYj4aMxglRKeL8hXNIKWg0agwHQ1/SVU786F4yTUzTu2Vf8zLAwt/jxFjo8pU56x0Q7B0d8CLDCh958GMlvB5w1g/mJNLjvJgSAnReftZPhQzCOe9PsKCEwxp8pCkP9zjnK0xhBUYYv01ayHwjPrLc+G0dRjic81GKjMy/DiKvvGW9n8TmkROhfOpEKtPcH0H+PlBIvLei0x4RBAGlcoUg8FEKaxKUUpSrMWEcU6vV6Xba9Pq+f8k4MXQ7A4QwWGdJRmNGoxHlSs2LVyXRSrGwMIdSIcZlCOmIo4hxkpBhUUoy06iDlAyHQ1rdAWG0SaU+h1bx0zqi8SWjM1y59+TPIyzjsEeQxdx27rWcO/C3tKoXaVUv0jv2YV5w4i37DeLgxYYpMd++ibnOsX0POWEYhV0ytT9/P1MJvfIG3fI6g7jJIGpiVHLNfg7HKGoDsLr5QoxMn7Rh/CuCE9yyNc/hVsP/vkeYPRr7mxZCaBXhSIGDxX4lP63DiiUGQcrlepf7DmwwDK6tSLb3XHujA15EKGaGMUebM6y265TSwPereJTrFAjCTHFse5a1VoO/W7vEmbnmtdGNGyBVlvuXN6gkIaFRedf1q1JJ8Z+b67U+Lz99mDs2FmjFIz67dpnPH7yMFd5ntnf/R0tPO9CtXnM/BQUFBV9L3LDQ8BMwhVYOkyVsrm+DMghpiVWVUrlMmmVkaYoQvnSodZIYjQwcJ8+cRCiHERnOwnDcYWe7j1VgTUqovd/DZClRXKNUjjA2JQwUFy6fZX5xgagUUC3NILXLzdl5VACf3iRyT4Qx+XbrsErvVjWy7DFA71nNh+mkFucweeoPzpuNbZ7e4pzD5L6QvUmxU8HhJpN55/tAWIuT3rfijMNZ66s8Wb9aJpSf4SZZxubWOpcuX0YHMQcOHGB5eRHrLIN+j+PHH2K71SWKSszMzRHpkIMHlpmdm6Hb7U7Tti6cO0e5VEJpzXA0RAfap+4I4TuX40MUE8/DrkE9T1zKPSMGg8krQInc6J26jMyCtQLhTF4a1mJthrWWzFhv9DYOMDiTgXMorZmfn0cj2Wxu0+x089K3jrycFWDzFCyLMMILMWEh7xkhpJw6wif/U15mYK30p/AKZ+qJFMJ7fwLpIG8kiJJIBEEQcuDgIpGKfaTM7b6evrjA5D0vyIyh3W6hdUCpXEYFAdYYXOojI7VqmVIcUhlWaXe69Pt9rBKISZqZ0ugg8OcX+FK5WvmqSM4hpc5TyLynSSuNy03ftUbVv0f6kla3R63bZGb2wDSq+OxBeH9VlLB/RvXE06nGQY+/u+MDPpUmrfLck9/JybW/ZHPmBJkec2nhXmqDA0h7nWo/4tpIgXCa8nj22idyMNc5kk+cM9bnH+Thwx+9tiO4AKNSHI6LC1/Yk8r5tc80QvFk345XTfqVE9TGEdWtkNlhiXMzbZxwhJlivdZnuzyYRhwmfouZYcyBbpUD3SqVJKSShLtRC3ED15i/tqFRvOT8KqvtOl84eJlOfOMlgSfm7nZpRLs0mkZ3VlsNtqp9xnpi3Pdjt1MecLne5dbNBRb7FV5++jBn5lp8ceXKNJIz8QVeqfeYH5QfddwKCgoKvla58dQp60gzQxBotIpZPXSMKxvnKMdLCAuN2jy1+QabVzbIzBiE9yLYdMC5c1cAKJcihsMucRRhsxTjLCbzE91knCKVRKuII4cOc+nyeYzJaO1sU44rXLx0htFgyG23v4BypYzOq/gAuUFY7vFS+K3WWDKZ99gQ0qdaYXHOT6qNNVMRMknrcXhB4CYr9M7upka5q9aVpgbfya8Tn0e+Mo/A5qky5Cv00z4cOGya4oBBv8vFs6cxRrK2doADy4sgHN1+l0dOPEKr1aFWrlOt1yiVYg6tHKRciWm1msRxiZ2dHXrdPrVaFYnI05nyLz6Tkqapj/yYjNRYsizDZhlZljBOM7IsI0sNJkvIMuP7YhgvHsiPM7np3jiB9M5536DROoQwOON8hAPp063yyE69VmO2MU+mYLvZ48LFC97XgPRfpsL3O3HOR35wDiOsj0zkkQ0pQcgAlUcDjDGkaeYb4SGwuU9nGgbJU/qccCjrsHljPQQIA3Ep4uDqMuVqlc76NufPnadajalUfTd6Lzikj4ggc69QRq/XRWlNpVQm0BpjDBm+4WGtViWKY5rNAJOlvhmi8p6Rvf0+nJMY483o4/GAIAiIozIyr/glhX/9MpOhpKJeraKlwlhDtdzIfSrPxIiGf08BeV7dY82i/P1nieDyyYDV21PkDczPJ+8Bo3wqS6YSjh+6h15pY7rP+tyDNHqrzHeOIa1Cm/jadKDJr4/yMjjhSPWIfmkLK1OMzLi4+PnHTAMSCJz0Ubqv5epQk4l9aCTtePzlM7jnE//FXoWFfj7BdoK71i33L29y78oVHI7AKJ53eZmbt+cIrLwm7e1LQVrJcq9CYxTTiZ94WeC9IhIHG7XeNEJxdWTiwaVNSmnAwU6NShJy2+Y8Y5Xx0NJWvi4kuGVrjlu25r60mykoKCj4KnPjQiNLcEGAtYIwFNQbVfqDOqPhiO3tK4yGY4Zpn26niZI6X402ZGlKKS6RpQk3HX0u281LrG9cpjcwCJGB9RWGBBItJaPUsNncZntnB6wjtQmpdQzSEVpEzM5vcTBYhij2C1X5CrRPXXLTOdhkJTwzZrecrYCJSdk6cvFgpxVuJylR0zSg/MyT46+d311vppEfNa1ANTF/iulzT5j03AiDkMbMAgLLgeUDCCHp9VqceOQ43U6Xem2Wer1GpVRlde0AWiq6HS/YLl+6DDhqtbJ/LgfjYY9Ou8U4ydjZ2uD8hcskWYrNUpzdjb5Yl0HuE3HWeTHhDEIKjJNIDNL5FCmEQyHzCl8+AuJsPi0SBqzM05JM3n/Dj3EGlKqVPIKk/fGCXEz4LuA4g0Bh9/Yacb5s1MTfEIchSgkQEiml7yWSGexkdV9IkAZr3DStTpKnWqXZ9LVwOIwNCKOYcqVCuTSk2+nQbm8SBTHlep1GvUGpXGJ/7xMvZm1m6Ha7aK0plWJUEJAZfy9hEDBTnyEdDQDf28WaLI+ATbw6PmWt1+1x+fIF5ubmOLBSzt/LfswGwwG9Tpt6o04UxlQqZYzJkEojrvuee2YgOg3IAtz85mPu55ygvaGozRtWbkmR8vGEyaOcR1i2G6f8c09SWqTh+KGPcdIpgqzMc06/gXp/hVQPGYc9qoOFqSCyMuPi4heoDOcpj+YZRW0G8Q479TN0yldIwj6TzjJf6YpTXw4cjnIS8Lwryyx3q5TT4LqVkJ5SxP6Ju7KSozszHF/YwkrH3WfXWGs3dlOj/IU+qWvaKQ/4myMXaMXXlrN9wgjI1J4oloPDzRlmhzGb1T6X610+dfQcK50aL7h8gJlhzFKv6oUGgHBcbHQ41G6w0qk9+espKCgo+Apzw0JDCIXWIQjBcNTl/NmTVMozzK8u0NzaZHXtKINxE1CApFqto7Sg1+mAswRBCRUESBFw8023s75+kV6/T5aOEXkZ1sNHbqbT3gFriELFaDQgCipEUYk4qPINL/4m4mrI9vYGi8urhFp7wzJiuroNkwmdT5PykzeT+xK8uLDW5SlXwHTyv9sB3P8+MXJOem/s7fw9Oe56kz6/jDURJnuNu3vL8uanQOLQgaY+06Aah2it6OYio93uUmvMUq1XmanPcPDAAYxJGQ6HgODChYvEUYRS/s6zLOPKlcucPHmC2267k0o5YsP6BnBZZnbv21ky7LQfhXQu72WSG66NL9WLdLgs8yVnMdNeFNYoP9LC+cm8zdMo8p4YApBIlPSdyKO4lBvPfbfsNA9jTQzZNjdJ+GBCXv1KWN8kz0GaOoLAIlU+wXPGe4OsQRrh3wPC+aqyuYNdOsD5RouGFOH0bmqYSVBKEoQhYRQiVUCWZiRJxmi7Q6vdplIuU6/XqZR3oxywm26XZoa02yEIfZWqQGvfoNJNxthHWaQkb4ToIxs++iYItEbJgCxLmUhln65nCXRElgqa29s0GjNEpSphXnzB7RGvzzRctfv4O+WcfyBAh5rDz03QjUcbj8nfWZ4fyCRKCVK6PHByVSoUAqt8qedUj3jo6J9z08WXs1M/y+bMcY5c+Ubq/WVSPWK7cZrLC1/0BQZMhJGJb+x3HVHxdBcZE5Z6VW7ems///vnKp+8IiIzyJvqxZrVTn4qMQZiyXu1xtDnzpMZb5hHvL4c4dMLRiUe84PIBbt2aZ6sy4Eq9y/lGm8+sXWSpV+FKrTf9CxcIBmHKuZl2ITQKCgqelty40JAB1do8xgzpdhI6nR0sks5gB4tjMOwRxCWk8yU5h4Mehw/fwsqBVc6dP4WUmkatgRM+wWZ+6SCNbMSli+cwqSWKSkSRpN6YwZgxAkUlnueO5z6PYb/N0aPHKJU1Z8+fZLa2jLOG4XCEDkvTyZ/NJ4JS4iffwk/qnPPpKCYzWBzWXjXhZ7ebt5uujIndNCcHQuyuSk8EwyQiMR2jqZDYza2Y7H+1KNnT0w+EpFGfQcuMTqfJyRMn6DR71GZmqVaqLMzOs7y05AWGcPSHfbrdDuVKKa/uZen3ezz8yMOcP38B6Ry1Wj33YCjSLNvzXF5IaSGQEhJrcFZMG+Mh8pV1MpRTOG3R5BGOyezCSjKXovKO4A4fqTBZXtFJecO+T3vSaB34nwVgcp+HchiT91p2ArBI59OjfLUpv7rve0dYL/ucj2ZY53tuJMYb1Z0QKLJcRygU5D4HctGhEE75195ZTD7pV0qidIAUEql8VSztwGaWQbfPsD8gCJvU6jPM1OtIvdudHCWRTmIzS7/jIxxx2ftjpBTeg0JedtlaRC7WBCCFQuvAd7Z3eH9JXu3MmAQhQoIoJE0szVaLijU0KjVMNs6bCz4zJq37EaB2xf1j4uD050OCCG550dUVfa7CKuTmEnZum07b8Ok/KTNoS573mhGH70qQ6rGXv/vxNvfe/Mf+aYXj+KGPIZzMP8d2J6OZ3l39fqaIiuvzZRa5zlejQoA28tqXxvnSuN4fkotH4dgpDbl3ZR0jHYdbDdS04dHjP9+0WV8enZkdlvjGc4f437eeJJ2+J58aBIJmecgXV67wjefW8kpTVZ57ZRmL4/jCNqVUU01CetGYyU24vV8Yz+S3V0FBwTOOGxcaWtAdtHAmI47LZAb63TbjYR+JZae5RRAFjNMEFYSsLB9mde0QVo544KEOWM04Sfj6F72EzfWLtFottnY2qZSqjMSYIFScOXuaMAiJgjIrB29i5dBhtncuMhwO6Q/bfPbzF2m32szONFkerFKrz1BBE4RB3uvBRzC01ISB9qvZ1vqJs8mQQvoKRj43Kv+8FtPUGGBP9GISkZhs2x+dAK4REXtTY3Z/v5arvR4Cb35uddqcOXGadrtLfXaWSqXCgeUl5ufn6Q96SCnZ2trEOUulUkEifN+NSxc5/vBD7LS7GGMoxxVmZmbpdNp4+7P0Rv0sRarAX5e1u6VmZT45z0u8TtLKlARBgMs7dQshczHi06hA4IRvUGeNIbPghCEOBEoaLAohBVp7P42QEiuET51y4FCTIJL3fUiLkL78L85X4zLWgHC+OpWbdGdXeQlbhc39IV5Q+vKxVuj8Jbbe+I4DkaBk3vQK5VOrpCIIQp9qZXwncSNkLlx8Kt54NCIZXabd2mZ2boHZmRm0jlC5MLX5qniapaSd1KcNTqNjFmPJRYe/J//aW4SUBDpECkhTg9SC4XCEkIJSKaZaq9LvwzgZ02l3EZmhOlOHUg2eRobhLwcOMKnAXFuM6BqyzHLqdJMVbTj9uYgHPlEC4OJDIa/+0S63f+Ojp8dMPB37f3dTkTHd9ixivdbnxMI2N2/PTSPITzVnZlts1HrcfW4NtacMbZ55xIVGh5HOyKTvzTEIUv5u7RLDIEVZyam5Jrdsze+3+uyuKV2DEY6Hlra4Y2MBbSU2Fy57q0Vd04T0SXJutkVoFF93cQVtJWGmAMdz15cAeHhxm79buzjdf6PSZxiklNKilG1BQcHTixsWGtI5stEIIWBu/iiEGRfOnvYrzTKgWqtRr1e4PDbej+BSzpx9EGsNlVKDam2GLMu4dP4MRjhOn3qIpeVDCByNmXmkdHT7HUyWgZOUSiVKkWJpfp4vXjjDffdvo6VGSkW320GgqVfrOJvhnEJJhST1jf+kN+IqBNJ4saEsoLyoMJkFfH+FydeHj0LsFwq5rSMXIv5bbvI7cI3IeLToxd5u5lfj8v8bdDucPPkInc6I+sws1WqVgwdXqNerdHsdhHCsr6+jdUigQ5xzDMZ9Tpw8zukTp0hGiW+Y6AxOOKJSjOh3yawhzdOmpNTIPEXHSQkmr+4kHAKbiy+FwCKcxGJ3twuBnQo0X8VJ5fckncAJQSLzmZ/zK/pCGl/eVyqss9i8BLKPWAjIjeUWgxXSJ105hxO5GHIOKXxHb+ssRgjfWT1PlXJkYPGRFSF9hazJYzYXitILFp8N5zt0W2tzoaEJpPb34fBd33PDjsFHYKzwUYdxktBsNpFSMDuzQJalBIGeTn7A+0qyzNLr9ej322gtqZTLZOkIRxkptBdy+Kpni8vLSCUwFhQCpULAYp2lXCoRKEm71yVJFJ3BALVxhajUIAiuqj7zLEMIePGbB8QVRxA/9gr7eCj47J+X4M9KBBE85+UjshRGPcn86vW7PT/mrPRZjEAwDFK+cPAKc4PStDnf1Cz/FA2Xlb7RXZo38rtS71IZh765nnScnm96w700nJjfoVUaMgxSBAIjLQ8sbzA/KDM38KISB4kyJNpQSYJrhIJ0gpVOdZqC1SwN+fzqZbI9la20lRzo1LjU6OCe5H0KfM+hk/M79MOElU6NI60ZolRPI93Kin37j3WGucGu9gUFBQVfS9x4RANLGIYIEdIdtEg7w2nRpUDFKBGwsb4xXb121rG9tUEU1jiwcoi1Q4cxMuOhh+8jCEK2tnfY3mli7ZgkNYShxmQGpCSzY+67/+/Y2JwnLkX0Bj0Mlkhq4ihCa02S9BknQ6SCzeYlxoMB1UqdY0dvIQokYKYpTAJBIBWhEmQGhmQYs/u9KPIar45dX4XIwxiTlPh9ooK8YtXVY/QoYuLxGqwNB11On3qEfnfAzOw89VqVQwcPEsQhg/6ANE1ptVrEcYyUvsnczvYWrVaTzc0dRunYNy40PsyvtU8x8k0NvZCwwiHy8r4OfIM96Rv3gQ/NOya57L521aRTuhMasPmEPV+tz9PVHMYLCHyJYZULEiEl2ubCRAqElQRSEuYN8AR+Em+tyUsOe3O5yWtXKecwDi9u8rCSFA4lFc4ahPOSCIH3gUuBySv2yPwapPCVyHx5YXBYL6ys8+V3sUiVlzoWDicE1qa+70luKPeGFYlCogVEgaIUhZw4fw4nDI25WSqVOkL41Cwkeelf4/OwhGAwGFGtGd9x3b8hCMMQrfWe958kDAUmHSOdRGAJQs1MY4bxaIixsReBDnwi4LMQNxGZcOg5ecqUE9f4LPZSrju+++famBRUAOFEmDxKCoqz0N7UVOcMOijyVK5Hogwfv+U0jWFMnGkOdGos9itEmSKwcjcK8USHzvn+EtvlAYMg5TOHLpIow3qtR5RpXnrmMPVx5FOr8J/r0gpSuZveJBD0ooQLjc5UaBhp+fShi7RKI1594hjlPVGBQZAyDjIqSYjNU7Iu13uMdbZPkJi8y/dTMdWfVO6yudH7Ur3LiYVtbttc4Kad2X1RnKmTUMAoyKgm4VNwBQUFBQVfOW5YaCSjFELNC573AuK65pGHH2Q4GOAcNGZmWF6Z48EH1gmUQusqh4/dRqeziXUwHg7Z2rqCdWPOnTuJyTLiUpksdWTpiJEYUXINwiCiVq5Tm69z4exZmq0mQU+hVYBIx2RZxghBaB0XL5xmMOqgA02n02Zt5RiVUon77vs7bj52GzOzDawVZFn+pZSLB6E0OsgrJQFRGDJTr+GcozcY0B+OyYyPQExL2uZ59LuVp3YrS038pHmT8eumTe31dVwd7XDWsbF5mf44pTG7RBgFNBoVICMdQbfXZTweUyqVAMdwOOD4Qw/T7fU4duyY785tHDiBFg4nFEpq368BX0lKIFAy9BNUBWQOZw3SOdI8fck6mfeycEg0CoUzhtSmgENLRWocTqQgQArJODd1aPADkI+PVIGPEkjrhY31wizNfIlkH11SedqWT2VSwo91mhovHEQucvJ0LIREy8B7GST4iJTOK1f5/e20jKyvWCXxHckFFiHzCIkRGCHIrJlGaoSDMAx8oQAjpillEoMUAeVqGWf9PuNxwvr6RaJIkWWOfruJGY8plasorRBI4jjE2hiEQElNGEZ5la9Jd3MFk57qzpLfkB8LIVGBwmXOG8OVIKyWGY0SwjhEBZonvaT6tMSLjPamIogc5fpkLB/rGO97iUoOStc8dNUP/u+ys6X4o3/X4Na7x3zDtw8IY0shNnaZpJONtWG95rubn5tpo61EW8lCv8xd60tUx76PxSRKsK9E8GMM51ZlwPmZNgDn8n8BRjrjE8fOoJyc9qNwwnFyYWe6z7T0rlWkyvhmok5gpaMdj+hGYy7Vu9y8M4twPkrw10fOs1UZEBjJ3LBMYCQXGh2fprk3giAg1U9O4E9EQ5xq7txYoj6K2CkPAJjLIzCT8VruVlnp1BhpLy5WOjVmhvGTev6CgoKCrwZPoGGfwDrLyVMPsXBgntGojzEGZ2Fj4yyt5iYLS4t0Wy2UkJRLJQ4efj5nTj3IxYtnuHDpFN1WJzfyBvSHI+JAUirHSA1KJkhRQschYVhCRZrRYIiu1JhfWAKRcvH8BaJSBQf0Rh2yTUOtNovNHMPhkAcefhjnBPXGApVafRqGts7lE1cIpESHEhEECCmoVSuU4wAEzNQjeoMRmzsdBuN0akaecI2hW+z/8hRXGcP3Coz8DN6LkPtCJqk9ywfWSEZj6uUS9WqVR44/TLPdYmF+gdnZBpVyGbBsbq7zwAMP0N5ps7y8TKVSRasA5cAI4XtY5JWetNY+RG8yLzqMT6lKkwyHxGQGrQXO+twfqUNfMceCUJJAhRhnCaVAKkmgte8ALoU3XOvAp6RpjUL4KErerySOQ8K8EV2lWmLU7wCwsrRAHD3PZ03lHg+Rvy5SSh95sXkZWfZWEbOEgTeVg48YjMcjTN7nQ+QpVplJfbREgjUOZ/xKpMGSZg5nLNb65mg2c+xsb5MkKbVahbGJGI0T0iwjzMthWRyhigiFZSwdQRCQjMcEWvl7l9pXlcos/W7HR0cQPuox10DlwhZhGPS7hIHyfUCsT7Ha3tlmMBiwsrJEFFf2FR8YDPogDHFcQihNFMcorf1753EiZM88HMlIoBTU5+3jzvvdJBDhJt6rxx8vawUPfiLm/r+KGXQkJz4TceS5CWt3Po7Z/FnMJF6M8CVcU2U5P9PmSq2HtpJKEhBmmrlhCWUFS70qi/3yY3oc6uOIF104iBOOchLicDTLQ+JM7/l9RDseMQh3XxtpJXPDErODEvODEnGmp8+ireRVp47RCxPmBqWp6AmM4hvOr9KNEqQT1MZh3rdinkQZdsoDtioD1mu9L6lL+AQvgPxnw3y/zPMuL3OgV0U4wVq7fvWAgoBaEvKqk8ewwqGceMrT0woKCgq+Utx4Hw0ExhjWNy/S7m5jMKQJ6ECQZQbjemzvODQh/bTN337mHnCCbGw4cOgY584/iHUJzglUEFFr1DFpjyB2pIklTRxRSSKFQSuJEpryTAWTDLh8+QJh5Fezx6MeUnnzt5IZYJiZPcDS8iHfj8M5er0+/cHIp3rlObZSON9ATfjUGiUlWiu0FozTBGOMr1iFYLZWQYkRveHIr25P2U2fmqRUTY0be/wenl2/h5/z5EZml/shdr+mUQgWZhosLc6TpAm333E76+vrRFGIc5ZWu8X21jrHT5yg1x8QBCFxqUQQBvueT0qNFJIojEiTFBzMzCxgReR7UWhFoL2IEFrlYkShw4AoECglkEL5gIAKfBRBeQGBcARq0hTRiwTrXF7xS+R+Fpd7HySTxCqE9KVdhWRubhE78TS4PNKUm7gnozQpA5vnGPmzWJ/GZaXyHgqcj3RMnOR78mBMZqZVyIzzOd4+Vcu7I3wzRjf9vi6VHXe/9G4y48isJU0zbwR3eQ8MI8hMQmq9OTwzBukMaTLGZpY0s6TW5t3eM2ItyYwldY7RKGU06iKVQ8qAQDu01oRBhNIKLX2vjSTLiPLhcrkolkrQarUZJwm1Wt1Xx1J7ZiLPMpTOI4fTnhmPPgZCOAYdycN/HfPcVw4JoscXG87CfR+P0SF8+0+1WTicEVeKnPgnwuQTLVOWTFmGQQY4LjU6RJlmsVd51JS1yQniVHPz9v7mdKud+jW/+1TPqw53e6TAnj8VgaCcBrspU/ljEkEtiagl0b7zlLMAHKy166TK8r9vOcl2ZXCDo3Cd23KCxX6Fu64ssdgvExj1eG/hPGmV3f4gz74/+YKCgmcINyw0Qm2JAl/tx9gMHQQYkZBl/hMz1hphHU76HPhOp0m9Mc/c/BzLyw3On8l8yocWmGxEOpAEsconloLUJKRDQzv1QsHZjGFv5KsCSUc/GVOJqxgzphSXCKKINM1IM1g9dIyZeoNarU4QxX4yLAQms94fIL1/wDqLMRadT1gTl5FmSS5GPEpKnHWUSwHGGQaj8bTBnRcOk1QqdoUGub+B/Hfnw/rTclUu75GQT8jzqfJUowShZmlpgfF4yHA4ZNDv0WhUAUu73aHV2mF7e4ssyZBC4KxFhSFBGPpVcClRAqKwwtGjR7njzjsYjhOq9RrPvetOpAMnBMb5akciL6UqrJ/YC+WdGkLs7Q2iyeWR/5+zSKkmdzi9r938eH+cNXkn770r72LSf8KvzvloTp5mJcT0WEcuUuSkbK6AicFbek+Gy/0gCAu552T39XM47abXKZwXHFKYaU+Vqc/G7jHvs3vfxpj9USgHxmQIKacGdQtkWYIUejo2DiCzIAXGWjLjaLebXL60TrfbIU16uRgBHShKcUSpFLO4OEugFM5mCBn4VCpnKZfLWGcYDkbs7GxRr9fRgfApb+rGA5HPDAS7t3xjM65S1XHHN418b5YbQCo4+oKE+/8qJiw5SrXJYkAxw/tSmQgPhyPKFHPD0uP61W50uCcT8S/l2Bt8AsCbsvXUM3Fj0TH/2S6Q+GhKlGlefH6V2WFcvKUKCgqedTyBPhoSJyVaR77CUZrg8s7LOtQ4KXFCIpVinI4YDoYkyTr9UouN7fN0el1SmxHYAC3AOEM5UGSZIEv9pWTWYVNDNRSMxxlOBERBiHaOcbdDJkqk1qJNhswkcSlGBpLLF49z/mzK/NwSd9z1YuIgzFeFAZFXDjGS1PmSrlbgV+/VpJt4PtEUAp3HGxyOQDnKYUCSZGTsNlQT+PruSTrmkeMPY52jHAXoMECFMWEQEGqNDkKk8pEDrfwkcjwecOXyJVCSICwRRZo40BgM3V4Xk/jywUk65OzZU5w5dZZbb7+DMIrJ0gQEaB0QhiWE8D6HUlxhbe0wz33+8zi4chCHZdgfYCxkWk2L6EiX97LIIw1Iv6onJt2186iC80lYiFw+Of9KI5C+iV6+TViHU5P0Jv9/vnywQKC8GMDijPXlcZmYu8HafNJv82dwgHAYY6cRCZeLIZdloBROqlzHOXCZ74ci1PT6fbUpA3ltrckd+RK3BsEesSG9+RyfscXk239iogcfERG5uPSbNE54b4uSOt9P5M9n8ziVQimHVpasXObQoTUGgwFbG9tsN7dIM0PmfIpWfzQijmL/HrSWKCqDM2TWoPLyt7qmGQ4HtFotMmuo1hfQamLhL7g+Xng2L2uWb0pvrBqwgzQR2EzkEZCCp4JJxabGKN7v03iaIIBvOL/K/Qc2uFjvkKprGzJOvCECLyy0USz2yxzdmWV2GKOsJDLF32xBQcGzkyewNBqSJI44rqBVSLPbQuMnjFiLSQz1RozDl3PNsgybGUgTMixp5ue2JktJpEZgGA7TPHc1QDiJcRlpNmacjFFBzEyjjJQKFUnkoMdg2EJKTb/bZ6B61Kp1okAzSjQuk/T7fcbDHrV4AaV9fwdn/cq4MWa3KziGLMtX2QX5CrSZrmpLJivrgkBprFJYY3AOTC4yHI7BsM/xR+5lPEqn5vGJ6XwqMHTEsaO38Jzn3Ak4Lpw/w+c//zmSNEXgU5MOr61y+PAhlPSibTDscN9993Hp0kW0DKjWG7RaOxgcxjo03rMAlgMra7z4G17CysEVgsBPfkejhHEyJEsz0sQ3fTLO+tLBeYqXwPpJfB6CMFaQ4SMwznkTt0SgsAhn8nvTvgEdvn+En6iLPKqRr+xbi5pM1nOnvLG+z4V04MiPR0yFj93ThX3SjM/lUSHfcNxHYnxqUf58NsM44cMnkxCTcL5Bh/QCYNIgz7m8a3gelXDWYnLRNEnzks6BsFjrpgb1/PJ9FEGQv7ZexKTOoYTMhUz+nnCT1J7dfaWUVCpVSkeqLC4vsrW1QWu7yTgbY2xGkg0YjoZUy2UqVd+Qz5qEOK4hhSZNR1SqFcIoYjAccOXSOarVxpP5m/8qM5nEf3knnULAwqEMeaPzO+FYOJQxezCjOl8YwJ8KHI7ZYYmvv3CQxX55Nw3oaYRA0BjFfOPZQ/TDhHMzba7Uu2yXfZ+NahJSGYcc6FZpjGLqowjtJFGmrvVVPP1uv6CgoOBJc8NCw2QZUalKXKrg4ox2t4NzBiV86zblBFpqTDaAvMwqwpuUhVMEynfkln4Wj3CSIIxI0xSTjXZXlrWvyOPMGJNmHL3pNpJsyHg4Ztjr+HQoZwl0yaeV0CLsdxFC0u5ughB8wzd8M9VyGSUlvsKP8wv4VmLwJl/pwOVpMs5apAODwTmJyb0Yzvo0GLubO+VTfZzD4OgPuqRJ5lNrcrO8s5OJpiSRAsGI0XDIcNBGRSW6+T1gLAZDZiVJmqCVQGDZ3LhCu9Wk1WwyHv//2fvvYFnSNL0P+30uTdnjr7+3fU/3eO/WYQ12YUhwBCeAIFZLCiAgEEIIAhSUFAqIIRKIIBghigowGBRICUQEFCRBCgiAWMLu7iywZryfnvbXm+PqlM3Mz+mPL6vOvdNmbve0nT1PRPe5p0yerKysyvf93sfURKPpll0gCYilVhiRIUKgU3b5xCc+jjHJsraua46OjpjNK2zTpNmED6uCGR8QcskPTuP9GDUypqnGkmUS2wRuAak7jCrF/gmNEqkIi9GvXufxFTQ1dom6tjxzkiuWXDpetiL5EFsrWnmXyD6mRiAV+i0lLUZCcK17VLKQTY/PjhsQaBshiC6AlO3blc7D0AYAxnbyIGMSiMdWl7Oa0kRJjJ7Y5oTQOo6FpS6k7WdCiITl619OfVp7YSl0O8mBxlqc9ynjRUq63ZJO5wLDwTovXn6BxlZ47yBIxuMJdVMxHA4xRtNUC0yWU6iSuk7NeqfoYJTE1a+fL/5241u/VnL6YcvGGf8DU7nvBzEup03HIvBWMkVW/iA9RyRGsZpYPfaJmoc+1JxMNN4ARCKZU3zk2hlOT3rpxndroS1aPUed8+TtHZ64s82orIgiMqhy1F3uWsvH3/PzBCc4wQl+B+O+G4219Q2kFIxGuxhtMEphHQSpibZp16nThd+H0BaHeSoXQwNREaJDSY1WCm0Kut0Oh/sHBCGRQiORxKpBS5VWxYnkecHDjz+KNmlicePGZTSGzZ1tvBOsr60znR3iXEO1qLh+/QpPPH5Er9NpU8ETRziERJ1RcmlRezz8lu1Kd4gZkUDLkMIRVgXlqvRYFsUhspiN06SEiArHK9nLhPAYIiE27O1e5evVATunzjI9GmO9TVMD0mQhzzOcrXn62aeZzyounD+LVAatczqdgiJLzUGn02N7e4eLFy/xgfe/j/l8jpAwXywYj1Oz1esNGPT61HWN8575Yt7qrtN+ScXq9af3JqJUa6+60mOIVucQVq88SU6SG9RSVnGscVhtrdXcqDQ5aY9VCMei+CV8S8+SYqnZaEMBw91OXctGQ7WNj0oTkWUTECKi3ffl1MLLNq8iHjcqMbSvo80QSQ2EaDM87g1dXJoCLHvLdJtL07G28RCRliLG8b4AKiybjpDoZu1PZx0gUEoilaTX63Hu7FkmkzHz+QzrLN57qtrhDo6oa8fG+pDgG4TMkFlOrkqaxYI8y4/PwXchfvW/6WHyyI//sRlP/vjih96eEHDtqQxXCx74YENTweVv5HgPpx9y9Dc9SkdsLbC1oBy02igRmR9JvvWrJaaInH3UsvOgIyvD6vNxgh8OVgW+fP4GD+1v8MDhWkq1fjcf1mUTizgOAzxpKk5wghOc4FVx343G/v4B29ubCJk463mmcNYnWk2waJ0RJfggVwFuAo+PkiYGdADTFqhrw4I8N7imSXkLQjDsb7G1tUVdW4pOwYsvPkNdV+zu3eHBRx7ioYce53A0oZrXeF+jZc77Pvp+Zkdj+oM+0/mU8WxMaBz7o0O2tncQIq2ki3ZVPkKyl5VADC39R6x4+TJ4QkhTCSElmdb4ELB+KQJfUnTSVaWqKpZJ04m1kwrRpTQ8tI9v6pqJiFy9+ttkRbkSIgMQI9VizjNPP8O16zc4e/YCvf6Qfn9IXvQRMaKlYWO4wyM//R7Onz/DwcE+JjP4aeDKlWsMBgPW19YxmUEKQbVYrATfSiiiTGSlEMQy+64topOGIrS5EcvHJJvZlPOQin/frtwLQstFDjGkZmlFXVrSriJBilb3sGw+IkoogqClT7UEqiiRIhCT6gVoizwpIC4F4Un9IIVAIlKERBQr1ykZZDuxSduLwRO8R0S5nIksX2wbjLfc3fTahVCrljO2Tcey8UxvT7LzXTUkYtloLAX0ywlN2/hAq0EJSfAeQGbJDtc7j3cNCEVR5BTFFiFsYm3DdDZjMhnjGstsOsbWc/qDLp1OH+0tJispO8kgwC8nhu9CbF90ZGWkt/FaMwmW7weAaG1r0++DLc/0QPHUb+R8+9dKbj2niTFNNB79RM2n/+AMpSKLseTr/7RAiMi5JyzPfjHnW7+aCsbumufhjza859NV0nWc4IfCMm9jVFZ899QuF47ezXS/l8FJY3GCE5zgBPeF16DRcBzspwC+ICV5rlksGoyRVJWH4DAxYoMHb1PRKT0dk2OkSvoAHFlmIAZ8aEBoVJbjnWVeHbG/5xisbdPtlKytrXNndw9JBFcTXcXW5hD76KM8/73vsGjm3Ll1lboGIRwRh4rQ6w6YT0YcHO6RZzn9Xi/lKuAhRLRRSCWJITA9mlCWXWKMNE1DZR27t25jG8tDDz1AUeYgWrtSH7D2mIYkZGA+m9HUDUILGmeJMXH1pWiLVpKAOstyup2Mm9enIATO2dbyVSFk5GA0ZnO9h9aGfm8Now1Ka4bdLr1Oh6wo+dCHP0CWZVhrmc3m7O7uUeQZ/W6HjfW1FQWorhuOJuOUKO4a6qpKDVXbTMS4nFosV8VjWxSzsq6NS50GHAuel8/jOJV36faUEtiTtTBSEAirDBNBJEqZRN8CREjOTDEJXfAionwEleYnWkgcAeHS1oNMKebRJ/cpT0SHFMIFEuEhSMeSxKSjwilL9Colk+OIWiO9ILTTHBUFUfr0HvlldqBKwpNWc5ISxgMiSIJcSuElUbIK80qPWRn5rqYyy6ZmqdMgerRKU7oQBK5xbZOXnpnnOXmRs74+ZDqeMDo6orGWw8MRs2nF+toakohUGYlC9+4thP/Q/3mEEMtg+fuv1oIXHO0q1k87YoTbzxtuPW+4/YJmuOM595jl8KZi94peNdT1XPCdXy+48q2MtVOe0W3FZF8So+DLv5wskoWIDLYC/S3P9kWHPqFNvaGQUfDE7W06zX3af73TcffpcdJsnOAEJzjBD8R9NxpaFUQctqnIeh1qW4MQNLYmCEWUmiBJhaZIEwEfBYtFg1K0POgMoTT7o4pup0j/dQvs0T5NXdEsKuaVYzydIKVg2B3goue5F5/ixRefRmE4f+lRdk6dYffmTZ577hmcCygRGPRL6iaSr2U465jNZviQ+PFZnlPVU45Gh/T6Pfq9ASbLsDYFPiltklNUblh75MG0r1K2q/aJKqSlAJ2CvZZ6iLpakOcp2A7NSp/BShieOFhH4xFS9KlrhzaWEH3i/UcI3pMpTZl30Kbkve9/H+PxLTaGA86cPc/ZM2fIixznPUdHR8wXc5CC7a0NpNKMDg7wvmYxb5jOF9RVTV1X1HVNCIG6bkOtWrelVYMhxKoZCAFo9RxSthqMEJaGUG0DkhoVT2wtadNz7nGzF2lQEFrKUArwTvzlKEH5SJBJH4NIxXoWwEqBiqBjBJloaFG0hlRJXkGQAe8FMkS8arn1wSe9hPAIr5EEvIkECzIEvKuQu7cYbG1SmZJZSI1DDJJgPBJJcAInHTEklywvBdIlyp0npImJcgRHK4KPyWa3bThYEsfukRu0k50oCC6gtUKqZTChQmeSWC2w1qK1Tpa6UqCVYW1tjcFwwGw6Z3w0xjnL/sEh86pmOBxgjCGEd28xrHRKXx/dUXSHHlPcn05DSOgMAqPbmm/+i4JnvpgzGx0rvaWK9DeWtKf2OSJNLMd7ivHesYWzEGkfAPqbns/9H0Z010KrGeG+9ucE9wcZBWfH/aRheLce1nj8o9GOg86CrVkH7dVqgnmCE5zgBCd4edx3o5HnJcO1PoejXXSeUy88MUwJKAbdIWWn2/LkI4TkaBQk+OjARvKyQ68zpDMouXVrF+8hy3MIFgLYJtDpd9nY2GR0dEBmOlx8+CGuvfgc4+lNjg5GBBdprKfs9Oh0SqgqvK+obcOdvQnnzz3Ggw8/yu7uTfSeYX1tg6qekpmSRTXl2ae+QZYX9DrrPP7k+xlPJ5w+dQat83bfA0K1yokYidGtivMYI0ouV6Uji2pBVTc0zq2C1KSSrcBZtPqMQIiBslNSNU1yjNKGaBu0SpQdJQzD/gCt9Oq2QW+DD3/kNHmnxNuGw9Eh3nu6nT5FWWBtcrmqqznz6ZjJfJy03S2zCyEQSiZdBbSv5S7tQ6spiHGZz720l01TBikgiERPSRrHdKMMEYRq9Rrtxu8q7EJoKUsxFeGBNI1AtrQrSXtepA5CCUGUARlVmpNIgYiR6H16TFDIGAgGhFMYUvowcfnaIkF5cBJBwKqYhO0h5Yo3wqNqi3cBPTtk3RisNswyRQwShSSYiHcSTcClTgoRI15E8IIgPPjUJHghUlcmklFAIBK8b5u22IYP3qU38Z5FVQPJTjnLc5QyicInFVL6lU6EIPE4pExajrX1NQaDAc46RuMjptMjqmrBcDCgKMrX/4l/m7F/TfO1f1by3JdyTj/k+IU/O25F26+GdLy//fmCr/3jkvlYpXPsrgIvhjTxgHtvv/f3yBOfrTn9sOXqdw0H1zWPfrKmtxHax5xUjG80gohcWTvivbfzd1+z0Z6WtXZcHY7Z7c047Cw4Kio25iWP7W7xwOFaWnB4N72uE5zgBCd4C3HfjYb1NdAj05rx/h5RSIgCYwR52WFr6xRH430CAusCUYKWGdIIfNOwcB4xnxJEYNg3mNxQ2TnBOhpfEyXU9ZwoHFtbmxzuj5lNxpgS5pMaqSXO1hxNDkDCbD7BR8tiNsMYjZKa8fiQ+XyKUoLhoM/Vq08jpES1+3H7cJ9+Z0iuh7jKorMSKUzSWLQMoXDXkujdxXmIaSVfSFASbFNRVRXOOlSUVHWVthPAGI3JskQRA5rGEkSy0Z3XFfO6olAa33r+xujJMsOlSxcock2W5zjrOTo8IsZAt9sly1Iaegr0W7C7ewdnHYt5hSmytOQrYGV+Er9Pp7ii8rSUrlWlFiG2FB/i0owqEaNWwnEQKIQMKFqRh0hdzVKjkDaXUsV1bFM42sfIpQpCQFQK1U4CIoIoFEpEAiJxp4TEa4OKgPREBComylJqVxSyFXt4QHkJUhKETza13oD2uBhQXqeEc+cJTY2bHiGjY72/huuu0ZSpWVFCgQwQLAKVGgqf6FJeBEyUBEnbaLVNEhJPsgEOIjVsyRZYHK+qC0i6l6RTsdYilEQplehPMSKVSK5YxOQqJiVapemMUhohBRsbfbrdkslkzmS6WDUv70Z4L7jxvYx6Ltm9qvEWKOG4Yz2eEi1vaxaSK98yfOkfdqjncqUz+n68Whbc8nTfuuh4708uePLHK6qZ4Ev/UwdvQWdv5Ks8wRJBRJ46tcvOtMvOtJe+Nt4NRXkELwM3BhOe2zzgxmDSfp8BCPa6c8b5dbpNxs60+7bu6glOcIITvJNx341GxLOoLWWvx+HokGgdCNVagrrUZARPJCQRtpAURmKKDK80i6qiaWqidHQ6OcIFhKxobIORILTCOcfNa1fo9gZcvPAwtZ0ym0xpGks1r4FIbS3zxRgBaC0py5SObXRGtZhx88YVpBR4WzGZHGKdBSS1rWmsJZZdyl7JdHFElpUrlykQLSVl+d/djUZa6XchpZ7HGJP2wafQwqZp0FIgtEagiSHgmqYt/BUEjye5O+VGMwjp7xptEFKzvr7JY489QtEp8d4zHo2RUtLrddE6bW9RL6jnC8aTKfPFnG7TQcTQukDJu1Zk40psHpaia+KKghRjWNV0S3E0kPI0lvkVy2LAty4rMjUTSniijMSQEuLjUvB9lwZELlf1l7cLQRASgW93IOCXxze09wMyttoND9FarA4InyyJnVnS0FKzFzV4F0nDjJh21AtUFPjM4gNoC7oo2P7Ih9HBI61jdPMWs8M7NIdHdKuGYmOTw6IkynSUdNREkdzGpJJpv4JAqTbFnEjUaZ8CEUJyMZPRr5o2IdvMDiFa7YpM+9ciOI+3jhACo/1dEIGi7JBlBVJKnPN4FRAGpF4S0zRlmdPt9AjBsVj88G5Nbxc2zzoG20kv4S2MDxTlwOKd4Jv/ogQBH/jpxYrGdHBT88//6z67lzXei1dtJl4NQkQe+VjNIx+r0u8yknfhQz+34Hdc0PpbCkGjPL/5wFUe3dtkZ9KlYzNKq48tYd+hjUejPF87e4tx0Z4zS9OI9lMpOKafnuAEJzjBCV4e90+dyhSChumkxrqAEhJiQJsC6x0mpBXmVJApohDUzuKq1HQoLfEetNYMByWLylI1DcIrUpHvCRG8a4iTBaPDQ5SpCLEheNFy0x3GFDgbKDtm5Z5kncAoRafsMp03hNBwZ/8WyIBrPLkxSA3eBe7cucFkNOKBBx5hc+sMnW6XTlmmFc92X4SMq4I5ZTCQGqoQQQRs9BxNpkymU4SApnHJ7vQuipIxphUFGzq9PodHh/jgsdayqGokDq0LOmXJgw89hDKayWSC1pruoN8WnZbJ5IjFIvH5AZyzCBGQErxP1KYISci8dDsiOSclkXfrirW0l22bi1UxfDf3aUmpYin2Tj9jjDg3x8WKGB1FvoaI+q7V++QGJUWiTHmRKHRCyNUJFkWbVyFa56lVyF1Ax0hUEhEgyojTAuXSRMSrlMQugkBLQaMi0rX0JhkIIoATKAROeURQKAtCeJAQtUTnHTqdkv6Z0+zeuIWta6IPlP0Oi4VlVM9RMaXVRyHAJWvO4JNwvBEB6QU2OvCgpUwi9giuFcmLAEiF962p2VKjQ1j2qceuVBEQkqIzYDYZMZstWCwqiiInzwtiTOYCoqVVGWOQEpSSaHKy7N27/N499xf4wO9z+PDb5Pk1iq7g8Kbm2S/lfPEfdFAmBecNNj2jO4rf/B+63HlRv8Qe+bVCavjI75nTXQtUM8n1pwzn3mPpb6bJ1QneHIj2/7Os4Wtnb6KCxHjJ+qLk9KTHzrSXQu5eKYvibUThDDvTzqrRuBvr85KPXj/L1rzzNuzZCU5wghO8e3DfjUZRGnqlobECrTep5g11XaNkJFM6CYR9SI5LMRJCg876GC1xTUOMqRBfzCyzPCMqgREaLwSN94Q23EwITZnDfHGAmyyorSdERcoliIR6hlEZRg1wscE3nl5njXNnz5B3SvZ299g7PMI1C5rG0+93iFFSV5bFom5zNAzXbt5C6S7raxPyvNMKntsQttiW2G0jE0Jc2aSGGJlXMybjMcNun5mYEULAB9WKqhMdpzA5PjisD8zmszZZPFmlGpMlrQMSISWLxQKTpXwFKSVNUzGfz1MWhnPAXdOVkAhEMUSaqkarNPFYkU3iktKU9iW29CYpj/MmgDQJuet32ucvbSmPEZBxwXy2T2UbyqLAqBkm72K9IbaichGX8XhJ6yKlILaZ37INaBSEJJBui3ARkm2tkxLjabcV0b59NQpsFBQ+Eo3ACYn0qYmJEjwKY1PX5JUnRJloWcpTi0jHC4L11KImREdmMopeFyeSFfG81XFIBDLGNB0JSUsSAqgA0cTUOMtIdEkb4gnpcXE5lQmkSK9lMGCr0WjvXx5bKQQ+CoJIsyZTZJSxz+xohJOBuq4xpqIsczqdLhKNUgrvjyciSsmW5vbuxODcT/GJPxr5yL/xE8wODqknt7jy9Tsc7n0Xqb9Hs2j4p/+vPq5JuRfevZEVZ9qWziIbZz06O1mOfquwbBODjFTScdNMuDmYoIKkYw29OmNYFazPS9YXBYM6BeG9bQ1HuxjzvlunqLTn+nC8cpoDODXtcWrSW35pvvNxcqqf4AQneJtw341GCvRWGAX9rmLY7+JDErO6xtF4hxCSXEnWN3domjm2bsg6PebzWUurSYLqlAqtUcYglSPaSOMl2miUkDgabt6Z0skVSilCiGgtkAF01sX7QL2ok42l1BAcs9mc557/Ls5b6iaiVMD5yGLRMDx9OqUwW0leZhRFl0F/g63t06xtbBKiIwaBErQFYiS4tmyOS9Gzb3UPgsVsjkCQdwx3DuY439JhfJs7HgXOWgiBPFNomQYOSqVk6hDCKiTPZJrBYECRaerFgsViQW2TY1Qi3qTCP9AW88EjQsQ5m+hMOiWfJzG7IAQHwUMIOFun1fs27yKuRIuxpfVAJCBFe18Mq/tpV5CNjDTNEfujEdFL6vkC17VkuSfLB+02WCnRk2heEmLbpUmBJk21kitXontJB1G2JrkxZXkEQUoyjx6nJdoLdEwuUyImYbkg4tIoAR3TiRkViCjJnIBM0EQwXhB8YO/oABcleRToQtEti1UMYeMd3nsUgqBCKyqPRBmRDemYhZgE6TJRp5ACUEiR7FElEYdsm9KkRwkCVnnpSxnMSoDv07SnPXLGGPKiYLGYQpQcHY3Z23cMeyX9YZ9ut0+eFyzHIt61VLF3MYQQmCJj7ewp4BSnHoePfO6nee63vsPf/w//NuM7hytt1OulSn0/goOv/uOSn/4lhzawmEg6g4A2JxXYW427Z1NBRiZ5zSSvuTmYIGKiQJ6e9HjvrVOsLYr0PRjFaxNd3/22vt5zSEC3yfj41XMcdhbMsmZ11253xjRr6FrDDzdrewsQodae2/1JMrm4C4++Tbt0ghOc4HcOXhM7OZLEq9Evk7aTkNaUGmkltnYUZY+NU6e5fv1FRrUnukCe5TTeInyiGDU2IpxFyYgWEiVyNtZSIFltp0xmY/pdTfCp6Dcq6RxCFK3o2VHbGbYJaKWomxl39m8DAqN1W5CDFBLnHOPJPp1On0G/iw+OjfVNHnv0PWQmaycwx7aYonWVok2dlomZz3K9HgTBpWC/prZUC9c2XO2koc1gCDG5Tvk6cDiapFXwNtTNmJSrEKNHScd0OmY2iwTvVhMG0VaoSQuQxMSipUFFAU3dkGUZwTvm1Zyj0ZjRaMRkcsTF8xco8oLbd27z4pUrxBiTw1faw1VTchxg1zZTra2vJAXUKQndXkbEUS0srrEIEalrR541OLufdB2ta9XyGCoh095LyDLDoNtDKc14MmVRJ32BEHI1PXHepW2ESFVZRkdjRJToXK0mOMNhh363Q2MtzgamVYOIEZMZQkv0EiIilEIJgY6Q50mc7kNqUpTRSKXR0qBzTZGVLMXaMVh80wrAVUREmVLhgyDIpMlQQJQBokpTBRFACBLhrm0mWoctgUTKNFla2v22HUernzmeUmVFQVUvUM4DiV53NPEsmjlmNKLb6dHt9uj3U9bK3VOoHxUoo3j0x97Hv/Z//Df5H//K/5vFePaSx4gfouuIUTAfy1UBKnVErZqMN6IqPcHrxT2FugAnAteGY273ZnQbgw6KT10+z6DO22nzSzZwL2ISod/pzSicZlh93/Ney1sswASJCvc+6aCz4PMPv8iPP3+Jfp2/806b9hhAWqy6ORjzG5euthYcxzt70mic4AQneLNx341Gcs7xqxAypWRafW7TlUNMzkEex2Kxy9p6SbfbQwjBPhG/mGJ9KgCDc0ilqBpHlmlscJw+tYWUiaLkqkPKToHKU6HmLDQ+WZbWNonCnY9EJxAyFf5EQa+Xs7U+ZF5VTCYN1ru2IaiQ0jJbNDR1ZLB2FiEVWrfp0y0lajpfkGc5tYvgLP1+t/2yXk41AAI+OLy3VHWNdw4hU+J48kjSiSLU6iN8jMynM4SWZMbQKTWLqsbaBucD3Y5jUc3IjFm6y6aCNYRWyL28hqX9tNYxn80IwTGbz7l8+QqHoxFNs4AQMCrnscefpJNnICV1XaeyV6fmQSyzNARpiiPFKlUbUv5E+mfKQZnNapRKwu4QBCF6lAuEWCPQ3O0OtNSCRGgzDFKjkWd5or45u7LmPUbEh9iK1COLWdLngEAXOjUaMSJoENEzmc2oG8dk1iCCQxlBDBLvbft3BUpKnPf0OxmucSysT/bBRqNVmt6srQ24eOE8kchiPsc1jtu7+yzqJgU8BghRYBQM+h20NEwWC5SSdNrGqZov0FqhY0To9BmQRqN1TllkECPWNse0teThm6huPqTpU0p/pOh0qMfj1Ni0bYkPElc11PUhh4dj8jyj0ynp9X40eeFCCB7/iQ/w0Mffwwtf+h6f/RM/R7nW49o3n+c7/+Ir1LPquBF/HU3H6Jbi5jOG809YTt+d/l2VyNunCeevgFpOi15t+9/f6L3Tqsx3N5aFsFOBo7KGCL/28Is8urtJxxqCiOx35pyZ9Dk77r/spGOvO+PzD72ICoKfeP5B1hcFtXZ0m/vUN71KL7901DssFxx0FqnReCehbTK+ceY2s6zhw9fPpOuKuLfJOMEJTnCCtwL3P9GILTc8RoieGJKLkpQGHyDimC/mlJkieg0CMuMJ0bOxaejUQ+qFZTGfpymFznDeJ6G3lly/cxXdFm/eBubzOXmhkUJS1w1KaUJIfHqldZs+HSmznKppQEKvLMhEwCqBMaTwQC+wtsE6j7MObyWzw0OapqLIDCEmD6RQNfjEOKKTyRRC6B1CJ2/+uNQuhFRQN7ZZpWmHGJPeIgYQIRXNrVtVFKnYdt7T2AZr04wkyyTRgtaW69ee59LFR5BYEJKIXK14x5i4+6PRIdOjMfuHI4LzPPTwJQ4P95mMR4hgMVISlCTThv5gDRkcy13OM41REiE0MfrV/nkfkyVtq5lIK++K6GOyspUpr0BphfCiDTAUNI3HGEFmDKs2KB6nYafjlWhFSmmyLG+F4MuAteNiMUSIXhBj0nP44JHtw2QUWO8QgI+hbTbTVEYIjyAQgyIGl+ItWkq3d4mKFUOaiMR25pEc0gREgVSRoshTYW8DhAZixHsLSRZDAGoXafKMRlimkxlKp8+AUZqDg4PWKStNUpx1SKPo5Bk7Wxs4D3d2d5FK0e0k0wFjUl6KynMy08NkGoSkXtTcaSyLpqbMEh1DLadjrQKnri11UzE+Gr2ez/q7AkIILn34ET7xR36KBz/2OEIKPvq5H+PHfvHnufatF7j97HVe/NLT7F25RTOv26b5B5dPQsD0UPGP/saQhz9a88AHGzbPOYY7HnPnNOq5x5H7O4RzVwjr+2Dsq/cPVZkmWvm712r4XQMBk7zmq+durm6KInJ1/YjPvnCRzVaQ3aj0XV5aw+3+DKs8Xgqc9NTas9+d31+j0brm7fZmzDLLsCpSfs89u/R9Rhp3PdcpjxeR3Om3rQcVwKlpl4XJ0FFSWoP2KhlmnDQbJzjBCd5CvIZk8GVBmgr9lYVqDEkMHGRLxUkhZkKlolMiyFRG1hGofgcfU0J2Yy0mz6nr5CoUvU/FZFMRXIoxyLKMoCMhSmg8QkmapqGQkjLLaGKKa5YEhMpxXlF5gTQ5QtTEEPE+oIyhqhuiE3giV25eQ37pC/zYj/80rqlx3hOJ6GBR5CmlOTdA6zrF8QpqiBGpNN6FFNYnaWk2LSUJgWyPlZSQCUmvXzIZT1FKY52nWjSECGWe0TQVz7/4At1un9CMyIoe2nQZHY05Ojoiz3OapuH2ndtAyuQYDjfo9dc4OBhRFB2qxQwpIsJoup0O6+sDpqMxgtZ+dTmVIQUIpteVLpI+pBVcFRPty0WBkmmKkOWGs6f7VPOaqqqQRJoQCNEhVdIbSKmPh/F3XXeXjlRKKqTJiMGBXB7HdorSFtAuNKsn++CIPiJlamoEydLWuciiaqhru9KvRJGyRYIISH1MW0MENBKhJdELVBAEUoI8KIRKTZTSBqJCZx4fWtF62515b3E+NdPL9zUZ8bYmAUndjVSaEFP+hdKKImunDsMBSmj6vR55nlF2OnS7BVleUBYFWsl2+pMa1fHRlINbN1FK0ytznE3nbjq/2sZVtNqad3Ey+A+EgE/98Z9J/2w/c1JJth86w/ZDZ5IDWm05vL7Hla8+y2//d7/CrWeuvXTG8DITDyHA1oKnfqPge7+ZY/LIhz6R89kHH0QECQebqMMN5GCEe/9Xoah4xUoxf6kT0QnePHy/SYVAMDeWzz90mW6Tvqsr7YgCzoz7HC2doqIg84rCaU5P+j/4D0WwyvPc5gHfObVLrR250yyMe9mHT/L6JZOCW70pe905H7px5nW/3h8KIh2fM+P0eqOIrC0KCqeZqne3vusEJzjBuw/33Wj46BAyg+hbEbbEe4eMus1xcAg8WhWr3IUAra44rVg7D0opstzgvCXTijIrMCbRqJrGoYQmYPExsljU5IVBaIGICuE9PjgWFXgESmbY4PExstbtsr7Wx9ZzrG3olCU+zPAh2ZQSRBKbRwi+ZjQ64MWrl7l+3ePqGUIo+t0Bp86ep9PrJ63CKp9iiVRqdsocow0xCBbOo1utgzKBUmvKwqBJ983nIbkdlQbwNLWltpEYLUJ6jNfMpnO+/fS36eaSTCqqJjCZWEKMPPLIQxRFEgMvNRV5nlOUHYwybO9sExc9jIAX93eJwaOkRCgFUWCDgCAoUihESz1RLKPEBanpcCK2aeaeBoEkMFzvUDUNtfME4am9I3goipxer6SuLLX1SZvhE10stDoFJRSCwGA4RGkBPuWk1E29CvmTbaHtvW9D2FrKRAzINuZPyEQt8jHiQ2oIA61wPIa2AFHJUpZAILSNlcBoIKgU2kiSUKfJk0BKhVKaGAVKK5RUIFPwYyQQvErndJsOp5Smk5dkZc5wOKQoSzbW1ym6PfpZRtZJtrPdskDrDG1kq61PU630efEEH9A6OUnZRY0Ljrws2kZMoHVGPhgg0CwmRzi/SBMfQjtBk+mY/Ajj1WhRSyH5zsNn2Xn4LBsXtvni//B5ykGHU4+c5+DaLi9++XvsXb5Ns6hfss3jTQuaCu48WxLOZMjl4nMUMF5DTAfEl7E1XT6Xu7ZzgrcHAkGjPY0+Lp4jkec3D+56UKTRnv3OnHnWcOlw7ZVX9CM4GfjKuZs8t3mwaiAq49pnvPR5NwcTnry9c889O7Me64vyjXiJPzTmmeWp7V1u9acsTKILRtLXf2lPAmROcIITvPm4/8C+GIjOI5VEqLQqrYRKt7fhZBGVCrfWKSpGiVCtkNqnVWxnLQiRnHRiEg6bXFF2M0KAtWGHRV1h64bgPIuqwXuF0oK6DighCM7RLKqU7aHBOs9iOsL3BFVT42NKbPYuIIVG6oj0Kjn2iIBSgoWt+MqXf5314ZA8B61Lbu/uEkzGhaJESIWUkUzdbSWadAaz2bzVHNQYEUFKtIlkmSZHEjw0ItJYn/5mCEgVaOpE39ICnEjTkdFhhdaKxXyG8BlmOMCGRTKGVZJuf8BiOkNKmVbNURhTpvRwIo21bGpDqQRXZHv85dIO2CNEGzBo7SoXJOVipMbANnVqFKVcWdFKrRAhMhqNGY0cUmVMZ3X7OsC5hrpRKK3Z3z9qG4xADGCtTzkkQiAFnD17LrlQBZhVlnllmc9mWJccn/JMt/uSdDtGqDQVEoJWNXJXhkKikgUfEn1KRaJQ6LYRETFlbbgYUVHS1TlZv8vm9im6RhGlTC5lArrdko3NDUKAxjqcs6xtDLE2uact5guss8yrmsxoTu9s0+2W5LlBaY3Sum2E5KrIWGpTQiC9HslqyhNjxDaeg709tJYEAp3OkPl8QVYU0DZmKtOUp07TTObEak7W6eKmExbVJGV8RImMd+kLfgcjxshDn3iCBz/xHoBVA+sqy42nrvC9z3+dvcu3Obhyhzsv3Eg5JksDhJCa0Um1wMc0AYuR1Iz3JsTOS8XoJ3jn4/ubiAj85qWrOBnInWJ72k30qZcRkENqNHa7sx+oZ4hEZBScnvRX301pB9IEJfPq7etBW43GpGj47YtX2evOV68nzYhBBcGPv/AA/NTbtI8nOMEJfsfgNYjBE08+BhBRtOIykmi5pabE2IYPpD4k0V5abjxtWjjtyq5uA/ygvei3k4+8yDCZpM4UMUhMUVFVFoSmLDXWSryL+Ja/b5RB6WQRKkUqnhGRalFjGw/SE2xqjqQQeJuyHGKw1I3jTtMwHHTITYXQOU899XUOj0Z08gIRBWcvXMBojQue8dGIxWLKi89fZT6bMBmPUVJgvUVleWqAJEgRiB5sU6N0htSGehqxtSUQyApBzxiEVEznKQujW2YYpRJNDMF0OsUjkcogpKSqasqiwMeIMslZKQSPbWo6O6cRTYVAIvVxkGHwEBBY7/F1IDgHwmN0jm0sPnq8cyij0IikNxCR6NPa/+3dMeDp9rtIlfQagtTYHYxmZMpQVZ4o0qQhNxkiRJDLi3RqOqU0QHKJUlIRfOv05AMxRJx3OOtROiViRyJKGZROwumkg7HYhvackWitUCh8TPkURskVb1oLUIDWgkcefYgnP/A+fJP0PaF17oI0EUlGUImadPrMqZYmRUqR94EYAuPxhCwzdLol1vmWErhsMO5a3Y7JfWuZ0h7bqUv6TaC0xOSasixBSJomIFtBSvSJMqaMot/psnc0RQTBYj4jOAteUGQZLnh8ePfmaLyRWE0p7qrohBDownDpw49w8UMPA1AdzfnKP/gNrn79WS5+6FE6wy7f/ZWv8cxvfIuHP1qTlREsoDz+8W8Tdm6DcpxMK34EICJVu5LvZOBOb8YDh+ZeAXlbmB905rywMWJS3J/uRkTB2myIi308JZIKzRyBb6mhb8/5E4lcWRvxpQs3km7lZZomLyO/ffEaf+Bt2cMTnOAEv5Nw342GEDJpD0gTjBCOA9+8S3SqmAYG6T5B4pSTGg6l0oq5EKCVhja/QcqlY1EEIVFSpiJTgdARrUoKU+Aj9HophXw6q3A2UNce60JbwEJlA0JFnIu0RlSo1gUpek8MMomtQ0SrVPB6b5nNFtjMgHQI0XA0+hYERXewxuFkTNPMGI/HzKZzvLMYZQh4NjZ6INNkQoTIokoF9FIzPxh0KYoMpRxZlpFpQeMaskwxnczIc4jBY0yP4AOjacXpYggoauuRqqX0AHVdUXZKGmKiphHw3rE/OkQoQa/sYoNDeZf0M0Rc9BAiuZJ4FWlCcgvLTI4EHJKg27DFdruZ1NQ+IGWkNygp8wzrI5lReOcYT9MERkqBbTxlN8fZJgUACkGeZUiTpgxEidEmkaCEQBLREvKiIBcCEaAsOyzqhmpRI6Wjk3cou+sM+j0GnZzJrKKuLUoJ1vpdolRkyoAIOOcpej2KsgTv6HU6jOcV0/ERR6MRSrjU6ARBZRPtK/gGnWVJ2+GPs1KEFMkNqnU1ijGlrruQjA9i8ByvCYo2YH0pCD2+iL+y9WxEKUF/MEBrjdIK0zTEYJCy1YFECM4yPzzE1RXdQR9mU5oYcVVDaD9P/oRm/ao4pkmln+Val8/8iZ+FP/Gzq9ve//Mf587z1xGzX0d8eUo8imlRYmOfqG1rNHevO9pdf+EteiUneD04tgiH8b5i/1q6zPU3PL89vE2nMZyadlvRd3ovG+35jQeuMmnF/T9omgHQn50jzH6S/dglIoGAZkFPvEARbyeirVjuyVuEtml6enufWrvV99XdWO7SUXmiMzrBCU7w5uM1kDSXEccaEX3ri5QKTilja32aEpeddUh1vFykhEy6DiWJclnI0eZGx5XHuRChnWxIMiWwPiUuaw2aRFUpc0OvU9JYR1M7FvOaRW1ROrnzNDXUrqFxKQdi0ClY2EDlAsGn4pF411IWaYU+YtA6kBmQMeCcZT495IV6QjOdIUSiFUWgyEGJiBBJsCyFZGErtMkIzrI+7FDVDdO5JQYwuiDTAe9hU3eZ+wqhZXp9QmJdhbNpdwqtmHhPlmtiBK2SvmVZxJt5RS4S1cpaz3Q842g0oVf2mNVzxAAmR2OausY2jkVV0+kWqLapSy49sc2FEPjWdUsISXQQTGoIBJp+r4t1DinSxMH7iNYGrZPGQylJp+wxnU4pypJup6DXLciLgm7ZQWcZ586dYtDtMZ1V7G2fRkjBhQsX2NzZwjc164Mei8ZxNJ6iguXUzhZZmTE6mpKpRE0yOkNKQZnnLOoG18zoFDl52WndniIHB7u4GLBNxPmG+WzG17/6jVXDW1uHMRmudpgsa8X6MlHJWkqWQqXAyHA35zs1knVT0xVDYmtJFRH3Oh4tw/qWjloxpa8fNx5LW+g2o0QIcpNsf0MIydlLaTqq5NLp09yKEoymaTwgOTPoYZ1jXFWsb/bu/2N7AuClug+Vac685xIxXsCf38VduUmjvorovx9kwM2fw9vDY02HzRB1TuxN4PuayzcHr0Xwf9L4xPY6IoMmE1vIbBvRv4h/bpdbX1pwyhcMd0o2f/5Jpqcd5TOHADi6QETZOY/c6PHNi1/FqWbVTNztLiWiQPuCzHbpLjZ56MaPkdnB8eUEhaXHKL6XglP0xTNo3noK3mGnYnTSRJzgBCd4h+D+JxpSJgoLrRuOSEF2AoGUgih8W3g7XBDkyqQwPCFw3iNlRHqxWkESShBsssoVS6FzSOFtSidqiDYm0VyCSO5SIq06u2DRWiJFTm4MZZMyLeZzT9NYTK6RMiBjRtnN8VXNvE5ZGbGlcwnRZhsQ8QFyLZEyTRjQioVtCI1FqSTcFUh6HYk0khhT4J7ziSrjXEMIHmJACxA46rqmqhq8ytjqCfIs52A+p7e1QX1wh0wrrPN4Z8mKDpvDLtO6Yt44pFQUebI9lSLpKnpFgSY1HsYotNHJrtcHIDCvxjjn6XbO0l9bYzwes7m2TnPxAQbDAUiB8pHpYgZSUVULgnd424BoE8hjxCOZTidYH7h1Z0xetKGF3mOtp9cf8FOf+TRlWVLPp+xsbhEkWOsoy5LgGoqyRAjBbDbj8PCI3cmEGCMXzm+zvj5kfdDFugYrLfu3rxGCI5OSLMuYjHZp9j3ReY6qI2KE7nATKTUHQPAWhedm45C6QIhI4xxZqDkaH9E0gRBBG0WeS8pCA5Eyz3DOo8zxKb8UpC8rBdEeg7vvh0QP9NGlBi1Gjs+cJMiISyF7a6S1bB58DK1FcVxpbJYWwDH9wdVqo1QSnRlsvsZzezN87BCaiOxvoWNgTkRFQd5L1L8TvDEQQqJPnUKfOkUeP7i6vT78l0y/8bdhf5gms6f2EcYQKwdZ3QY33rXyvXpLXq1BaNn8MSJ1jxhqYrDEKAgO6vm9lLjRbcX17xlck563dcGRd45tVrMysn2p1QIJkM6kc9nY1Omu/uxxs/ujOJmJRFQwnL/zYXYO3kNpN0AX2E5OPpnzMx9eutQJ6lsdhA/M4mB1LWo3wtrBRT7QPMje2rPsDZ9nXhyLyrXPePTqT7M+uYh2BSokp6vvP4TpCEsahndtP640jG8erSppIkEgIy8fbHiCE7yFWNvUmPxVjD2kYLCmka/ABBYSDu5Ydm+caBLf7bjvRsNa1+ofJFIqovcImVa2kywitlx5gxYaHyLSqBXNJD01NSuItCIshENIjRegkr8oCHDO3ZXv0I62o0ORil6xul5KkIE8lxjdwVpPbRS1qwku6T0an1K8tQIXkjOokvKYmx9DO0nxQEaMEq3Sa/Q+NVXOB/JMsT7sECX4EKnrCuc91kbqpoHgCAgyIRjPI1XjkhWpEIRcs1g0hCA4Gh/S2Kq1wwWjcqTKcCJlckQC1tcsFjVlWSBjwDcNQkWmk1F6AVIghEK1FKYQFEYr+oOS9zz+KNduXMVHz87WkMcfeYD90SH93oC1wYBf/mf/gqqap5BBBEpofPC4CAiNA3SRQ2NZWxuS5YbR4RgfLWWny+bmKYyW2GZOZgI3rj2HDTCuIp1cIKJjNG8wSpMZgSIQrE3HGMl4/waz9Q6Vg6Szd0QXIUpqD2WR41CYLDVW3jYc3LiBjRBdwBhBr9CMp/NVArmNMOzleOdonOfF6yOaxqGxPKAMSVAeU3Oq80Txa5uI1HyyOk+PSRcJUrbOVy62uSOt0W0INI3jzp07GKk4feY0UcmVJqNxr/zlKNvKMMbjonFJcVjUnhBmRG+JiXxBiBHbWrglAtePsL3t24i7px752qcISJoXv0MMAdV8EP3IRerf/grxPXcQl3KkWiM+H7H6a8T+GIJisqd54duC2aFESNi55Cj6gVMPOZRqC84o6J39RfZf+A2++g+f5sYzhmYhmOype95ZW4lEk1tNfO9935WGzfMOZaBXGn7Xex+jm+XEcpE4rAJi/4iwsZf2TzvEogMIYnfyfQ3IuxORiPY5j179XZzefxIR5VJIhl60lrStaK0ZltTrHVxpsP3iZben2eDC+FFE52tc6fwTok9UKmVOM3Tvo6hXIqyX2xmiiDi1oJDfw3OE9goRwdNhFN9HxgF9nme5KnFv0/FKn+tXovDddWsUTHkQQWRt8Txnx32urI8IJyF9J3gbcO6hnL/41y/R7atXfIwQYDL5ql8/k0PH3/yr1ym7kg//2IAv/uoRX/n85E3Y4xO8mbh/MbhLtBSlQypQlSYQcc4TgiKGNG2QSmEytaKWCEHi6+v0eyRlK4Q2JA4hWt/RdEImy9NICK51CEoFmIoihc2104jUIyRdh5CJJiWVwBhD4Q0L0bA2KBnPJswrh1IGgW9dl1L4YGYkdUNyElIGZQyJVmtY3xwyHu1T2xppBEWe4SME67E+4G2gsQHrIlrkBCHIM4MUgrquiB6kUGxvbaOlpsgkZVHim4rMdOn2Cvb3D8i7JU888SRXr72I0ZrtzR2aJiDiDC0zdJYRheT0QLN/eECnHCKRBBfIdEZZdBkMBvQHXTqdHrZyXDm4SmNrijxj7/CQr3ztGxTG8IlPfJKFbWi8RfpU3EolUqq1DXR7Hbq9HiLTHNza5fTZs0iVodQdtjY2UQLOXLjAwnuEhNpbGgUz23Bj0VDNAtFJBAVSK5QxKCLE5OQkfMDYBY0PdLY2yfMyreoL0Ei0dQRtUEowmc/IzAYxD6hCIENMwXwxMongewOWdgMuRA4CoD1owdaFdY72blMd7qNkWnkUrQGujOkCvyqxRPubFEwXCyaTGRBZGw4ginYKEVvheCAx/yKjuWNYSk6fPc0yVE+0Go8YAkqp1TQj3KX9gLsmJXffT5qmSS0ROsMjwVtEq31BJvthgUq5Gid4UyGkofzsJ9HbO1Rf+Crh2ojm5rOo9Qt0P/xvIdc3AEFYG1N98bewu89jTp2DjQvMpt9gtH9IjJFnvnwbu5iyc2nG9kVPd+M8lz76abYHH2cyXePL/+gm3nl0phme3sQ1luneEcGHVQjm8YX43iuys3D7BUMqWGFx6yaf+cBZzm4N04IOwGgDefUSZA0xa9pGIxKHI/xj3yV2p2/dQX2DkbRVGSrm1GZKFCE1GvCyxUtUkt0PX6Be67zqdqX1ePV+ius1i71/ggCyjU8wfvxByi+88Kpl+6zY45uP/D2CmiMjnD3q8+DhOr3JBRqGWPpkHGEYI1nmBy2/DwSOPo4uAUPG4UpcvhxGBTRzzhGiQYhIwR0UFY4us3gRRUWXy3zyynnOjvt8+9QdxkXK+jjBCd5MKA2f+3dOsXXasLZlWD+VDHp+mDZ3uKn5g3/6FLs3Gj7x00N2zmU8/bU50/GJUPHdhPtuNJSWSCVSkrZUK8F3SlpOjlTJ0ce1lB+zEtdGZLt4llaNU25CS5eKsaVexVYsLpAiicWTQHa5xhwhSKIUyeJTpNX/Je1KAEEuV64ztEqTEh9ThsLasIuUkrqySGVQQqN1xmLRkJkMUKBUErUjWO+XlNkpZvM5ZVHQK2E2nSO1TvuVGfLCkOc5hIamaVIjFCIuKEwvIwrFw488Sr/fRyLpdntIEZlXlrquyPOMxlqUVnR7A6rxGCngsfe8j0sPVDSu4fqt28ydRaDpmiwJtMuSta1TvO/DBU9+8CNYl1yi1oY9zu5s8sJzT3Pr6lWmLrC3d0BdVwg8M+twPtGthBBJi+CT7kWqSKfXZevMWaq6xq5ZtEwJ4Xfu3KE/HLLW6zGfz3CNJUR/rG8QoKQGH1LmCSLlXtTpy8AnjhAZgljNiN0NyrJLXqTMCaVMyv4gpZb7mBypgnfcunyNweaA7mCdIGWiqAFGqkRHUholFd7a5Mjkoa4dudIEIzA6psmYhOjj6oKdxN/Jclm01/Gy7DC3ChEtWgr2Dw7o9/ut+1R6DS44jEj0OefaKUlbDKZJWzqPZatXSjguJpYWrMufS6S8jdYPN0aUUslooHXIOm6PTtYn3yoIrcje+yj6wfPUX/omzde/Q/6hJ1EbW6vHqPUhnZ/9ObAOMkNXCP71j39o5UBWzxaMb4+Yj6YIKdh5+BydYRcEPPKZD/GH/9qfYvf5m5x94hIXP/Qw3nr2Lt/mzrPX+cb//AVe/MrTr5grcncmCDFy9faY//FXZnzsiVM8dG6N67tpkcU6z8eeOM2gOzheED/YgkUH3sWNhhCCwelf5MLXIJ/MjpuMV0B+MCMbV1Tr3XtIZGK5+BA9EAlZhgCESIsUKhh2Dh9jcnGdna9cRrhXbvS9aqiySWuEEnl+8xAVMh4Zry3/GqP4PhQVffEMedyDlmpVxy3mnCei2kcGNBO64jJ53EcQmMRH28ek6+6MS0gsSfFoUCRthg6CBw7W2Z52+frZW9T65QMHT/DOhzaGzqCLyXNcYznaO/jBT3oZ5KbEaPMK96br0bye3jNpfy2QUvCBT/W49FjJTAz4Tv5hhv4O5+1TyNc5hRdCcOHhnI2dtN8PPVnyoc/2+Ze/PHpd2zvB24P7F4O3mgYhFDEkO9AYIiZT2CYlgItIW2iDkiEJqMXxl/+yuQihDXezApOJFMQm05REIAg+piwGqVKDEQUuunaVKq0ux5Y2o9tANSEEMoQkOCdCSLz44bBPkZUQIpkp6ZRp0jCf13QKhTIaRcR5gdQKKQwXzp7nwoULSdBMZOFqnnv+RfJ8jm0WRGFQWUG/yDmaHBJ84PTWGkWn4MqNmwibJjyNzhlunKXIFLPasj+1NM5S1zV1PUcL6HYKpIUiL5Br63jnUdrQ2egRCXz5C19gf/+ADz26xd5UUQeNjTBvHC4EJtMZtl7QG6wxPhrx7S/+Zkqo7nToZordfUFWlCzTE2WrqfFRkpcl/X6HjrfJStdkZFozrz2u8bggqG2Nd57+YMBwbQOjBKILi2qGravUSEqJMR7plxOA5XueSmNtFwhlkM6SFSXDrQ2KToeiLCjzkojAOcfezdt4V7O2s01vbcB0PGa+mLEZu/S6GUGW6T3XGp3nKKVYLBZMj8agJYXOcC7SiQG3yPETjdQtb70N6Vum2UPb4N7FHpEk96ncdDEaNrc2cS4lkcdgkbS0OyHZGqZzysdWqdTqNASKKFzrEBXSBEIkqtTLFYzLqUaMd9EoZPs5iSLphmJMFL/0UXjVQLsTvIFYJpN3SorPfJTsfY8h1wYvfZiUkGd3PU2sTALKQZdy0H3ZzetM896f/ejq92UD2t3oc/GDD3N4fY8Xv/L0fe9rYgx5fvvbN/nid27hw7EZwWhS86/9+MNo9erF+LsBsZ0qCgQ7356xefs80g7utax96RMAGD67y/TsGsEopPWsPX0LtXeZGxtfYJrfQplNemd/EVddph5/CQFsHz5GubZGcx/7JqJERoUXFhEl20cXOXfzJ6jYXj6CiG5F4++j4DaekoZ1aMuxYyKVomENG4fk7JKJI+ac5W4yVETj77qMe0osAzJxiAB6TcZnLl94Tcf3BO8MCCn53J//JX78c7/AcHsDIQTj/UP+3t/4W2yfO8PejVu88M3vsXvtBt57vHV453lk+xSPnz3Pv/red9ifJppRrxzyyLn3tY6fL48YYX98m8u3n35dzYbOxGqSOptLFkWHQnZ+6IUxbSTDjWX9CZunXqlZOsE7FffdaMgo8DZRZrRJmQquzVtYeXOIlMJMjEQXCC2jfOnGY5QGmRqNtBqdmglawbNrGmJbyMkAonWoSivSMlFGliJun1x9QogoAYqloxRIpZIYOCvoKole09RVnRKwXZW2KwJITS8zCKmRUpFnJZvrO1y8cInhcA1jdMqriJH1jS32Dw64fecO3gdsChZha3MHFWu2t7fpD9cpdUY9n3J974jgBAcHe9y8fRvvHVoptNZMZ1MO9nexTUNZFFy8cJ4zZ85hvcdaSyEV0lmC9zz06COcPTPBdDqcHl7k8PCI8eiA4cZ6+nJxlsVsjlYZmztD1jcG7B+OUE3N5qlTvPeJxzk62mc2XdArck6dOkVV1+ycOsV4NMLWFarooo3CupobV55nNF3QLCqIDbrscPrUGU5tbTIdT5HdEqU0RVGQZRnRe+q6InMjSiEIskNQ6XgiQTSe4GvyUEGIbK53EcGxmE2pJiNsvw9RYm3D0d4uLlicm7Fz+gEKnXH61A6bp86AEJQyEnWJkoJ6UTGZHxGCp8jKFGYH7N+5zXhRQd2QMtlM2o92ChHbJoO4nCjE9pSSHByOOTg4ojCCbHuL5FjZ2s4ScN61trcREVszBLGUhbcCDxGRop34oVZNwVII3p7QiBhQSDzhrgbcorKIMIoYJUokIXogQlDJqEC81K7yBG8+hFaojbU3dpvf1zDe87uABz76GN/4x1/g6NbBfTeXaVJ29zQNyn4HceY0z+oBH/rU4/hbd7AvXn1DXsNbjdSMafL1T6Pys4x6mpH+R2S15sLtj1LWa/cIoaMAXxiaYcliu8/Rw1vE1hFRje5wcOc/56j7Io2dE23ECUEz+QZER/ALBALjOgiRE7TC5xpp61fUafQWO7z3+d/PuHOLbrXB1tEjCG/4/h5IAAHDnPP33Pb9W02XNEHNNnXcvqfJ+H6kxyrC3Zd1cTIDfTdCSsnP/onP8Uf/8p9hcNf3zqlL5/nLf/Ovr86/elEx3jvEe081m7OYznjoqT0+aAv+zN/6L/jn3/gqMUS2186ilXnV7xEhYGt4ilk1Zu/o1mtuNn7uD21y5lKW6PG9DgjJhrvJqxtkvDbEAJefOXFUe7fh/sXgwWKkTqnGMRVqSiXBY3AOQURJQUDg/dLhKSCkIoSAkKJNC08BbEvaSiSFlimhQFisC0iROPOS1DQgQUaJNooQA0omK9DgA0YrjM4wKocYyLICKWG2mKOUpKoW1MHROIsQkiBARYWWisxkFIUEqclVhzOnLrCxsUGW51hniaRGKITEM9zaWKff67K/f8je/iFoT57laFmyaMAezdjcOUeuJe/5QI+nnnmBmXN0u32cq7HW0jQWJTM6nT6zeIR1jv2DEWfOXmJjY8Do4ADnHZ1OgQ8eYzL0YI0QPTbGRF2zE+7cuc1guI4QApNpvK+5cf0aTV0TvKexDYv5nEcfegj9yc/ym1/6KhsbazzwwEW++9SzVHVNrg25EBht2BsdIZWg0BkPnD1NWeQI4M7BAWuntxiPR9jaoZTEZMehjC7CweE+k+mIKDVRBzAlZF2k0ERqopTMMQzXOvQ3enS7OXVjiUFxdDRGG41SGt0tiFWgcQGVZQx7Q25ev4Epe5w6dx7rPJPplMV8gQuezmAjWdT6dH5VTc3mmTOY2YLxbViIg6RBQRKju9dRqrVrji39TwnB5uYaGxtrBOeS01Tw+AiND2idJbqeFPc0DVLQ6pGWX8oSon/FInJZCCbKWWuPGwS2cVgbEKVCSr36rCytcJcS8NTUnKzo/MhDCB76xHtYP7vF0a3XTpWIMaIzzWM/9gF+8t/5vZx+7DxCSqIfU31hBtc8aJtKgLtMN95piHcVKYnOJCg3f5be+X8bITOa3rc4Wvz3RL9gb+0Zdg6eYGN8ERUMeTPAyE2u/NyTLLZ7aQq/CnmE/YXnV3/5iCd/PDI8l26MRHDT1R8UURK15ujBrZfs20sgQAbJ1uhhtkYPr14BrzBoud8jLl7y/xP8KKPolnzu3/sl/vBf+ncpu/fqiZbT0tVjOyXFxfL4ATGSPVFxddHwp372P+FjX/k6//A/+juYmN/XZ1wIybmtB9lZO8v1vRcZTffue7+7A4U2aSq3p88RUMzlgA1/47638YMQI1TzE43iuw2vQQzuwfmUgyGOmwaNASGJ4piOIqUkOpdSmwkoKdEyUZ2iaPnxUbRWoYkWAqlgVsmNNP0vhVcgYkQLQ56XGK3bQD+Dtw4tJVILnG+Yzhcs6nk76bAQ0ixFColWup2ipFTnbidHSkFmSozKOXf6PN3BGtqY1vI04r0/5tOH9O9MKc7u7LC5scbe/iG7u/tEAmUZMCEQ6oaxiMxvH6BNwU986sP4wEocnILrgNDgnE0CYylZX9/g87/+a1SLQ7YGQ7IwJyhJkWsmk4rR7m32JxN8CGyf22HQL1EKjiaHeO9YLCImemYHhzhfo3td6tmY3du3GZSaj3/sg+ztHnDr5i4xBIzWrA8GKRNFC/JuAd4jVBLgN4uaSbVgazCgshapDaGyhBiQSrbUHhgfHSClxkeDrRpU2SRL4EZjsiRyDXpI1h0S6xlSCm7fuklWlpR5DwIc7O8ilgGQgPeCxaxmNplxuH+bZ33FeDZHa02IgXq2ABFZtLkmWmvqukZJxWJhqW2DtwvKTgej21N8tbKXNBAR2qDC1HQsrW6FAJQiSkGwfmUlHKJrnciSRuTuSUUk4nyboCHiagEnRI8Uafq3ChETS1OEZHqwvN05i2+tcmNIKeaxdRcDSQwNIstwRII9+aL9UYcAXOOYHb52h5Xl98xn/vjP8bv+7L+OzvTqtmArqt4v4z90izicU278NPX4KwQ7uktz/s4oZyNQHg44883H2XvkMpMze8Tbl5hUl7jZewoiuNl36epTdNWchjtcPfUlru18BYDB7AwXyr/IYqePnAd6T80Yf7iXVgeA3tkz/Bt/9H/Pma98C18dMOncJghP5jrIqLi2/VU2jx7iwVuf4sbDc6qNLqp2r358XnLf23Esl9PVt+FPn+CHwqlL5/nT//H/iU/+3p9OOr3XCiFohiXNsKTHkA8Pe/z23/xVxrcP73sTRqdAW+vq1/an7/p3TDQC9vU5VLToesLA3qF4ebO3+0YIkaY+uf692/AaksF1KgJtSMnQQuCDoK5tOqnadAHvAsEkZ6nCJAqJlG0BJxLFKlm7pqC4dJ9EilS8JrqSTgFzSiOUbtPGQ6LIEIjBU7kpi/mMKJahdpHoBT4GlBBkmQahEqdfCoQxeJ+CBhUSpQ1CZSgKzp05Q6/TbRuocNxgeE9RlkilmFd25R4kpUQJwblTO2xvbnD5ylX2D49QRjLodsmM4dLFc3Q7PU6v73D+/FkQlqLoErwlBI/zHucd1lmqxZzGwiOPPcl3vvkb1NURVQUhWOqokyWvEXQKyXhWM60senpEJ1sw6BYMeh2i9RyMR9y5Oafb79EZDJhMZpjDI9Z7JVtbmxAlV6/fYDKZ0rjAaDQCH45X4wUonWEyRWEMa/0eENDaMBxuIj2cPXeOzBiQgrpx1Is5jZJ4dwe0QcgCZQrs6CZbOuDJORrsEEXEzyfcWNxif/8ODzz4KP3uGnUzYzqbYxdThID+cAMjDQd7t4gyojLN0XhEuHmDvFi6kMm2TE/aCyGXQXkC0ARnsfWc6CxaqaSfaCcCLE0J2vM61RwpGPBYqE2r59CJ4uADMXry3GCtbG2ej7dzvEepzhBSJiOEpXZIHgvA23HG8Upt6zqltUYKiaedYEQwWYZApSBMb5NwXoCt3/oQsBO8tYgh8lt/559zcPXO63r+xvltPvXHfwadpa/4ZWOs8lP0H/1TjF/8TzHdD9O/8GfJxl9icecfAh63uJwoQ29jsxGXa0xe8aH/9vdx6tuP0HQXfO933+Rf/r6MAzFhsvvl1eMvbf4hHmKL61f+NtP8Fjo2dP2CncPHyLxF+EDUglDeq0+RSrKhDWtiHTFav2sKAVFETh08gQgKFSRrT99m98MX30gWyJuCiKCK2+Ri9+X7jLv2/8SJ6p2FJz/9Uf7C3/iPuPD4Q2/Y5+/1bkdJTZ6VzKr7XOgQsLZl7r0BKMKUTpxQobA2UhQ/3OtyNjI7cZx61+H+Gw0t0Eat3DRCW+BpUq6EaDMjhBAo3XbiIqKNhJYeJYVsi39F8BEjNWWeo1SaIiipEEK1zUhITlTRs2gqmqbBUQM+MWBiROpjWokAMBHtUycdoyAzss2mSMF3ymikiCiVo2RGjIYzOztM5paj6QGh5b8vC07v7goTFMer0cnqNBWLSinW19fY3Npi72Cfg939RC1D0X2wx9e/8U0mk8TpfeqZb7O7dwfnLM476qqiaSz9osdHPvhpbu7fppMPKTKJdxYfLXXjiK7GlIqSnBRQN2MydoyFYG24iSk3sNrx1Jcvc+PGLd775GOc2t5B6SMuXrqAkpLJfI5SirIwyI21u76AUlFb1w1ZptBSprTwGPHBA5pB0WXY76FFRGhNZT1RCGbjCbP5nIVtiCpDmxxT9tKCmnMcjEfMG0sRJU0ISFcxmh4Qnedw/zadwTqD/hb7t24wOjoiyEgUyRzg/LlLDPtDpDuFl4roPL0MGi9oXMpPEUiUbJ3NhEIpgzY5VW3R2mDtAqV10kBI0XK3Y9v8LqcMMaWzSHWsfzhe2sVHhxRparLqx9rx9dJxKzH9ZEuxWzYTEVpL3ZXDGqwcqgL3TjkgNShCSUSWpdRwIYkxedBkZZESxKWgNCfUqR91xBh5/ktPpe+g11Es5L0yuVu9DHT5EOuP/TWENAhpyIefIut/kPrw88xu/V3wix92918zVo04Ch1zzMbH2dz/NFvPps9tPi158CvvYfJvP8J1dZMXRt8FIFM5F4ePIeIac/EZXnC32MbzwZuS7vCzjB/YJkpB1JLJk+l43D25qTa7uE6Gmd2ruxAIqjl0snQt6944Ihs/jfDv/NVUT0FlHIXV3B3cOM0aFsYxLmoq7TjszPldb+N+niBBSMln/8Dv5n/zf/8rrG1vvqFNfjOvaOavT9PwsR//SToPdnnu69/h+jMvMN4f4ezL50MJwGRpvwOKRqbRRRHnDP0d1hTQ/+FeV6rJTprjdyPuu9FoakeTyzRxUCnobckhj40niogWGiUFSgtSSFp6nJEaKQ3GGJRMdqSNtYioKEqzShv3scG7Cu8dMToinsZ56rrG2UAUDVLolbORlAKtaB+f+LdKJt2IVAIXHCpKtFLJbUkIpMogZJRln7NnzyKkpGwbA4RA0Qa0cTx9XiY5r2x0w/JiI1quf6Qsc57ceZTZ+Qs8//xlZvM5N2/d5oEHLjGaXGdjc4utnT6/9vl/ztHRBJ1JbNOw3l/jwU9/igUH3N69TqENeZ5hjCEHeoVAqGRFK4KjdknnEbDYxnFweMT1a3t4Kbly/QaLRc10VjGZLxiNpzzzL36VT33iEwzW+ty8cYejwzFRpKC/RbVgOp2CUCyahnZgwNr6GnlmaKqKouhweHiIUc8xmS4QZclafw1jFJdfeI7LV5+j0+vgmhpvNaGekOcZKtRE7wnArBpTCo+QgU7Ro+gN0FmGCYEYLUVeMGscg+GQsuxS1xUb22c4dfoc0/UpvX4XYzJ80+BiIATwIQVIiuCpa4v1yWp3Op7QzCYopQhtDktqdNupAgIpluLw1u1J0hoECNIvicYlLIggyTJFlplVGN+SShe8RxpNbMMdhUiNR4TWmCCdQTEGlNLtPgTCiiR2NyIQMDqjLLsoabCt+DyNXQRCgo8e/0pRqif4kYEQ0N8cvknbFgjdu+d3746Y3fxv8fbwLRMPLycXaR8kpvd+uqf/MNvPC6LbxJd9pHuapd3ajbVDfvPoV5FaM8g32NQbPOxP0+gezijq7SHD53c5ffGDNOcf42i7n+zQeWUW0ezcGodPnGbnS5dfcp9Rqh1cCqQP5EeLdwyt7NVg6XGoN8j9Jjc2XmR/+AKRyEFnQaNd+/10UrC9E6CM5vf80h/ll/5vf5my13nDJ4l5r6Tod6gmr23xQAjB2sYWv/hX/0LKrjocceuFqzzz1W/xzJe/yXPf+C53rt5gMZ6u6qHgI00VkIUmkAKXdXxt9KtXw2zi+S//w+sc7J4khb/bcN+NRrdrKMuSGFy7mu+QbbFvkcRokwBRFZSmi9EGITRy6US1zLyIgRA8jasS/aUSuNC0FqLHhT2k73SjFHm3k1yCQmwbGlKAGQLnBRGVpitKpgIzunbykNylJBapMjqdIQeHE9b6JQ89dAkfkx1ptqTBiKQvkW1ht8Rx5kFcNR3E41W4EALBecajCUVR8uSTj/Ern/815k3DzRvXGR3NuPTARYoy45HHHuLw4BAjQIuklRgfHXD5ygt859tP8f4nHqdTbhOkREmZKDku6UykLumZDqIj8D7pJQrTwQX4V7/9RaqmwYXAs5evcXA0ZjAcsLe3x1e/+jV+9md/EkFkupiR5SUCgXWO0XhCDB4tFeMYyYuCGAV1ZREIfDPDiMh0vM/+wREugl3rY73jcHRAp6uIYYGSAaJF+ppm3obbCUdDxNaHCD9HSUXZ6SBlQVXPuX79BVywHB2N0NrTGw4puwM6ww2EUsxmU6q6wkiPNx1CsEilKPICZEY7+yKQBOF1U9FUDf1Bn/nYg4godWwtmzRC7dSqtbZNxU7bWIZAGxrfTkFSevpg0KUoCoJQCHGs29Fap4aF9u+EdjsiopD3CllXE7H0WXiJZqM1wwrOMZ8cJecgIjIqiI52pNZSxH50EYn42/uonTd2Ze/dBiEl7/v5j/Otf/alZHbwJh8LlZ0mG36Mxd4/fRO5/fFuPwak6tA988eR2SbRV+Rrn0SoDpP3OHRlWfvKaJWjBPDNjxyy2xwiGpAODipNf3qK9Y98lmg6nD39JHbnETbKbeZimUXxyljRJ+s2YyKmiWrUEuECuVLpUcuATSHS/ryDz8u0vpHB/OPUCPqj01xdO2DUv5YCDVei8nfua/idgrxT8sf//T/H5/78L2Hussd+IyHlsv567RBtGaS1Zn1ni/WdLd7ziQ8B0CwqDu/sc+W7z/DMV7/N01/+Or/1K9f51X9wiz/9V84gixROpaJ9Q860GCNf/rUx3/itCa/DefcEbzPuu9GIMaKkQuk0lZAiOQUpqTCypl4c0bDAqJJO3kv0qhDwwbUr255IwDuLkrBYVKAkXTQoiZAK522aOkC7LJxO8hBCGn8Hj26zNoRMRZpM1WAq3ARoCUIonG+dq4Qgz3JU3uNobDlz6gKbGxuMJwu0Nmh9TG2R7QcyBFb//v5jkEIF292LKZeCGFsr1chsMSd6z53bt8gOxoyrOTdu7LI/OmBne5tnnn02jR8lBGfBR06fOcu8mrO3v8d4fJ7z50+hlEItxc5KtU2dRCiBkQrT2rYW3S7TyQKjDEIKZPRkWiOkZvf2LXzwzOcLvvCFr/HgAw9w6cI5pvMFs6rBZJpL588ym8/SdArBoN9L2RAi/d3JeIwjIkIqAgOOvVHVCukVMniEaPMpQmoiM1OgoqTwM8qcVm8TKbsGKSJCCWxdMVscoY2myDVl3qOa7+KbESJqJru3IARsY5FaY4WkX+QUZRdjDF4ojJBk0oLMqFVBnhfYyuKjpF5UmDxDS0Wyk038ZdFSnqQ8nmhAEusv812ESL+L1v1MqBykQYm7AgONQWvdBjXG1GOoZRPaulMh2zolTTaEWOZtHDc4xynhgSBIdCnnCc4SlcbFgAg1WuXUtiZEh85fnhLzowJ//RZqe+MdXdC9FXjgI4+ycX6Hvcu33vS/JYQgX/ss1f4/580QIqTmXKHL85juexAyw3TfQ772aY5zdxJ8rgkmWcm2lTNRCh7lUcSZRM28dXSZW/EFrl9Yp98tkMCw2HgdOwbl7jS5UQkYPbrD/vvP0bk9oXflADNv0LOaxU6faqPLxnduoiv7zj43lxTQCHnT46EbP8bXH/27OHU/SSAneCvQXx/yv/5r/z4/829+7vWJvt8CdNZ6L/lsLhc88k7J6QfOc/qB83z8F36KGAKL6Yzdq89gmv/sTdkf7zhpMt6luO9GY9DZZNDpo2S7ysOSFuJaXnvSblhf0TiFjw7nbEtLoX1sJIpA7QLWW7TUNF6gosAHlxoGsbQLTfqHKBP1yceAFwYXIiJ6jNRtcQ+KFPgmhUSY1IgoCUIJMp3hneRo74gHzj/EqVPbzOZTjMlW04hl0Qm0+wgEv+Lb372aGNsVreVNYbUiHqF14mq8JQrJlRvXaaxjczBge2OIEEl3YsqMpkouSY1vmFYz8JEy1xxNxzjnEARiVO1FVpLnRat9SYncAfBNEtQLKTl79gx39kZMqTFG0e12WCymGFOwtrHJ9dtJXN3rdXjm2cv0h2vs376DyQzaGJRS2Kbh5u1dlISyKNjaHHLxzCa9TgetVDpdZMS5RFtyHmq3oG4sVV2zaGqqpmE4XKefD7hx/RodbekCIQZc46maCuslVVXhXINqaorSULuAkhnOLpIzmJ/TVDVNbcnzLsNOD2c1B5PbaC1pYqCQEkuiyz3x2GP0+jk35mO2zq6Rn1lnsZhxZ28fow2dPEdplTIwlqudd72vSimyLK0qeW+JUSKQSJX+E3edI+l88TiXzmspRepOW3F6lKkx9q1AXMmUHr6c1gkhUDL5TgklW8crSJOOgDEZMsuxIhKtQwRDFIJuf0AznxDCj3jK749AqNwbgbxX8sTv+hD/8m/942P3u/YcNEWWzucYaRZNClH9IaGyLaQeJvrUG1hIt2tAdE7/Ibo7nwP56labAohSMHrfOjc/V9H/9oy9n1mn+v1bnO4lOtMw32C7f44zvYv3hMK+Zgi4+WOPkB3NiUoyOz0kGEW13uXwsVMQI6p2hFwjXKB3bYSuEv1oNW15pzYd7WXJqRov7ckU4x2C7fNn+HP/j/+AT/z8T77uacP9QijB2pkNjm7t4+1rE1HPDqftAuSrN0JCCIRSdIcDyu5j3L7yWRqbrn37tyxiXjPY0JRd+bq+V1bfefnJ+ftuxf3b28aItxEnmiTUjr5tIDxNcEszH2Js8C5NGpQE6y04oC3atDQYFci6GtoPmQgxBfC5Vtgdk5gbEUF4tEw6jzw3KCOpmykhRpROF50QY8rbiBEbPLkxGKXp9frUlefW3h4XL15kfb3HYjFDKYVUqhVtR5bXaBXTajbLcaNoR8wtz3fJRZTtpGQpCD9eGU8CZS0Un/3kJ/iVX/s8L167SV3NOTg6SMfNW4IXhBgIPnHFRqMjgg+EaFkfbnBwMKLTKcmyrLVcTfkkumlY5o7ESDsx8gTnKLslLjREAXXVcHBwwLyq2dnYYHOtz+WrV/ju08+kaYDzTGfTNFlZLOlgEdu4Vdp7nmlmkzGTwoBPmg4f4OypDcput+U+KxaLGVopbAjsT+cE7/AhEqJhffs03VKTGYX1nsl0QQTm8ynBSYzK0jQkRDKlKYocGSI2RILSFEKjdEVZFAwHA4SU9M5u0e+vM19UTGdTgnXsbK7z2EMPoI3hzPYW1qawQ+t6fPuZ57FNQ68s8NZijKS7NqDb6aVJkUoWskqlhjaEAEqidYF3KeCvDotVoXdc7MnWgbkVeS9dq0IrL19eQIRIJgOidbYKyxyZ1g0IUgOKAGHwCIwSSJ1hIgidI6QkeI+zDVl3iDFvzpj9nQAB6Evn37nF21sIqSSf+l/+DM/+xrc5uL4LQGfY4/Gf+CAf/L2fpLvRJ/jAzaeucOXrzzI/nHL1my8wOxhz9vGLr7mIUfk5yq3fw+zW/3d1cV/iXpOEhO9/zOqx37cwI2RGPvwE5ebPINT9+VsKwA00L/5vzyN8JGrR0h4TOqZHx/RebRP3ByGo1jtU68d5Bau9bym0vpOtHOVuf/wSqnaYWU1xMKd37RDZOKS7t9ETL3ds3oZzOorI7Y3vpgWWk0bjbcelJx/lL/znf5X3fPyDbwk1NCtz/th/8mf523/+P+PK15/j5v5lJosRJsvob6yxtr1Bf2Od7rDPeO+AyWjMYm/GWra1Svl+LZC6T3Hmc4Rrvwze8U/+m32aa4YPfPoUv/AHDxAiNTsxRvZvBaxNgc3zScNor8bkGcF7Lj1qWNtKRi7TI89v/OMRv/VPj97ow/PGotV+mswglaLT75IVBb21Af3NNUyWsXvtJke7B8zGE2xV36X3/dHGfTcajhqLQsSAb1e0aW1lnSOt8seIdRYXkw0ngBK61di2K7rLELK2kCeCx6NN4pRIqVd6CIj4ECB4vFTpZ6sDEnKZR0Cyow0huQMpQ1Z0UTLn4GAOwAff/156vQ7WBbyPqAhCppTlZKub3mwnEqUG75O7iGwpVcj0JR0ivtV+iFY4TjsBSSdMbK8lkaLIefyxx1F5wXyW3KVCCKhMQRBoKWmsxdoKLQVaGUym6fe6zBZzup0S7xxRpiK4aRpcTGnotM5XMbYFbgSjDcPBkMzkPHDxHIu6QkSwjePWnT2apmE+X6QGsbUSTu+DoChSMVvXDd4GYtSMJxO8zWhsoCwnXDizw+baEGU0i6ZJx7ks6KxtIZWmKyXDTQ8hNVGRRHtb1A27+4do1zA8NeBoOkcFOLe5lXQ1WUHjHUoZOrliMZszXkwxeYfFokZJ6Hc6bK5tIrXCGEOWF62uIVGdgg/c3jtYifeNSoV+FILHH3kQCPiQ9A91kxqh2azCWUtoqXrdskQojVJQlCUiJIczgcBZi9Jtc+p8mnih0zkY7nWVQqYL+t1UPOTLrOQsf2+bVO8aFAFHgBDwjQUpWjctgcpMoi0KgX2NK1PvJsQYWHz3GxQffD+68+aIod9N6O8M+cX/4n9HM0vOMVknp7Pev+d82ry4w/t+98eIIVJN5iwmc7rrfaR+bY2GEIJy6+epx1/E17fI+h9E6j6+2cVOv00M1V3aIok0A7L+h5AqFemRiJs9jVu8uDIKUWaD/oV/l2z4UYRYhiTdXwEjIH0GfoCg+4fF/Ww3TVkks7Nrx7fFiF5YhPPkh3OkP7asLvemSNd+TkOkf/UQPW9WzzveyJtYbLZ0XqdOkpTfCXj0I+/nL/3Nv/6G2tf+IAghyHslH/5ffIYmq9kQWzzwvsd4zyc/zNmHLlL2uiitkUrincc7y81nr/LdX/8aT37mw69r4uJava5C8yf/r3+Rh8++D60j4xf+Gs30m+0pL+mc+bfI1z6CAEa7B9jiNpeeeIT5dMazX/w7vK/zPYqO4toLBX///zNh9hoF7W8KhMAYQ2fQY217k+0LZzj/6EOcefgiw60NhpvrDDbXyTslvbUB2hiyIkdqhRACWzdUsznj/UN2r93k9pXr3HrhKjefv8KdK9c5vLPP9PCIerHAux+d67yIr7Qs9X34+d/zE3R7OSBWjnnW2SRa9ZHJ1DOfzxj0cjY3eylrA2isQwqB84nmY1QS0ColkboN8SMV/b5dFUphgG0xr2TrEARCRKRIqckIksOVTE2AlJq86EDQHE3mTKdztjY3eeI9j9ApChZVTV3XSKWBtN20Cq1WIYO0zYsUaWKxpGZJKYkhImMa6cdWFC4jiS4luOsC3DoLxUSjSjaxAUJ6TJYbNje3qKoFSipu3b7F4WiU6F5KoVQKiEvHLBXSyRPr+ALtQ1ppT+uMiXrjvWc2G1PNa7KiRGlFYxucbTg8GLF3OGY+X1DbhhB8m0qd8ku0SOJoJVP2SFFoijxHq2T7qpXm3PYap09v4QEpDUJEemWB0hlVk95jIWE+mzE6PCLGSH8wIASPUingUWlN9KkxGPa7KY/E5Ny6s8fO1jqDfg+tNT6mlHmhUoaEc8fHIQSPlJrWXBbvEpUsLqdP7RklRRvOKJLdcggRYwwxhvT+SokSEuscLnisdbjGJ7qfURRKovIMheS551/E5AYlFVVVoYxm0O+121tOM5LF7bL5tdau3qPjczlN0XwI6RwmNajOOe5cv8Fzz71ItX6WPO/ig8daj/MWQUQpQ5bnKGOQQvP3/sv/+If97L8t+J8Xz77ifTF6pvYFZntP0dl8hH72+O9oQfjbgZRHNCWGCmk203dvcPjmFnb6HezieYTQmO57Md1HkWbjHh53cFOqw8/j61uY7qPo8mFUfuZH+n38QRdQESNmWlPsTxEhUuzP0JVNP2d10nzE48ceP/H1HbMVvTfCUe8633jk/4fV1ctONP7kZ775uv7G24l327mktObjv/CT/Ln/9D9g88zO27L/L5lQvon74ILj9uxqmvgHwZnhRUBQTe5w6+m/z9raAVIZumf+GLo4+7L7enj7Dt/9V/8T7/3kOYq1D7F7fcYzX/km3/yXX+B7X/w6t168ymI2f1NybYQUKK3Jipzh1gZbZ09z6oHzXHzPw5x96BI7F8+xeXaHzqBPlmerDK7XixhjWnB3jmq+YHI4Yv/mHfau3eLG85e58dxl9q7dZO/6Lcb7h1TzBa6xrzhRfqtxP/tx343GT/3Mp8gymVKLpUZrg5TLlXXDdOKYzyuGg4yd7T7a6Ha1Np3UzjmW4WWhXY3OjAEBgUDwflW0LdOXY0wTCqUUPnhm8znRC7LcUBSGXtmjyEqCMEynFaPJFFc3GKO5eP48F86dR0jRFn1ptTSFqLUp3XIp0E3TiSCOP4DLYD7ZNjJLCr5Sx88RQiRHmLtExXAs8k02gkAkeRDFVBwXWc7Zs+c5HI3Ii5yj8RFXrl0jhrBaCVdKo5QkuoAnpKamdR1KNq3tZ6xN1A4xtLkmop3EyNV1ygeP96lwtb7G2STrXorptZAr8bcQEqVlEv63zZjWCi1BC4m1TdJoRJBas7m5xXQy4WB/n8l8QVkU5FpTlAUoRa/MMEoSkLgAWkBWFKnpCBHrHE3jQQYypXHeY51DKb3SvyxpSUIsRaXJwjjJYmJyI2tXUHWr0VmNy1rhdfApiNEHD7GNmJQSFzxSKbTWGKVWdDyWIvcAztX4GHA2ULuGqmkI1lPVyYbZx7SPxii0VmRG461DqHSbFGk1Q2m9Og9du88xROrGsn/7Ds9dvobdOEeWd/DW0czHYAq89Xhb4b1HyERB/JX//m/dz8f2HYdXajRiDEya7zFzlxM9EMMge4JcbSNF9q4rLk5wN+5/ivGjiu+/yIoYESHpP4q9KbKdUuajOWbekO/PyI8WaSoSIy9xo/3+z8Nd1x+faXY/dIFsUnMr/Hfc6v92m3P10v06aTTePAgpOfPgBf7IX/oz/NQf+f1kxatrk34n4IdteGKM1IuKO1du8MxXvsE3f/2LfO9LX+fWi9eoZvPjbS4JBu21fXl72eumSINeh06/x8aZbXprQ9Z2Nlk/tU1/fcjW2dP01gf0N9bSdKIsV9lwb/X7tzxewXuaRc10NOZo74A7V29w58p1bjx/mTtXrrN3/Taj3X1mRxOaqib4t24a8oY2Gj/+kx8jy1oHqCROaL/sBEJkOJfEsWWh6XVzisKQaZA6hYv5kBK7RVoTBwSB5L0sVCq4BJApnSxLY1o5D/iWGpXE4VoXaJkTAlS1paotLgSED5SdknOnz7C9vYXWCu9TAjekpMulW5RArFxGIFmYKgReJFrUsrAlRpRIE5XQnmBiSa9aXjyjQIjkuCQ5ptAk96HjyUxcpSckR5CyLLlw4QLXrl1je2eH27t3uHr1avvGHQuHl5fo5Hd1DLXUiCzF6m0TsnLQCu1kRsuVQ1ZsbX+l0BiTXKpotR7Hp0HKlZDCpNcZWyemluqmlcRIRW5kK0RP2Sne1jQu0O0mu10pk/C6ritiCHgfSFklgA8rCtqymxdSHDd24njCJJcTNBGRBIKQQEr7Xu733TzH1KSp1etZprzTHgMfA6qloN3daKZzLInCIykkL6Y/ghSseJexPeZp0pbOIWctIULT1DRNQ1U3LGYNjWuwvi0UREzaoSxlpGiToQhorWisI3jP/nhGna0zqzzTOqKkQ5oSHyy+mtPYmqrx2KrhV/7uf3U/H9t3HF650XDsLX4TGyarRj3l8BQMsifJ1duzEniCE7yZeLmLb5ImBvS8IRsv0AtLuZ+aEWEDxeEMaX2ahLQb8blmdnaIzw2TC+vMzq0BgsMX/xFXvvS3yEpPXkY6w4BUcfW8P/nZb71Fr/SNwzv9eyArch7/2Af4+f/VH+GjP/cTDLfW3/H7/G5FjJF6vuDOlRs8943vsJjO2Di9Q1bkABSdkv7GGpDOm97aMGmFswyTGZTRq/fm3fgeLfXBzlqq6ZzxwYiDm3e4ffkaty+nRmTv2i0O7+xytHdINZu/4dOQ+9nWfWs0jFbkeRLvhlb8LYTA6AIfDJnJmM3nOGtZ1BHnK8pc0imTHiFpKkQKcmEZfCZROhXTRksC4EJES40iFY25yZHCIJTBNp7xdM58ccR8UeHqhlxnnDqzw7lzZ1nrD1KDEQOLpm5XvlXr+LPsKsQ9J5ZqNQ6paH3pRCMVwz5pOUh0KRHiXY9LKeEiQmxds1bbj2I1Co/I1TGLQNXUNFXN2XPnuH7tGo888giTyYTRaESMiUq23J+lLevdBbWPKatBCokXcUUJSo2Db5OnQTRto9XSo0JsmxbnUVK2rzscNyirIInYUn9ItLa2YLcOmggLkSYdSjm0VoAk+MjRaIYnUceKokiCdiNWHbZC4EPAhohtaU+0U5qQqviVvXFwfmUYkLQ86bUtG6YQjvdbtPQlQsAtP3zOrZoXH3w7+Yj4lsa0dBtLou/2S6uuibQJ3u05iCCNaYEYYppiSdLULgSUlEk7ojWdokh0E9KKFjGF7DW1xTUNk8WCajHnYHSEtZbtrQ021oYURc7ps2dYLBr2DkZ857kbTKOm6A0weYHqDVEIMhfx7kfRdUoxyJ9g3DyFD3No80pcXDBpvocphihxf0LiE5zg3YKXLW0ERC1pBgXNIJ3zo0d30l0xInxEOo+qLKtlKKNwnSwtjty17WLzp7j+wiEvfvkpqskh25c8j37CIMUdBts/it8jbw+EEGyc2eFTv+9n+Ok/9gd45EPvxeQnk9g3G0IIim6Hi088wsUnHnm7d+ctx7L+yfKcLM8ZbK5z/tEHgU8eT0Ocp6nTNGS8f7iahtx68Rq3X7zG7rWb7N24zejO3pu2n/fdaBRFh263k4w42gI0xjSaahpP9BWDfsFi5pnOJsx9wHdztFGEZbBZXFrJspoEQEzC8ChaLnqG1hlKGXyAuknJz3UzxTUOay2ESKcoGO7sMOz16Q36ZCbHhoB3Ke/DmPweipOAlbBpScNJrkESlRau0eLYRWr5uOVPubwk3GWPekxhiqsJBIAgIkkr+rHVFghA+IiPif4SRWQynfDImUe4euUy89mcS5cucng4Wk3F027EdvX+3q4xkgLdjulEy2Yq8XMRx1ORENIEIQU1kAyTnMfjU3EvSVQp0nEQ4ngSs1y1N1qvnGe8a0XSIdI4C3WTKFZKk2lDmSX6W20t1lqKokiaHSESTS5KEDalcd+lx4lStFSniPMuid7vmbak+7TWq/fTL7MvltONdpqRtpkaE+csy1yXdN4JnLfE0FLc2vMSjicgS03FkmInhWy1Q+mc996zmNeE6Ak+Jvev4IlBkGeSzY315A7VNm9FniGKjP6wSwzpPWmaBt8K5+eLCm+nNNZCaHj47HobwNiwsAvmC0n0nrqqifFHr0AQQpDJTTaLTxFiRe0PsGGU7sO8vTt3ghO8DXj5JkQQtcBriS9e+rm45zkRykGHX/hLfwxXWya7R6kw65Vc/trT2PpEJP7DwOQZ5x97iPd+5qO8/8c+wXs//VE2Tm+/6Za1JzjB/WBZjyqjKY2m7HXZPn+Ghz/4JHA8DfHWcevyNf6f/97/hWvPvkA1nbGYzt/QfbnvRgMESpAC19pCNgXWRXSesjWijBiV0e+v4XyyjW2aiHMeokvibZOKdCMjBNAmR+oUACiQNHXDwgasnVM3DU3dEBHkRc7WxiY7O1usDfoEERMtJYCQ///23jXWtiy76/uNOedae+9zzq1769VV7qfb7qfbL162edjEDiTBBPIABIoUEkhQgoQSPoAURUoURZEA5UMiBRCJIBhQbMA2GDDGYPO0MU8DJna3G7vdVdVNu7vedc9j773WnGPkwxhz7XPbgtxC1S/3+re6q+qec/Zee+19qsd/jv8jLyfUJz2/G6VP0aG3Xkmc7rtcy0VJJgnEKCkvZvNOJDISJvAHNx7SiwLjNLxvHDbFTcKyrDoAE1J274PPu8I0V0oeSKXw2uV93vq2NzMMI7VO8Ry6nLTDKV5XxJMPcs5+ao+QFJrYLXkYS1JWEjmddC3EQ0+vQ8FEaUvKlrgJvPj73r2eCd86TFNFo8bdNz3xEqmkJEjcv+1mE8V3CZ0r18cjDWW3O2McR4oaOrhHoc4VVLGcadq3M7ZIoW7LoRb/y6f5YgD3j4hhTeN9cSKYhKULQ9X9Gg94bYKw9MdPKVFvbQ6WDUpOfq/tQbnWOA4Mg1DKwFiGhQT1x3GS0/daXvCXc4nkMvWNlyTGknn0zpn/PlmNT6cwV/jUiy/zyU+9xgsvvvDwv7ZfQPDf2UKSC4pcAG+/9cXP2WWtWPGFib7ET4lxt+Hxt79p+dL7v/kXfo4u6gsbuzvnvON97+Lrv/Vb+AXf8st5+1e8m+3Zbt1crPiCw6Ji2Yy89d3v5H/5i3+M/dUN//gH/g7/++/876nT/IY910MTjRR+DO1DICEtqm3xL2jzr4mFXCq7xj/lQm0VM8EsIVIw88buqcL++kirB0juvK/NT8zPzy74kjc9xRNPPsb5+S6MucqxtviXaPLY0Vu6/uXm/St0d/3kustoTgN4aPjbrQH8ltyqR9qKnR5Pw1eSEJdM5ezSqmVdkjHPvvUTdfHkKkiIGKUk71QAWp0ZxzGK82z5P4nbm4XbbdZ9UI4X6b0cIT3qL6hfaZcl3b4Xy2k9bjTUpm4sF4kuh2i7jk2JGpR4nYup2m/bIi8yPInszLYMpdBEPbo2ZYZcODs/x8QlTcdD5Xg8shlHzrYb0sUZV/evltfQSV3/b62VaZoefA/7JiRkUv39thYEIxhXycVlXinRmj7wvf2edBlWb2ntKVf+WVHEWJra/V4r23Ek5wJJ2IzD0jOX+8YrOVn0+2XMc2Uo7vlp2rA2YVopZWCaZ3bbgZJCGiXRv+GRZkhTHj0/Y757h/zzvbAPVmKxYsWKzz0E7ty7yzu/+v18w7d+C7/oV38jT3/p21ZZ1IqfVxARxu0WVeNvfMdfeENJBrwOotFadclMZJpbyJRaa6ScIvLVZVE5J++ryJlxsyPLlmoNVWGeK9fXew77Aylfu4xH/eB/tz3jiUef4PHHH+WxR+8yjsWH8CYcpgMaZl0fuvOJUMiDhMLo5vMUFmwv2VvSoG4N3WFHcBmX9DI2dUN2nKovJ9kRYdTjA7ucymIkNVWQtKQDWRQAeoeIv1DBB+DWlPOLR4JkVMrZOTmXB4ywt7cnPW0pRQSvqnI8Hh+UhwngNmbMZInBvf09t7cjEkQCvBxM5PRagCWS2E3TkAc/qU8J1E4ehmXw90ws5taQnP1+BHlrzU/3y5A9Es6E7bhhnmeOhwOp9DbzAy3el5S8kbvWukimlk1G3KOeZtaJU98c+Hsh0WT/oOyse296wpmF7KoT1dZaxOhm96eEHK0MA9thWIzqpW/Qovn7dqFjv8amGpsmZcxCM8NapbVKbUoKYnE8NK6u9wwlMZ6dMWAkKnMzjpOTrFYrZ9uRsy956mF/bVesWLFixevA5mzHm97+Zt7zi76ar/7Gr+OrvvHrefKtT5NLYSUXK36+wsz4oe/+Pn7sb/29N/yxH76wLzYNReSBVKG+UUgilGF0ApASrcL19Z7X7t9nml7FICJNodXJG8OnxBOPP8abnniSp596nPOzC9fBixOFq331AV2MlAZKPnkGQEjRgWGcJEVxx3yodoVXDOonqc3tYd7C+9C0nozHGJhg8Y9q5luJvq24tTnwcjpbDMAaOYTJfAtk2tOmPM7wFEwr3L13j/vXV8zTxGb0lATVtmxXLF7L8lNq7vv4tNfRfQyuaMuLx0PiPt2WHJ0ITH8cN5tbkCBP+slLRC7d2yBwmCuiFtIyCRFQSJNSJnHaFtXZT937hqC2Bq1RW+U4zZSQRA2bkWIljOcnOdI4jgxDYZ59yO7yJglSS2x1Ojno0qcejeyekdN2YjpOJ3lVa1T14sKckhMYO70zKWeGsbAZx4g3JqR5Ep87WVrl/T1y0q30zxtxHX5duRPhXBAzFMUaiFamubG/vmHcbiOLzRBVmglVlWmemaaZaT7S1DgcjxwOq7Z6xYoVK94IlKHwxFue5v1f/wv46m/6Bt77S76Gp7/0rWzPzpaD1RUrfr7j1edf5Lv+tz/6GSkKfHiPRorys5TIMjCMGSTTWi9wEy6vLnn11Zc5HA7egZAS41DAWORQ2+2WJ594kqeefILt7oxchNZgPzVujq+iaviBcfbhNXoDRLznYtlkLDIilo3GaYQ/EQKX4DxIMgDfVDyg8U+flmIERNcE/TFCOoS5idk3IAQNYDEc55RAPaJ3sYrH4G7qCVYXF2c8/sRjPPPRnwEzLu7cYZrm5fS+I646/l4XyVd/LbclYyefSN9eKCLe44D0q5Sf87jQpVSeQNUMvBjdCWQnLK02kgjDOIYMqS4Dd9yIn7M5un1PnXR490iN+9bahIhvVMahcO+ebznmeWaO5LChJO8qITHV9sD9mebZ78zcUIyhlGUzNc3zgz4M82I/UmLMQs6DN9Vrw0TYjlvKOCx+n9sboE4gwKVykpxkEh4Qo2J2KuLr919NvQcj7lUndagyVeX+5RU/+8mPIwm+5OmneevTbyGV0T9iOnOcKy+8/DIvvfIiTeBw3DPX40P/2q5YsWLFihMkJe7ce4S3v//dfO03/zK+5ld+A+94/7s4v/fIsp1fseKLCarKX/jDf5LnPvyRz8jjPzTRuDi74PziDhhMx8r1/QM3NxOHw42fY6fCsU5I18ejnsIzFB679xhPPPYoTzz5BPfu3nH9ftMoa5voMaqppEUadVs2BCeyMM8VAXyeFFQl+g0M04ZFipVZjfQllwRFSOyyZVjOzkMCpWaotRjA08lEjRfhuak7I0kYQjLjJuNEimG7ExeTRlMhWe+vBiFF5KyfaL/pqadJknj55ZdJKfPoo3f55POfWojGKcHqtiysexJ+rgwKHpQxsRAgAxrSj9ulF96dBunbj2V2awuTjMbJ5C4poZKoqpR0InbNNEhhdrla7eInKMNAKeWBf4F3mZNvJDLaDJ1nDtPMmDI5CyUPlDwsRGa/3/vPJjdc5zDcX+/3HG72DDmDNqZ5cu9M3LOhlJB7uexsMw5s4ucXQoFvrbrvpW84Tv9Vcvbtjpkwx/bF5WARipAEUWEy93OYGdY8OrcGyVDUvS21YQK1GVevvYK2IyIlPmeybEReeOFTfPTjz3F1c+XeqJRotZJk/T/DFStWrHgYSEpc3HuEt77nnXzFN/xCvvKX/xK+7Kvfz6NPPekFr6scasUXMcyMn/mxD/F9f+xP3/L4vrF4aKLxiU++gHzyRabZI0KNeWmP9kF1wiyx2Wy4e+cRHr13j8cee5Q7d85dFiSCNuN6f8TsCEEIJBX8yymkKQ+mCS3leQFPkpLwPXTZjpMIywMYqLXlZB26cV3QJrjXOyJ2UTBlDjPxXGda8y4P1YimDY+Dmg+545C5e3GHzSbHiXULwzQ4EQCaG6iXrYFF94K4j+Pi4g5vefOb+eSnPsn9+5fce+Quu+2OT3ziZxfS5QZuPC3JFJO4zzQw+Tlxt/1nhEiCEqKUMLYLt7c6RnSHhKF72YxEx4lYX1AskbZhZAExP51vTlaGnEnmxEzyyafQ44zFlIItzebd5K3WUIM2z9GNYVgSqtR4qsxmzIzjyO7sjM1myzQduby+4bXLVxnKwLjZcL7bcufMs+ZzSmhrPtDXBuIN62OXQC2bsNgMcZKoicAwnBpEb8vRugTqQZma+HXjhYLNQsCndkq3AiQnsmRy8vCA3DKN6k33V/c5Tke+5Om38Kanv4Td5oySEtWMuSrPv/gKr11fYTrTWo3PkveQrFixYsWKnwtJiYu7d3jLu4NY/AonFo899SRlHFZisWLFLczHiW///X+Qy5df/Yw9x0MTjfv3r9gMmz63u9QoCZthyyN37vDYY4/y2KOP8ciFE4tcincpTBPXh8NClEoqpH6SHJr3ZTC75SVIsY2IpFRy6hKm28ZnH4QXP8OypwjvwW1/g/l2IqUukfLBOYs3W+swQDojJ2+GXmQ/nDYFy/Afz/Hgv7AiaQtBxQdQi4uXiID1BKTMO975TlISnnnmGZoqb//SL+Vmv+eFF154wKwuBikVet8IQJICtwZ64IEEKo0T+D4gd7/K6a6E5Kc9KME6SZ5SyMr6aw8Jm8Q9lURKZbkfkoQy+B2Z5vqArK03pk+10qQx5oJEJO80zbfkby73smpo8nLInN3zcjgeaXvfKIy58Pi9ex5vrMbNzQ2Xl5f+vQZnux3jppCtUHK0tcc2K3d5V6uIJSylYO9Kzp6C1j87auq2dmtLEpeqABo9I12mZmituMQuLbK+/hkW6ffWezfC5EMZR/b3X+Ollz/FzeGacb9BDEqGWhvPv/gSL770PC+99gK7sWAts9e9G9fFPiMayhUrVqz4QkQumbNH7vDmL3/HsrH48q95P489/aaVWKxY8a+BmfH3v/cH+Ud/9W9/Rp/noYmGkShlYLvbcufOHR679yiPP/YYF3fOGIYBicbvpnCYZjTKgNxQLORcYphNS/BqCz+GmSdE9TP6JCepi4V0pg+w/nVZvBnA8j0dixToliSotfir9SFNSNFqoJEq5K/zlmypS7jcoLFsLVonQjz4fc0s9hjemYHgUhlhkbs8+eSbeOqpp3n2uWe5urri7t27PP74Y3zoJ3+S/c0BbRZbGqLJ2wfUlIKwiMvLPBlJHxjsgYVIne4LD3RE9BP908bH4uv93qVbA/KplyT1f855SXbS8N3kKGQkZdo00+YpYl/9fnWiV9VA5+iViBvayaUbPBavjJOnKNFTJxqaDZGZ7bhhHDPbYaRpY3/cc//+Ja9MvikbS+Hizh1KGSAXtFWOU/WtTDYSg3tlok28E53TByoIWciV/P21kGR1KZtv1bJkzG5J1vDNhUZwgGHeJh6PVSI44SAC6slsdy/ugAhTVa6v9jz73M/wyvWrqHp7+zRX1GAoA2ozF7u1IXvFihVfnHj8S57iqXe8hbe978t519d+gLe/7118yZe9nXtPPr4SixUrXgdee+FlvuMP/OE3PM720/HQROO97/5y3vaWt3BxcSf6HjwitNbG8VhRnSEGNtfEh2k792H8NAzDiRYY4sZpMYj+AYkB1cOd1GNkb0l/XPICPVaqbzRSl/zYadMR65d+aZScSD0hSv1E3lOM9MFrDFN337Joj4nldEodVxnfHl9PGaxi2hB8sPRGaeXs7Ix3v+vLub664rlnn8XM+NJ3vIPDfs+zzzwLuCm6a/779efsyUidXPTtj/VOk08zX9+OwF0M85wImDxAkhIp2S3SIj+n/HCRCpkxH4+L4b2U4n6J2FLUeUYExmFYnsMUqtYl3ctCljSURKbEgO0f8lx8oG+qSCeKsCRMefu2cjgc3TxePImsDIWLRy7QBvM8sb/Zc3X1Kc7Oz9lsRjbb7UJyTZV5PnrgQE5R9hhEKRLD/ONgYcju5CJIGm4A79suWYiI3GoVhyzJiaek4DC+Qdnfv8/NYc+r91+h6YyUzOXxAM+/wP3rK15+7WWub66YWyPnxDQdvcgvZ1qrzFqZ5fR7tGLFihVfTPifvvv/5Es/8B5yOUldV6xY8fqgqvyl//NP8ewHf+oz/lwPTTS+9G1v5969e9SmTNMcA7if4rq5GIQcpWZ2Ohy2kJxYpTdVGz58pZSCB0gQBfAyO4IIKISMqX9fSsl7GsTjbS20TCfJCv69KQo68O1DE8gx/OaSo5DNlp/JkunpVbpsNbrLImGii2/BexhavH6N0+xIoOr9FObkJHVykjPveueXMQ6FD33ogxz2Nzz5pqd44onH+cmf/BdcXl3RY3pTSuFf8XuhqtRIWzLzU3AxLxu8nULlr8VQldgaJZLgcap2OlW/jd68fZKUPVha1wnf7VbyEv+C15CmzfPM8XiM4j4f2NUMbQ8a27v5O0X791wncikMw+DvhypNXbYkapCiFNKUphaJZ7HxqI2jGge/8wylMA4DZ7sddy7O2e+P3H/tVa5eeZW0KeQycH5+zvnujHEYqdaYm2J1IkkFcSlfEiGJE9Cc07LpWZq9w7AtORrkb8nszKL8kCCIIcOr2hYCKXng+Ree44XXXmRulbNh4OaVV/jU8V9SzajqyVg5pyBeG6wp2jzoYJMLpuv/sa5YseKLE2ePXFCG4XN9GStWfMHCzHjmxz/M9/5f3/7A/PiZwkMTjf1xZjxMmHXdvp9+92ELYosQGvc+lHUJE+ZDqEuTlCQhpfLpPr7X1w5eheCipjorKRWGgWWY82HZ8II4T0mCWzGmqljTRbrC4pWIlul2q3fDPDlKUgLRpRl8yc8OX4aExEsNtNXwmKhLnU5rFNqyqol/jqSrd7z5aR559C7PfOxjPP/CSwzDyLu//N3cXN/w0WeecYN1EIEUxYcWw+UptpYHSEDvFOmyrF5gZ9Y4Re4+KKfq33dbGrW0hMfP9+SrvllJ6STZ6puceZ6W7yklOiZMqVWpooscDrNb1+oRs4TPYT9P2OEY1+H9H7lvipYtTRQvmr83yolY5T7Mm/tDaquUkilZ2O02nO+e5jhNTPPEYZq4urrmcDguHR3juCFlJ4haGzY7SWxNY1vjKWMSJNcJRnHfS5BM98OEsVyNLBGBTPdpZASXf7300ou8+uprHKcD0hrn45bNZuB6qsymi4fGYrNWq3dxpNiKNa3cbkFfsWLFihUrVqx4PajzzHf8gT/M/Zde+aw830MTDR/l0yKFcl2/PHBC3hOFUkigbn+tafM+PAmfRnKJFRCtyr1M7hQpKkjEky4OdH+spm4gD5LTfc9YW5KTRLw+zmKo9s4G89NqgvDgg6GfwDcw9yo46SH+G3GvGulSYkvLdWu9Z8MWLT5AFsDSUmr46KOP8ra3vYP7l5c8++wzmBlvedvbISk/9mM/zvX11XK/zHyghFtek/6al/v+IGnoJXUPGMkFBAVzT8wpNrj7MPxe1lqZ5/kBmdTtx3ECd9p61DrTW7q71CpFc3irDVTQ8DOozUtULLhUSRGaGohRW42hfqCIv5+tGankiCG2kLx5XK7Fab/260tCWchjYzo2OLp8K+eJcRjYbrZsd2ec1ZlpmpiORw77G+6/NlPKwDAUzs7OGMdh8Z+o+X2Z5tnlVfEZ65I5J2clNjtOkA0np7b4dPxD2VvR6zxzPBx56eWXXLq1GZhb5ebmgPNyDytQU7R1eZz3lii+GRtiEzSM48P+2q5YsWLFihUrVgA+J/7oD/ww//Cv/M3P2nM+NNFICU/nkTAr49uIflr+QPt0Ti53j22GiDDEqtOlRCdfAbB4JNRScInTkGvxmH12WwZQU9+E9MfDNypqCiRMo6VbTj8nJciJRAt1cgO3qSzpUjl3+VFsQQAJ6U4pCdQ7EHzz4dubXih4O53KzKhauXv3Lm9581t48YUX+OhHP8p0PHL33qPce+wxnvvYJ/jEJz5x2iD4DXzA3H27RwROcb8Pphv5Fqd1n4V0AhhEIQjiUiKnLmu63Xp6WxrV//rAhiiu6/YWBGBeSvHck+ClfwRhcilRCeKQkzDPLQhDv0b3egzDsBjVmzW0zk74VJkMJBdv+jZzC3/4INQ8urfLmehyrqa0NrHfT+TYcGy3WzbbLZuzHWW/53h9w+X9+1zev+TOI3cYxkJOmXEsjJttNHLP1Do72RCXU92+B5LyQrpOBY+3i/683O84T7x2/5Kp7qmtubQsyCsoRRK1OaHJJaHNwxK6dyQlwao3ire0ejRWrFixYsWKL2bclj09rFfp6pXX+Pbf9weZDp+94t+HJhrC4FsNOxXZmXi/hJ8uR4EdiYQsvQ1dHvXgcBzm2qV4zDcQLEV0PQ3odKq/FJbFMO2SGvG40ZSXbYIP2rhMB0gWJX3xH4LI9LBXH9R9X1ObuYk3Nib0BFQqqQxgDTUlS3nAYL3AegmdE4W7j9zl3e96Fzln/t+P/DSvXN5nM25597vewzQf+MhHPsLxOKO0xVhcPF4KpXsg8EE8htnb96Obj5eNwS2yMwyn9I15nl12dKsDwswQPRGJJR53ub+nrUbOmWEYvPE8NijTNC3Pn3M+SbqWBKcTkmQIArK8BoGc3Mx98nDM0QwvWPJ7UHIhq/q9aA3J+dShEn92Irt6emYzWjSpNwO9Vua5sduNjMPIUAaOw4bd7sjV1RXXV5ccD3skDbz5LW9GUsXEvNApedb0zf6GVivjZvRWezVKVkqOz3Hqv1InAq00tCmX91/ltfsvMGvF0i2jvio5+WfewjgupiRRsnTTuTDHvW5WSayFfStWrFixYsUXI8yMm8srPvgjP8qP/uAP84Ff/ov5mm/6eu48du9fSzjMjL/yx/8MP/3PfvyzeLWvJ95WXVrUPQMibnymS51aW1IgfBMQaT2AT2Dxc5wkQlFp5qfRnMzX/fg+C6dT+lsn+SKCaXgzBFJIV0xuDf6pV+Ypak6IEKHEKfnCFHJZTuCNiFXtDc1qQHPPhDZUnAx0X0pPocphBE9J3OhuwvnZGe/7ivdzfn7GT3/4X3BzfU1OA+97z3u5c3HOP/vnP8Xl5X0QXTZDPW4Xk2VodmWWLfKmPvQDQTZqpCCxvBceEdt8c6G6bEy6lO12VDD9/SDeq9iMfDrJKKWAKofjIQr2WEoDT36ORM4nktLJR39sJxH9eeOTIYK2xmGaUE5mdPeLzMwilJzJKbs/pnoZY8a3C81gnis0dWLIyXtj4ZPJkYA218r02oSIMG5c+nQ+nrPdbdnvr7m5Eg77mRdf+BS7izscpyNmLt8zjJdeepnjdMDI3L24YLvdcnFxRmuVzbgh5UTKGmWITmC1Ki+//DIf+dizYMLZ7oybmz2q1b0x6uEKSI3fFzf6i3hwgrZORKKnJBk6rz0aK1asWLFixRcLzLxD61/+9DP80J/7Pv7u9/xVPvbhj9Bq4y/9kT/Fm97xFr7h1/7bfNNv+Fa+7Kvfz7jd/BzS8fF/8TP8hT/0J5aD+c8WHt6joS3aiXvqEmFEjqhXX22E1OkWITDoRXo9IlUWmRAxXD1YQOeSJHdRpOQN3H1A1aYYGrKf6HzIme6U6M/TSQ2xBUkxfNbqG5mEIJbQ2SVE1pwVqUFScMN6WNJjeO8JTLcjXyW2OilkWNIq4/aMNz31NM+/8BIv/PgHee2VV5hq5a1vfivjZsuPf/An+emf+gg1Yl0lyuOUTi4kNjp+It/vdTdrq+oyjJu596EI1Ih/RTwSWNU8BnYclwQp8A2DpNMmoW81aju1nPcekxzpSsfjkTrPS8xv/7luFu/Eq5SCNSc3niClzHVeekdyTpRhjCQpZapTGOo17rN/AnJOUaTnf95q9Z/NxQlWeDp8K9CQJZr2JDuTLqlKp01Kw5u1D/uJUpRxHChD4c4jdznb7bi52TMdJw43e5o19jcH5ojWTWbstjsM4TjPXN1cc//6inEcODs7Y7fbsRs3nG1GUs6eGpbhOM+Ukjjs99RDZq4ziDHPldpmep7CkNxgrhL9Kd3Qb6DNCdZa1rdixYoVK1Z8ceHV51/k//hv/kf++d/5B1y/dvnA11SVT370Y3zPH/w2vu+Pfgdf/jVfwTf9xm/l67/1W3jq7W9BUqLVyp/5X/8IL3/yhc/6tT/8RuOWRAfiJDrSpWIp4cPxMoz3DopwVS9Dk1FvNVn7X241W+ODY5Y++MfpurZbA2RCJIN45OnpsuKUPLmuvTVFKyG5agtZMYRqitBorZHx9m1DfEAUJWdnHU17GlZ4P8zImZ/TAk3IxMbtGR/4yg/wxOOP8bHnPs7N9Q11qnzJm5/mPe97Fzc3Nzz77Ef9eiT761PFUk/k8jyj5V7HfXApVIptgm84+td79Gq/DYu/ZJGvnczfbmI3aLpsHDp5SSmRY1vSt1a+NQm/RLy3tz0kcOrdWC45vrdpQ0xIaSCX4dTvEZsxtMV134oabka1imry1KcgdU3VX3skQvWAAItSOydeIceLpYmFgdwDAnrSVjxmbG2myQsEh6EwlA2PPDJSa+V4nKhtZrfdcXNzw/E4cTgcOByPbl4fCtny8hj7/UuMmw0Xu3Omix2b7ZaxZI5T5eWXXuLm5oa5VrApwgsS83GiCew2A9n8Xta5UpvSdMYMzjZbUkpcTtd+z6sgfGbLdVasWLFixYoVnz/Y7HZ81a/4Oj75zMd57kM/9a88dJwORz70D/4pH/oH/5Tv+P1/iK/+pm/gm3/zr6fVyg/9+e//LF+146GJBnJLlkKYkk85nh6jKi4FMU5xrBYSKuy2sflBr8FSuJcTop0o9K83MCjl1NDcH7A1HzIFw0zwZFRZBkzrJXuSIo60D9p1uQY/kfcBv6chuWzKX5ekTOqmkl6Wp37i3qN7W/gStmXgKz7wPp584jE+/vGP88EP/QTHw8Tdx+7xga/8AGaNn/iJn+Dq6mqRHS338pYsrMcEO2Prngp/DotV0GIWD2lZ690isR1ZmB8sqVTQ5Uu+DbrdFu4eDO8Saa0xzzOmvU8iL693nmd/Tzj5QXohISLo7B0rardlUsvHBDOXMJlWcvIkpWEozPNMrR4vqwi1nprPc0qg7pvRIKqjOBEhu16q5LwQtdtLQU0SqU6nrdht0iyRYlxTF/wAADJKSURBVHU8TtS5+X3IA7tdpqn3b+SUuDhXbg4HbvZ7DjdH5mlGUiJnY7vbMhQnHfcvL3n+pRe5uLhgtx0Zxw3nZxfsb67Yc+OfMTVUGtvtCHjD+lxPRHo3DGQZfPtijePxuAQsDMUQzQ/9a7tixYoVK1as+MLG2SMX/Ie/6z/n3/3PfhP/7G/9CH/l//6z/PgP/0MON/t/5c/cf+lVfvjPfz8/8hf/GsNmw7Q/fBav+ISHN4On5MZYOBlxuwl5SS/qJus4ESckQL2B+5YRuZumb5OOdHsBgvszckpgKQbphOQUcbjqGv3mJ9ZGnI6boL5aoZRhSa0y85+3pfsg3Tqd9+/JIQFaTOLh81AVVBs5vOTJBG16klOpUrZb3vHOd3K2PeNffPineeaZZ5nnyna34au+6gOMQ+EnPvhhPvXJ55cX2MmE9B4Mv3G0+HrO5cH7KT3V6bS9SHH/lnbuWhfC0U3aSwSuGbkkavXkqf71/l42deOyWgsfi0vVFs9MpER1DpP6ViHIl+HvoRvYvXOjy7D6tTZrDEMmSQnjsz/umAdu9n7q3wmgEyTBknlbvOCGfDXmfQNp7LZbxjx4oV3Iy4JC3ErVOjWb+z7LZXSLVCskVnOr3si9bIBcrrTb7gDYbLacnZ2jj0xc7/dcXx24ubmm1ZlxHJfXWXJiPh6Zp4mUbtjtNrzjnV/Ozc2eV155wdvLjzeoqG99BDckxb27bjMlFaY6Y1aZzftCsoXraE2dWrFixYoVK76oICKcPXLBL/11v5qv+/e+mY/88w/xA3/yu/iRv/QDvPKpF/+VP6dNOf5rCMlnGg8fb4s4EZCeUhR/ntz8rCGBEcLTQF8CPGhCJjYcbibWU8t0a2jIk8TCfB5Do6B+2h7DYWvquvzsRXKWwFSYm8bw649Tm2I2k3NZhmIEcikxQLMYusG/rPiJswCiS1c3OWWSQbWGSZABaxC9Bm9605sQg489+yw/+6lPcDjeUMrIe9/7Xsow8Oxzz/HRn3mGXBKSDPCeDdW2bAtO3o+I3F2ITyduTvhyl5mFLEjEOyx6alUKUphvybvK4F0VKkpP2bpNMrw34og1H9CH4hKqZi6xyikFqYlBPQlJ3MchvV0cIQOzhbypp435C/DPwC3PzEKAwhB+cecO83RkPx2oVU7+CzIlF3oje2tGrUdUjcP+yDAMbMaRYSxBbiQIpLix2szb4+lJZKftnD8//imLrUdt3mIvyQlTHtwbAjBqY54mymbD+Xllvz/j6uqK4zTR1P0/283IdjNi4tuXy8uruA+Ju3ceZ6579H5jPx2RSMZyn0wQyGQ0LLw7hSHeo7nV5b6tWLFixYoVK774ICKUceA9v+irePcv/Ep+0+/5r/g73/19/PX/58/zsQ//DNo+v7ycr8MM7uZdQn8vxrJZoFuxb0mpQJayPP8jW4gG5DjFj9hSvJcjpxIkxnX8fvKcXYbTiwJbRVLxhmbwpKbauxnsVgu2ggk5j2gU7PlA71KssZTwmPjleyN3dllVRPcmEil8CupNcV46ZxHxa0YaBt793vfw+KOPcTwe+ehHn2G/P7DbnfOe97ybs/Mdr7zyGs899xzjZuTsfBu+CcIE7ElNGKRcQsKllJSx5klEzYy5GUkbJuKFbUEOTDy+dkgnOc1QCopRhmExTKc+6GdfR3Rp0zxXam2M48A5W6CndQHiRXglZ4aQEIF7OpxUpiikq5QcvgkzhrDuVG0L0eibDTX1P8djgk8JWd5Hshk3mCk3NqMzgISRvJcV+vO3MJyLCDSB2Tdg4zA4YZBEHkrEBRMmdo9eJiKXtbVF9hbRAm6iD9+RmTIdZ0T69mdg3I5s8kDbbLnZ35BzZrPZ0lrj6vKS65sbbvZ7prmy2229+6MUJ1hNeeX+fWpt7HaPcvciUduB6/2ey6tL1GbMlJwKVD0VMQLazBvV86k3ZsWKFStWrFjxxYmuCHrq7W/hN/7u/5Jv/e2/mR/9wR/m+7/tz/LBH/lRjp8jqdSn4+HN4CSQHF4B/5NTm3f8SchsegLU7TI5kpBNwqjcWAzgqqSSqSjFBIsT85SHxch8KpOLro0+6KvROEldXKblMbxCQU2Yp5ky+OCd1HyLkITaKoZFdKgP7Tmbt3qLxJaGZcgViG4Ho8QwamaUXHj1lftcX93w4osvcP/yPjlnnnziccbNhv1+4rlnno3H9+1DRsDU+xBS8jLAlMJ0nTxOF2Oz2/nmpVUkgVkGjGHwAX0cN+6H2TQ3Tkd6l5gxTRMJfz+qGhUjSSGJnfo3avNG7rBid49GKmkhjTk2KB6jO5BSEAYEqw2sUYZMTxrLCKkkjsfJB3ltlDyQg6Btxw2J7vOA1iaPbY22+CyJnAeG1Gg53oMY8n3krrTqccZLElknwShVJ3KCMRfaNEFsdmKtsnhjbnuN3PZyCizoj1uby+L8OWCaG8d54mK7YxwHhvIIc5uYpsrhcGQcBu488gjXN3teu3+fl156KYjIhnGzIefC2dmOaZq5ud5zvx25uDjnsUef5OL8Ljf7a167f59DnVBzEplJbHOmlUjfaqdQgxUrVqxYsWLFChHh/O4jfON//Gv4pb/uV/HT//Qn+Kt/4jv5+3/5r/PaCy9/Tq/tdRANTx9yU7YX5BmncrcOn9MTTiTSQky0KSQfaEWNJCDJpVXWxI295n6OlnwYLykh2cmJn/InhpLQ5jGfgp9cq3aJjw/jZhmf4S1SpGxJhmpqqInLh0zQZkh2yY/g7eTdoF7yKRUql9i2AKSENT/R16p88mc/ERIk35yUlHn+k8/zsz/7KSQJRU5t0i6acqeAoU60kg/Z27LlbDOSy+C+EfENjrbMrC26NIZI3fJ3ZciRwEViKJk6zy6xKXmJhRUk0rkUmkuJfHPkrzWFTMppjNEaUbQImN/bhqE6g2XvDSGIgjg5CN0ZVRVL3n3hW4yQdrWG1eobBIO8+Du6N0GZZqUZ1KaMw0iSxLFOzHONluzYkAlh7neZXI/59SQrYdbKPM+cn+1OGzV1g/ztz6uIkCVH0lmX+3kiWTfUd+Lbuzlabbx2eQkJduOG3W7L7mxk3GxprTLNE9ud92tc3dxweXnJ4ThxdX3NdjsyDFtSEkpJmGX2+wPTsZJS4u4jj/HEvSd55f4rvPzqi+HRaO7JsITQyOOI1DV1asWKFStWrFjxICRUL+/7uq/lvb/ka/gtv/d38re+83v5G9/xPfzLn3rm06okPjt4eI+G+P9Y+Cd6sNFtb4GFBMlieO4n5WoKatTYKIAnCyXr5l1Da4Pkp8jJEikbKkCLk2ZV5jCAa/Ro+IDtRu2ytIaDq2UiCclStGy7z6OGryGnjKk3iycSRK8DptEkHi3N6oO0OidACTN2s9Dz6zK45uKvbRgG5rky1+oEytN2vXE8J0yEeErq3MhFkJQjwSoxBCk5HifarAgufcrxPf1+51x8iI8I2v3hcDqpj03SUIob5HFfS4sT/V4ChzjJ676bZCfTfqv1wYLCeG+bGUMWGhbt7i73mnuzd3ZyJWHK75svw2h1xhRUEh6QZUuLeVNlPk5AFBRGl0bvmejlg621xV9CCL1SbNDMlFp9M3S9P7BVZcguXXLic2pDBydzBpi4GEzCC1OXz7B/zusSjyyLpOzq5sD14cA4jux2O4ZhYBwH2mbDtNmwGTfc2e04ThP3719ys7/h+ngFt8zuOXpOAI7HiWMwzMfvPUUuRp29kbzG5meaZ8YghitWrFixYsWKFZ+OfjD79Dvfxm/+vf81//7v+E/4xz/wQ/zlP/rtfPgf/VjMWp8dvC6PhqnScFMw0sv5+pAXaVLWIpUqIoUEcnRetNZuafVdxlJK8SbqMGGXVPxxm2ISnQfhvbDaqBEDmnIOiVGYosVV9jlHi7d6ilIp3o6YxDmLIOG7OEXbtlqXwTvJKUlJzCVEzRpmLYoJe6RsxOrqTCp9GHeiMdfKcTpGT4Ux1dmL70wZJLv8hT7c+3AL5tKzwSL+1f9a54qYUa2hRhjEI9moVmaRMJU7mWsoScJCb6cIYY33JSNocnmUzrp0W5jqYnxv1ihJljQmN4D34ODelC2oeBFfbQ1T77BQKljGUqYkg6TuwBEhx/NYurUFE6OGh8ZUGTYj0zRFsSMutUrCNM3Le9p9M93I7tuqwjC430MiSkpV2R+O6KjuMRE396dU6B4eOMmopBPpSAIz8+Qv1eZSuSg9NIMsTt6amkfeHg9st1vOdzsQ9wDl80zdbJjnmbOzcw6HAzdXLo+6OR7YHw7Ruj5xtt2SS/bggpy5vLym3sycnW159NELmjaur684HK+xtm40VqxYsWLFihX//xARLh69y6/8Tb+WX/brfxU/+Q9/jO//tj/LP/6rf5vLV177jD//QxMNL0zzQUty8WFfZImM9QQkL0KTXuimcRKMD6GnU+iuhncjc87FT3lzT6nKWK3UVhe/wtw8nacU1+x70qnRxOJU2jBJYTR2KUwiIeIyraoWkqMUyVmNNtsSDesSqtRV+phVWhiPh2GIaFRlHDckUebDRHc5t+YSLtfPZzdnD8Mib0olI5JjyxKbkOqN5LmUaC0n/nzGWqK2Ro2I2aEU36xwihK22O4gQi6J1kBTco8ETrQkdWNzQ5unRDWt0Ii1j1Crstm6Cd1qQ8MEX+fZ5WNFljhgEWEYSgzcyjAUWpjFU8m+SYjyQFSpyaVJHtkbA33OEZ3r/SSKF/TZPFHr7OlRm9G3Zy1ka5IWKZZawsIk7QSrompoU1pqdHe/Kk6QxJhr89Sw5O+HWqLphBDxv3Sj+Yn8eH3LLY8Rdvr8i4C4J0aWrZ1xfX3NzdUVKRfONltKcaP/uDljU9UTrHJid3bG1fU1L7/yMjfHA4fDkel44Pz8jM1mR86ZRx65AOD6+prnn3+R7W7H+fkj3Lm44HD4/DB4rVixYsWKFSu+MCAijNstX/WNX8dX/vJfzL/8yLP8zT/9F/hbf+Z7+eSzH188qW80Hr5HQ05jlVqYiVNCzIlEN4Q/kDIFrt+XnuvDcgLtPRg50qC6m9wH6Lm6Zr1V9SG5Kkl6kZ/7EfzEW0k5e88FeGxsU8wSKRUsZE6zWhAglyzVGjKY8Dp40pQTgalWJKJnwcmTy6OUlIU2TeSSaWLkW4b4lD02t822NHubuWFdBiAZKQ1M8xT+BMDcH+KyMy8MnGtlLEJtBpJQjFnb0mSdxO+Z5IyJvxbivdmMQ6RXja7LQqi1Ms/QtDLXiZKzG9610XtOaq0hhfNtThIh58I0Tb5Rwf0uQ8o09W1Lay0IZaSO5eTSuVyYjwfImSwZKaNvA5rSaqOZ93kUcb7TFEQVbe7BqZOf1g/DgKbm3goTNsNITRWD8Gy0JQ5WJKKUI1HMPyNOGpIktDXq3KD49i1JdLIQkbLa/HMUJFhbtMr3jgtOpYrBoTCJzYlEy3jVxQc01yPH45FxHHj03iOUUlD1rUbOZUkL2+w8uerm8pqbm2vuX14z7I+cnZ97PLAkdrsdpQzc3Nzw/KeeJyVhG70eK1asWLFixYoVrwciguTM297zZfyn/8Pv5tf/zt/KP/i+v8lf+xPfyU/9kx9nnt5YWdXDEw08Wai3fqeUEXOpUu9I6Ke9RkiQQjMvkpatgzdztxi+nGy01hu36+I9aG0OcpPJQyFH+lRrGoZ031DUqvQQHjPvKpDUpV1+ku5yIENjiyAp0WqjROGfpITRSOaGdEmgNCS2HESb9mbcMh2PTMeJkpwEQGHcDIsXAZyQeAGgYOHbMAQvDYRxGD0dC2+FLsN4in1Vpam3WUsk1qo2MpCH4iZoi0c0i94MoTXlZr+HFm0kkXDlsrFEKYWMy5PmSAHrpAJVNuPoW6vWmFUpZWCz3WIt2tRtZq6NGsV/zTQSqPzEf56OHkEc5v95rqhBs9mjfLXFjQ2SEsWBRGs3GLMpyWCaZwa87VvFh3dVI4uwGb1NO9VE9Syt5T0Cb2mX26WGETi114mNDQwkGnN4NUIipV4I2LcXft9638tpq+EeFb/3vuFwoia4CqukvrHzFLObw4Hp+ZndZsPFxTnb3ZZxDGmYKvv9nnR2ztl2x/5wwf3L+8zHif1+78+bE9txwzgU0sUZw1C4f3nNq6/df/2/6StWrFjx8wBJ1tS9FSveKIgI9558nH/nt/4G/q3f9Gv54N//J3z/t30nP/qDP8T1q2/MrPHwqVOxiWgouZtmRSPN57QBOJEOh58Ep0hQUsg+pIY2hdQUDfJi1oBMEiGlwjAMi8E3KszIuTgRmafoXwAZcvgPvJhtnma/niTk7CfX1uNv6bIjNx97x4T7NeYWp+QSKUVAUj9dx5S5zszarcf+uJI7oXJDu6qiccKeSJFcBZIiq0mJKFaX/gyRQDTNjSTRoC4FlYa1tpiohUSb6xLr6ifrbooWNd/KxDCcDKQ1Zp1RS4yj92DMOnn/iCpJcpAjJz/TNC+N5ENKy7A+zY2UQXKkM7lNnpJcYiXh2cmFIImJvBVqNao2ap0oKZOGTJubN4GLICYMKVNp4ZWQaF73MsRaaxBa34hVXEZlzc3dqjOiRJO5xPYo7mukjBmEId8b7Q/zxNyEIWdyhiQFk2g+by22S06iS85UvEjSSymd8KTs9975lgXRiOfSU0mOhAStqXJ9OLKfJs62IxcXd0gpcXZx4T8zzUzTxG6zZbPZME8z968uub66Zq6Va645OztjHEe2uy3DZuR4PL7uX/QVK1as+PmAJ9769Of6Elas+HkHEWFztuNrv/mX8dXf9A18/Kd+hr/+7d/D3/muv8zzz33igXTZ14uHN4Obdzl48VuC5OV2vS+jeym6HKpGFGyWvMhyerpRSYXW/RcJl+eEqVkkUYb8QASXS1bSEjfbWlvCqKTIkrwkxdu7U3YZTZbip+WCd3dYXoZQKCylgdqI6mjKkKht9s2IJI+FbWDJqLMPv96SDdvthkcff4xPPf8pjns39kpOSE9SCtP7kLNvgoChZLzj0KkK2Tsp8jBSa6VqI4tvgeb5NFBazt6i3jsyolW8l7p5KhSM4UORiJ3NcV8TMA4DtSpqbg6v1pCmTK1Ryri8P5B861GrkzxNIf1K0ZpuvkDAI75KGanzTIpYWJOEFKU0H8oxQ+eGmFMmEY/r7S3mKYirWkJKYlL/XieD/tlKNKwUFI/r3W1HjsfE0eYTSQ2i2n8dUkibtDVP4QqiPEtiGAo5956Q5J+npXCyMUTQQIp+EotsK0/JTaTUPUvCcW6kFBuYhUT65k4iOKE15eb6wM3+yDAObMeR3XbLZtxwdXXJPLuxXIaBe3fvsttsuL66Zn+45ni84fr6ysnGdsdus3343/AVK1as+HmEMqypeytWfKYgIuSSecf7381v+59/D//R7/pt/L3v/UH+2p/4Lj7yYx+kzq8/jOZ1eTR8IJcodIsTZnBDdJyud/gpMC6XCWjz9mcEdO7JQ4myGKfTclqfFqlUO0Xiii6RXTmfTNc+IIv3YpiRcnEjcUSihlWcklyqRMrurYjr8rheIZdYeeAm6TIkUhHaVLHWvFNBhJIyZRgYxg2Xl1fUyUmT652KD5bqp/5VKxaDsoY3RK1FXDA+XMYt6klLJSeaNobsEi819dYKM3Lq8begXskXGxI35jeFkoXtZkQZmI4TNlemMEk37NbzFMqQ0FYBj7wVvDwxidCSvydEo7bENSISQQB+79xULcxTdVP6wnw9sasP+iVlb7aWRGnG0hkeMbJmjZQLpbkRPKdMTj6kF0k009gmNCwpm5wpZ4nDNDMdjbk1kjiR7RImf09SRBebJ0W1RmvKOI60FGZ2S2RzGVrVhpQBS9LrQehENIK4lvttZoxBjHtT+e3PqJknZbXWSMXjiI/Hif1hz6YMnlR1fo7qjtYq19d7bHaien5xxtluZH+YuNnvubm+4ubmhs1287C/titWrFixYsWKFa8bIsKjTz3Br/ntv5lv+S3/AT/xI/+Y7//jf5Z/8jf+Ljf3Lx/6cR6aaPQnNaLkDl3K10opHp/aT48RJMdgqM3L7cQQcTNxKok0FIY0IAhNa+jcDax1G++yCekkxqNpu45egpQYrbnWvsWgBy6l8WEv0qDEtwukgSJGFU8qEgipFotUp8fXbrfnzMcDGpsbxRiHTQyPys1xjzU3jHefheAdISLQ5ikazNW7PySKBVMm5cQ4FE9MMuvWBiR7VK5LuIQ0OGly8tOoeMM4qoh4H0gZBAu2MiY3QtfWvL1bhDKO1FYXWVuRTMNo1Zha84QnNUwr4CTHU5pyRA+zSJEMSDm2JgBC9GcoqXhsrNYZNVlM76kUNmUgGTTcZ5Oz4OWKDUMR80hYT/AChnyKru1ctZknbKmiLXPURh6E3XbrRO9o0E7FiZ0Q9L4QcN5hZszzhJmGPA9K7rHFvr0wnX2Tkr07w31Hye9j/L2qLYSmfxa7R8lXHyzbPin9cXukbmKqleP9S4ZhYLMZGYeRRx/1ONybmxsOhwNzreRhw263Y7/fcn1zzWF/83p+bVesWLFixYoVK/6NICJsz3f8wl/1K/jab/6lPPeTH+EH/tR380N/7q881M8/PNEIjbokoc3V04tiuHLpip+slyAbboZ2GYul8Gfgg3ROBeJkunidtxu0zdxfgHnJXx/bpRuGS8TJelP0MIxxaaeT7k5AoPsP8pKE5eZnl+jk5IWA1pQqhugcBna/vmmuXF9dugY/5Fc5Z6xV7t27x9XVNdP+4FKxnL0QrrkJO8Updsl+HabqcbJJXF6ELDKdJL0gT2nqJKsBOWXv0zCPsU2SIGeqOtGo1ZOoSs6M48aN5/4SkSTUWWlaORwnhmHE2oxIQnuyFMZm9M4SVXXzvEa6VkiOBDeKuzleEW9kpKlhVpdh3uJzUVujIIybgWlqaJjNU8ohCfM0sLkeXYIlFjK0RKszrVX/sERsMikSqSKVK+eMoeRspALT0ftIWmsMY2aeXfZmahCFjktruBmoObHpqWcWnSUGNcrzUvSaJPESxYEREYuCxdbDrFA7dZBokJeeTNWjnv0Njs9AkCZJERpgERCAMc0zh+OBlDJn2w3DuGG7O0NSIk+Tp1SNI+Nm5M6dRzjcrERjxYoVK1asWPHZg8uqCu/8yvfyO37ff8dv+G//i4f6uddhBg8Biak3W+PkQfpA1b0AyZOH6lxRaaTipuNeRKY4YTHVSJ6KTUjo+1Mk9/TT8pxKSLYctdVIUPIxDe3pPyeJ1kJgREJ61ZOuYvDE41RFEpadOCVJbHaum79/eYWIuaTIjDIMlCG73yFn5tZ8M5ETpWQfbJv/s5nGIO5N0z5Q+pBaUnKfS8oMpSz3NeEbiXEco3zQJVGGn4oPOcdWIzFGWd44bpxs4EZuESct47ihNS+Yq7X7XMybx4GEUMPA3+a6DPOSMqQU0bzuW+mDtad5uW6vm+RzglLyLblQQq2GRAmXKdV4/wW0NsqYl41OONG9uDEkcZ4sxpKIFW+vG/rFH8/iOjx9rESSV0Iks9m4pGgODWFvUO/3U6N4MElCwv8hEfM7DCVkde7LqbXR2kRrlXEcyGSPQw7pmMb2QkJC5jHHXaon3p5eihPvSEvrJZfdt+EMKPlrFaG2yuVlRfKe7XbLdrdjs90u2w1DQNIab7tixYoVK1as+JxBUuLxNz/1UN/7OszgFlGy4olFoqSSFy9FjphVw0+Hja6xd/kNIh4tazDXmSwpTpEzrbocxcyYqpOKUopLhQQ3eJccUak5FEAhnUoG1ZOTPCq3d2TIcu0+hLt/QRLu5YgoWrMoxBPPrTocjsxxDX7CDcMQ3hPzcr7Lq2skiScT1bqkahHa/1zc0VBbo+QSp/duzp7qFF6WxnYc2G427OejbzKCOPj9s5D0FO91wKVI2mbysKHkzFBcIiWRkqSzn37vp8lJVUpk8uINqUHqxlI4Tp7ale0kQatNGXOP0FVyKuQiTLWGRKoi5kN0Hkc0PDAJobYZjc1BJnkqVfGsZsOlZt4grmQSqXgXyDRPbkq3ThbdaO79GY1m4o3xUeroRgmD7NsnVygJFj6QzritKU3CBN4aFmRWY/vVCW/fKNVaF4/RrI0cyVvH6YhqYxxHNyHG5+q0veiljxIyPl02MgkhJ09E88+cE7H+c0uUbvwnSUKTUuuMXjWO0TbeCdTxePTtRjp9tlesWLFixYoVKz5f8TrM4AnEk5NSchN3l4ksptdIQBIRSJmUi/cuwCK9CuOFbxu0RZTsKWEqRRKRaiNHPK23XCfExe9YnALXpuQ0kLIP9WY15EMCTREsfAP4Kbkakn2wq5Nf5zgUNpsBEWG/PzLXCmKU7P6RcUjkUpimmTy4p0DVE60E75Fo6iZz7/1wMpLNZWIJc/N8nHonwleRvCm9ins0kpdS0GolYdRmSxRvUyMNJWJkB5dTpcI8V5dptYaipOTDeMkFrTOzurHcq9mNMWWP+QWKJCS73O04V0/q6t4N9VZ0l7ClMNq77yEXwapvjXJE9x7n47KRSAYaxm+zipiitXrZ3+ySN4bsTd7qW6bREnNV31ZFL4a3uQvzNDFjlDSzKSN9GeAfsQRVXS4lArMb6EsqzPhQX7vWKRKvCOlYN7ZryKgQJ2RDAsRTt4ZhgJS846TW/rFFKNEj4p0tPXTAZXuyeDeOcyNrjl4X3+pJdlIjCCUPQT7ash3B4jNVjVobl5dXHju3Gdntdss2cMWKFStWrFix4vMdD0002q1EHTML6Ye4bCl5cVytcWJLdD2E9CbnYfFPWCT3NPVT9iRCvaVvh97H4MlBzVwSJNHg0JpSkst7JAm1HjHzmNhk/eeVPHixm1kLKRPkpVHcwhsiXJxdkEvm5uaGKeJke09EjkhZNeVst+F4mKnmxGooOSQ2RhYjx1bEzesVbTBIJvWSw5h3/fsUE98wEClNBZAyeMpUU9+yqIFVlOSEwjyBqpTMYZrA/JRbzcsBJWU3zZtn/EoY9D1y14dhazUG7L7lCalPdXfIQWeGoYB5AeA4DAylkFNGTahTZRwGjnONTZQ3hksSapjhTRVNyZOyFPdXlMx8OLqPRxWTXtIIQxI0Z2qdOVoFrZSSGcaBWitTnZ0kqmFDZiwFbbicKmRHOSVaMY9TrjNVTmRoMbJr731xD4yEtK62Ro7P4KQgybtVWmxStEUvR1UoGd8vuRxOozRl8QGJYOKN5mLdn+OSO5MoT8yFSo3fCSdVHgoQcr7oNrmdvrbf78nZ5WG5vK4MhxUrVqxYsWLFis8JHt6jwSmVyeHDVSku2zFVylCgNvcO9HRRM9AWhu1I5onhSUSYpiMSp/IieSEcfTvSC9taqyFTyeFdcFMyiG9ZAOx0TZ9+9W46Lmh1L0Uz5WyzAW28+uqln2qbhO+EfvEcjhOGMU8zSGIzjr49CZkUArUZQ3ZpUzWP0iUlJCWPtzWQlD31yhRpgtE4tnq6nwa1Tag1VCzKCMMXkyP1KHUyJbQa7dYJsjl5aa0xTUeKZCxBa8aES7+SJOapQjKGssFEkQxjHjhMk/dbWKMIXhRoxtl268QwJxSNLRQ0C5t+vH5FsWpYdF7knLw5Pd7XJCm8FHkhByk2K9oalRT+nkICmjVaM0xcXgbuOZmbwTSTNk4C6nH2PpC5UsYBEd9amCQ2o3tDjsejb4TC+3PqZwnJU3iIDBbfhamXSs7TTMkuo0okptbI2AMxzvM8LYEAPae4p67VVr2PRaFZXUgzcW+6kV57W7r45xkiThqLzU1GG07k2+wkecWKFStWrFix4vMcD000UkpeGMfJZOtyKZeUtOan8xan6z4MWZy2Z5pEMV8SchzXLoVm4begn74LiCVSdoOun0JnUjZa1fgZ9YbwFuGjtUKWZaDvW4upNmqtbMYNIp62NKTE+XZkrsblfs9xirK9GDxzSdTWTb9RSkgiiTE3T19SM7ytHE91MvcgIEIiU4biJuCQ0ZSU3eMsTpKqVeqsJDGXQ+XihYVJaXWOE3aPExaE1mZa83slItQ6U/IQcake71uK38c6z+RUGKjklKkRctXEsD78Ckwxcw/DgJOXuhBAay3K/pwENHNDs3s5GuCFeqY+3Hfzs6TEVBuH6xtyiZ9tirbGNDVSwbcqvTMl/oMqJXXyEVuhqdEQJBdPIxOB7JuYoSilbBZJk2FOVEywMTNNE5txAG0cp9nlSZxIrKrL8vyzd+q8uP1fbYq1mZYSLStDRA1rLxHM/jtx+n2Iz2bf0OHdM5ISdfbPUcqJTImEr24o7zHOihHyQ7ylvMX1xS9hmMZX6dSKFStWrFix4vMfD7/R0GhCFiHd+nOBaKjuDeFCzhFDWyvgg52hpOyDsqfFphjIfLhf2qyz+zMUIu7Ullbnk9w+tPVh7B2HFKfChrZ5kdK4sMdlTbVWZEgYQhk2WBKO9Zp5qihQqxu31Ro5D0jx4W7MGw6HI5JcIlSbMtVGKT5QuiejRVqUuEwmncoNXcMfGwkEUSUVUBPOthuwxqyelHSYjuSc2Q5DbFYSc2thYHeC4klLRhlGjw/W2U3r4WepzShp8AG1bMniz1lr42x3xvFwALElsSrnjNRGyjl6LNw3s9vuICQ8VRVaFNbdkvR4MpifyLt/R7Dq2yskuiW0RxXDuCnuV9AgF7Cczvd2eItT/M0wUtWlYnWaKbnEJsqYj+68sJLIkhnGMfw/ireH+LW1qgzDFkmZ/fHIPM/xvDHEt/DG8GCjt8ZnXYNEuS9HFwmWiHtfit3+3OYIkorekeYEI2W/B53IuDGc0+OZpzdI6lfd+zl0CSwA3xgi0QnywG/gihUrVqxYsWLF5ycemmjkUiil+KmrWgw90ZWhp6KynHOYvEPSFKbvlMIA3jxpJ0Wqp6kXvpWh+CmuapxO5zipZjGU55xiMPVhXnLywjxJPl6Kn+znXGi10qyRi7eAp+LXgRk3N9eoVWptaJhGvLHcyGTK4APkcX9AMScVpJBX+cm2WvVkIjVSIjotBpd2iZHAZVTmfoCck2v86YlKRqUtpKSab25S9mZqPzX3k21tyuHYGIufd0ewL3WeSGZUa4zjJnwbRjNhGBJtniGM7KpGTo04wycN7neptdFE0eMRiXtseCHfUArDUCh4b8RUq28qJLtvYXCfi2kLyZIwlAGdJ2eg7syPuFldiBIQfz+A6fKzTVt4GowxBvOcM5orSsM0iJuoFxw2IMvSQq8qZCyM2tklXVp94O8fOIh73lOi+mcr2ufV5WApZW9kbx7V7Olpp22HAGUojNFM30m2RIKXhfm8bzW67DAFccE8icuM+LPeG+MbOpFIVOufF1g2ICtWrFixYsWKFV8IePh4W3UPgCTvIPBNhht9tWlo7tWjSVHm2ZZhzU9zXdPfVei9fTvl7F0YtSK5oBFV22Kk9lN0P+s184jSknKYZn1I83jdAUOpc2PIToyU8IjEk7aIZ53DrGvmvRQqbixPOSEmTFODkEwhhjY3nFgkb41FMBtIIgw5EohGP3XOWdCqaE4QUbumFWt+Lyw6LFJJp1N8gZwSm7NxiUgVcZJiMWDnkk79E+F9qbV6+3Z4SXKRSLYCqgIamv6EJGd1lmWJpDVcYuYxvn59Es3oqk4szM05SPEiO3EnNdlgrt4Cn/PJmzD3qGPzzdb+cCCXgZIFIy19LLXO1JAXeZJTwnciHgk71RnVyiAjOQ8usGr4Fik7qax1RvH0LssutRMNuZNr6qKlXdiM42L8FsHJUrxOU6XBYuaX1LtD/P6oOdmV2Hp0Un08TszTzGY7UnJa/EWlDE6QYqsFvrUw8JJIw/1F5v4LzLDWvNMk0siAk4E/BSGJiF5vI1+xYsWKFStWrPj8xkMTjVKKD80xpLVWfbBDMTHfHMxt6bHoUaBJ/Ay96+hznOq6y7i3iDthEbU4eQ9duqSQpvQTXV1MskUybZ69myO5vl3Vi/pSETeeL5GzIfsiRURtDGoCkmGTi3s6xpGSRy9Hi6Sqqt4toWH0lSj7UzOkFO+mAAjiMx39mrRWfz3RRC3EEDnEiXf4HRAv3Ntst6i53CxJdpKkupCrnN3QTTRwqyopO1GiD59qiCmSvYNEEcax4FuFRukxq1HeV7JvgmoUImZxY7Sbt406TUx69Ab4KbZIKQMJSY0hZY/YVSdKTVskPSXvPameHpUTjCUzTY05zM+YMU9HhmGg+QG/bw0Ak8SQcUM9XjBoOAEQdd+CJKOqolXZ7qBQ6I3yppBSL/MTpmkC8ELE44RF0eTSIr/4MgAMCQuES958Awe9e6XLrRxNlZubA0PEJJdcmOc5TPBeOJkjrtmQxefSn8+7UTQ2bxr9H+H9sEi3apw2MivHWLFixYoVK1Z8geChicbcGjWia2u0GndZismpjblDwldRmw+kasZmHKM3wxOKVBrjMGCxlVCt7tfgJBNprbr/ICJUU0pM07Q0UmcpHKcjom5qVlPmuSHimvbtbsd8PHI4HMKoXBH1+NgWJ/stivVSHtwH0Kr3IBxjEA6jdM5xDTlHm3hmHAfM3OjcUErcDwFIMPQBssf0hqdBzVOpwNhutmRxwoAZ83wgDRuG4iRtnmfqPCMpU6PToeTCOA5MYXRPEfMqTanRTzKOI22eo6yuq/1xL0NECIsltoMTEyckdSFui2ciufdCyNTWlt6Hbhz30utGGnxz5QV6xmGeEISpKXUcljCAJAOSXQDWvRHDMDBPs2+omi3P3yykcaouUYt27lqNoYw+d6ss8jskI6kh6vIjjwP24T9JYiwDk07LeyK4r0bN71/qkbRC+CT8+vTWa+6SrttxzMd5pmnl4uw8tlhKizABkeQhBVrDg+JSMXBvikV3hveuEJIt9zX571InnYR/49/od33FihUrVqxYseKziocP5FdF7NR30Vu9TdWlOj2y0/oWw42zOczLqkozDV9GbCZq49Aau7MzTyeaZ5eL4HbXLBCH78vpc9e6d8nLPE9R4OZDacZL79xcm7i6f8lcqxul55mUXa6Vh0IZWUzcasbN9bU/rrhH4OxsGwOhkQ1KyI60NrbDgA/mk8e4Rps0SLRNg2Tzjg8KiIah3Uv8CEN1GYp7E8CJGL7dwRRX3twqRRQf+EseXAqmTjrUvI1bci86TKQYrlXdY5AkihDNG8jdbu9kY5rn5T3yfC0nN2MpLhGaK9YqOb5q6t0hpQzuL2kVwQsBMwUyTMeJJEIZBm9EnyaGcUB1prWJnAWRwmG/B3HimVPxjU1/TclJhZlHEpu556JIwlJCxO+rh441puMRU99C0P06KaO0eO1KCiO3t68TsjuPVV7KJcMj4Vqm9oBHon+9d3REiySiQq3K1fWecSxshsG7M5ZktvCK5OKeJjTIi5MMs4aooNZjmzWCCRI1CLFEatmSQrVixYoVK1asWPF5jIf3aJjr3YkhqBMF/yMlpxEpaZHxmKrrz+FUxqeRRBS69O12RM04Ho5sh8ElR3GybMCsShFBSbTagNoTPvsM6Br8SDLKAkRxIBVyjuZx0YWglDiJTrEdyTkxDgM3N158Z9bjao0iQlXfnNDN5+Et0fAI5FxIqXgDt3kXgjXz9CPxVCSPwB0WggQe9zqrghUIg/D+UBdCJzEYe5rXKcIWyW6OoUuCjGSKmDKrv1FpM0QLt5IF9zSkxLHOsWmKXg/1FnARYUCQgm9vmse3NnNzi1qLE3biHoQpuQHYkoRlkshjXoz/erP366ozIhY9GplSXCq1PxyY5oMP+lKw4iUdao2UYgM0FKxWl88ppEjGaqZYPaDa2Gw3LqWaPco4l4Q0o6p7IHIZfHslCU3KdrOhtRYbF/cUJclLxGz3EJkrpQBOJmwJv476VqX7OlJKmPk2bL+fqdPEuN0wliFM4rLE+OYoM3Sfk2LNbsnvvHk+ZSes2tzb1JvXh6F4SeCKFStWrFixYsXnOR5+o2EGKjSrYd6VGFpbDJrqpuXcS/lcYqQ97jMSfpoIwzhiRCeFuO7fUiZJmGLdukCKk2c3G7tcK6doDK+VzTiSBeZWo5ANSkoRTes6+lYbObl+v5QcvRAeIXp+ccbZ2Y5XXn45Tpd7I7mbBqYW3pPoTtDoDLFkTLNfV/cmdEkNxMZCIEsil8RgbjS35gM0KZNy4nicl3btaZo4P9uQysjh6GlSPW5XsidMZYQhSZAnN8n3AsSmvjHJIszzRFP31UzTjBoMY6ZhVHMBVbLqHoSQD7VOSCQhg5NHxFPAkq+myImQThn748wj5zvIBVVjrhP37l74dQ+F47FSh4qQKalyczgg0hAT6uLvEXJyCVgZ/KNYa4tMLTnFIZcIE6Chtd4y+EdUsaXoeXEWmrLQZsWoYcjukbZQhpFkvv0ppsxqtNqotIWENnMTuHVCrcbCM3Ci5suNSF7Lp8Hf4lelNkP3Ezoaw1BIloKUN0oawruj4YnJHpEcfhCXSPn39uds4fdxh9NKNFasWLFixYoVn/94HfG2GbMWm4oww4bZe2kACKO2D+vuN1hkTUTaUs6gUK0nVblZuMeKumbeh79QrLjeXSuQkNa8RVqEwzyxHbYICVJ0SSTB1CNFj8cpSEoO6VGmtsqmDORUmA6V68uXaeENSSK0Jl0N40MhhZi5GYYBw+UuQx5Prw/zhu4whC+xrGbYrGQX3kfqVY4iwSNmXp5nlijDyFyVgRnC0D0fZzc8a2MYhvh7Q5KRSkbiPhynI+O4YehyKyqleBP4PB0ZNyM3+yOlyNLwPWwK87GSegpViteZZJHszLWS8FZwl0f19uuZofg2KeuEpIyYcnWYGFOmtMkJRhlAfbNQykhryqZATgOzSRT8eSSsX7dSJC8dJi6Byy6rq7N3p6h/vpIYdfLUK2sNy4LkzCYVms1YcdIH/t6pej9K90+YGee7HdfN0AZNZ7Q1JGeP4G2eYIZBjs91D6ztnR0m3jCv1T0/zkYyRJHjVL0o0Gxks9n4Z0obrZpH78bKZCij+zY0XDQSZZbJFv+3pIIQm6XVpLFixYoVK1as+AKA2Dq1rFixYsWKFStWrFix4g3GWjG8YsWKFStWrFixYsWKNxwr0VixYsWKFStWrFixYsUbjpVorFixYsWKFStWrFix4g3HSjRWrFixYsWKFStWrFjxhmMlGitWrFixYsWKFStWrHjDsRKNFStWrFixYsWKFStWvOFYicaKFStWrFixYsWKFSvecKxEY8WKFStWrFixYsWKFW84VqKxYsWKFStWrFixYsWKNxz/HzaVQHSpoKFoAAAAAElFTkSuQmCC", "text/plain": [ "
" ] @@ -118,15 +272,43 @@ "\n", "from perceptionmetrics.utils import conversion as uc\n", "\n", - "image_fname = dataset.dataset[\"image\"].iloc[0]\n", + "image_fname = dataset.dataset[\"image\"].iloc[1]\n", + "if dataset.dataset_dir is not None:\n", + " image_fname = Path(dataset.dataset_dir) / image_fname\n", "image = Image.open(image_fname)\n", "\n", - "label_fname = dataset.dataset[\"label\"].iloc[0]\n", + "label_fname = dataset.dataset[\"label\"].iloc[1]\n", + "if dataset.dataset_dir is not None:\n", + " label_fname = Path(dataset.dataset_dir) / label_fname\n", "label = Image.open(label_fname)\n", "label = uc.label_to_rgb(label, dataset.ontology)\n", "\n", - "pred = model.predict(image)\n", - "pred = uc.label_to_rgb(pred, model.ontology)\n", + "raw_pred = pm_model.predict(image)\n", + "raw_pred_np = np.array(raw_pred)\n", + "raw_label_np = np.array(Image.open(label_fname))\n", + "\n", + "model_idx_to_name = {\n", + " class_meta[\"idx\"]: class_name for class_name, class_meta in pm_model.ontology.items()\n", + "}\n", + "dataset_idx_to_name = {\n", + " class_meta[\"idx\"]: class_name for class_name, class_meta in dataset.ontology.items()\n", + "}\n", + "\n", + "pred_ids = sorted(np.unique(raw_pred_np).tolist())\n", + "label_ids = sorted(np.unique(raw_label_np).tolist())\n", + "\n", + "print(\"Predicted class ids:\", pred_ids)\n", + "print(\n", + " \"Predicted class names:\",\n", + " [model_idx_to_name.get(class_id, f\"unknown:{class_id}\") for class_id in pred_ids],\n", + ")\n", + "print(\"Dataset label ids in sample:\", label_ids)\n", + "print(\n", + " \"Dataset label names in sample:\",\n", + " [dataset_idx_to_name.get(class_id, f\"unknown:{class_id}\") for class_id in label_ids],\n", + ")\n", + "\n", + "pred = uc.label_to_rgb(raw_pred, pm_model.ontology)\n", "pred = pred.resize(label.size)\n", "\n", "plt.figure(figsize=(10, 10))\n", @@ -136,45 +318,1474 @@ "plt.show()" ] }, + { + "cell_type": "markdown", + "id": "bcb40b82", + "metadata": {}, + "source": [ + "## Translate Dataset Labels and Run Evaluation\n", + "\n", + "The evaluation call maps ground-truth labels from the nuImages ontology into the model ontology with `translation_direction=\"dataset_to_model\"`.\n" + ] + }, { "cell_type": "code", - "execution_count": null, + "execution_count": 6, "id": "a1f6ca6b", "metadata": {}, - "outputs": [], + "outputs": [ + { + "name": "stderr", + "output_type": "stream", + "text": [ + "100%|██████████| 50/50 [00:01<00:00, 35.81it/s]\n", + "/home/tejass/Downloads/TUDELFT_ROBOTICS/GSOC/gsoc2026-Tejas_Stanley/PerceptionMetrics/perceptionmetrics/utils/segmentation_metrics.py:147: RuntimeWarning: invalid value encountered in divide\n", + " return np.where(denominator > 0, tp / denominator, np.nan)\n", + "/home/tejass/Downloads/TUDELFT_ROBOTICS/GSOC/gsoc2026-Tejas_Stanley/PerceptionMetrics/perceptionmetrics/utils/segmentation_metrics.py:164: RuntimeWarning: invalid value encountered in divide\n", + " return np.where(denominator > 0, tp / denominator, np.nan)\n", + "/home/tejass/Downloads/TUDELFT_ROBOTICS/GSOC/gsoc2026-Tejas_Stanley/PerceptionMetrics/perceptionmetrics/utils/segmentation_metrics.py:203: RuntimeWarning: invalid value encountered in divide\n", + " denominator > 0, 2 * (precision * recall) / denominator, np.nan\n", + "/home/tejass/Downloads/TUDELFT_ROBOTICS/GSOC/gsoc2026-Tejas_Stanley/PerceptionMetrics/perceptionmetrics/utils/segmentation_metrics.py:222: RuntimeWarning: invalid value encountered in divide\n", + " return np.where(union > 0, tp / union, np.nan)\n", + "/home/tejass/Downloads/TUDELFT_ROBOTICS/GSOC/gsoc2026-Tejas_Stanley/PerceptionMetrics/perceptionmetrics/utils/segmentation_metrics.py:240: RuntimeWarning: invalid value encountered in divide\n", + " return np.where(denominator > 0, 2 * tp / denominator, np.nan)\n" + ] + }, + { + "data": { + "text/html": [ + "
\n", + "\n", + "\n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + "
animalhuman.pedestrian.adulthuman.pedestrian.childhuman.pedestrian.construction_workerhuman.pedestrian.personal_mobilityhuman.pedestrian.police_officerhuman.pedestrian.strollerhuman.pedestrian.wheelchairmovable_object.barriermovable_object.debris...vehicle.carvehicle.constructionvehicle.emergency.ambulancevehicle.emergency.policevehicle.motorcyclevehicle.trailervehicle.truckflat.driveable_surfacemacromicro
tp0.04.578900e+040.02.280000e+020.000000e+000.00.00.07.052100e+040.000000e+00...3.026940e+051.243870e+050.00.02.299100e+046.436000e+037.031500e+042.559201e+06NaNNaN
fp0.06.403000e+030.02.690000e+020.000000e+000.00.00.03.562000e+031.000000e+00...1.643800e+043.742000e+030.00.03.728000e+035.490000e+022.202200e+048.568000e+03NaNNaN
fn0.05.353000e+030.01.494000e+031.254000e+030.00.00.03.064000e+031.146000e+03...1.231400e+043.680000e+030.00.01.579000e+031.489800e+049.345000e+039.131000e+03NaNNaN
tn3326172.03.268627e+063326172.03.324181e+063.324918e+063326172.03326172.03326172.03.249025e+063.325025e+06...2.994726e+063.194363e+063326172.03326172.03.297874e+063.304289e+063.224490e+067.492720e+05NaNNaN
precisionNaN8.773184e-01NaN4.587525e-01NaNNaNNaNNaN9.519188e-010.000000e+00...9.484915e-019.707951e-01NaNNaN8.604738e-019.214030e-017.615041e-019.966633e-010.7892790.978781
recallNaN8.953306e-01NaN1.324042e-010.000000e+00NaNNaNNaN9.583611e-010.000000e+00...9.609089e-019.712650e-01NaNNaN9.357346e-013.016781e-018.826889e-019.964448e-010.6760580.978781
accuracy1.09.964656e-011.09.994700e-019.996230e-011.01.01.09.980079e-019.996552e-01...9.913558e-019.977686e-011.01.09.984045e-019.953559e-019.905696e-019.946789e-010.9982320.998232
f1_scoreNaN8.862330e-01NaN2.054980e-01NaNNaNNaNNaN9.551291e-01NaN...9.546599e-019.710300e-01NaNNaN8.965275e-014.545358e-018.176305e-019.965540e-010.7891150.978781
iouNaN7.957077e-01NaN1.145153e-010.000000e+00NaNNaNNaN9.141120e-010.000000e+00...9.132528e-019.436913e-01NaNNaN8.124602e-012.941096e-016.915187e-019.931317e-010.6177490.958443
dice_scoreNaN8.862330e-01NaN2.054980e-010.000000e+00NaNNaNNaN9.551291e-010.000000e+00...9.546599e-019.710300e-01NaNNaN8.965275e-014.545358e-018.176305e-019.965540e-010.6904760.978781
animal0.00.000000e+000.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...0.000000e+000.000000e+000.00.00.000000e+000.000000e+000.000000e+000.000000e+00NaNNaN
human.pedestrian.adult0.04.578900e+040.09.000000e+014.470000e+020.00.00.05.200000e+019.140000e+02...1.235000e+032.140000e+020.00.06.370000e+020.000000e+006.580000e+021.288000e+03NaNNaN
human.pedestrian.child0.00.000000e+000.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...0.000000e+000.000000e+000.00.00.000000e+000.000000e+000.000000e+000.000000e+00NaNNaN
human.pedestrian.construction_worker0.02.390000e+020.02.280000e+020.000000e+000.00.00.02.000000e+000.000000e+00...0.000000e+002.000000e+010.00.00.000000e+000.000000e+008.000000e+000.000000e+00NaNNaN
human.pedestrian.personal_mobility0.00.000000e+000.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...0.000000e+000.000000e+000.00.00.000000e+000.000000e+000.000000e+000.000000e+00NaNNaN
human.pedestrian.police_officer0.00.000000e+000.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...0.000000e+000.000000e+000.00.00.000000e+000.000000e+000.000000e+000.000000e+00NaNNaN
human.pedestrian.stroller0.00.000000e+000.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...0.000000e+000.000000e+000.00.00.000000e+000.000000e+000.000000e+000.000000e+00NaNNaN
human.pedestrian.wheelchair0.00.000000e+000.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...0.000000e+000.000000e+000.00.00.000000e+000.000000e+000.000000e+000.000000e+00NaNNaN
movable_object.barrier0.03.130000e+020.01.500000e+010.000000e+000.00.00.07.052100e+046.000000e+00...3.000000e+007.490000e+020.00.00.000000e+000.000000e+003.480000e+021.260000e+03NaNNaN
movable_object.debris0.00.000000e+000.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...0.000000e+000.000000e+000.00.00.000000e+000.000000e+000.000000e+000.000000e+00NaNNaN
movable_object.pushable_pullable0.01.100000e+020.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...0.000000e+000.000000e+000.00.00.000000e+000.000000e+000.000000e+000.000000e+00NaNNaN
movable_object.trafficcone0.09.600000e+010.00.000000e+000.000000e+000.00.00.02.320000e+020.000000e+00...2.000000e+019.700000e+010.00.00.000000e+004.600000e+011.360000e+024.700000e+02NaNNaN
static_object.bicycle_rack0.00.000000e+000.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...0.000000e+000.000000e+000.00.04.000000e+010.000000e+000.000000e+000.000000e+00NaNNaN
vehicle.bicycle0.02.070000e+020.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...0.000000e+000.000000e+000.00.00.000000e+000.000000e+000.000000e+000.000000e+00NaNNaN
vehicle.bus.bendy0.00.000000e+000.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...0.000000e+000.000000e+000.00.00.000000e+000.000000e+000.000000e+000.000000e+00NaNNaN
vehicle.bus.rigid0.00.000000e+000.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...1.400000e+028.200000e+010.00.00.000000e+000.000000e+001.290000e+021.900000e+02NaNNaN
vehicle.car0.01.746000e+030.05.300000e+010.000000e+000.00.00.02.890000e+020.000000e+00...3.026940e+055.070000e+020.00.07.370000e+021.821000e+036.235000e+034.619000e+03NaNNaN
vehicle.construction0.03.890000e+020.06.820000e+020.000000e+000.00.00.08.730000e+020.000000e+00...8.930000e+021.243870e+050.00.00.000000e+000.000000e+008.140000e+029.100000e+01NaNNaN
vehicle.emergency.ambulance0.00.000000e+000.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...0.000000e+000.000000e+000.00.00.000000e+000.000000e+000.000000e+000.000000e+00NaNNaN
vehicle.emergency.police0.00.000000e+000.00.000000e+000.000000e+000.00.00.00.000000e+000.000000e+00...0.000000e+000.000000e+000.00.00.000000e+000.000000e+000.000000e+000.000000e+00NaNNaN
vehicle.motorcycle0.06.670000e+020.02.220000e+027.190000e+020.00.00.01.760000e+020.000000e+00...1.369000e+031.490000e+020.00.02.299100e+040.000000e+000.000000e+004.260000e+02NaNNaN
vehicle.trailer0.00.000000e+000.00.000000e+000.000000e+000.00.00.06.000000e+010.000000e+00...0.000000e+002.900000e+010.00.00.000000e+006.436000e+034.100000e+020.000000e+00NaNNaN
vehicle.truck0.05.660000e+020.03.350000e+026.600000e+010.00.00.04.200000e+010.000000e+00...5.173000e+031.542000e+030.00.04.000000e+001.282800e+047.031500e+047.870000e+02NaNNaN
flat.driveable_surface0.01.020000e+030.09.700000e+012.200000e+010.00.00.01.338000e+032.260000e+02...3.481000e+032.910000e+020.00.01.610000e+022.030000e+026.070000e+022.559201e+06NaNNaN
\n", + "

34 rows × 26 columns

\n", + "
" + ], + "text/plain": [ + " animal human.pedestrian.adult \\\n", + "tp 0.0 4.578900e+04 \n", + "fp 0.0 6.403000e+03 \n", + "fn 0.0 5.353000e+03 \n", + "tn 3326172.0 3.268627e+06 \n", + "precision NaN 8.773184e-01 \n", + "recall NaN 8.953306e-01 \n", + "accuracy 1.0 9.964656e-01 \n", + "f1_score NaN 8.862330e-01 \n", + "iou NaN 7.957077e-01 \n", + "dice_score NaN 8.862330e-01 \n", + "animal 0.0 0.000000e+00 \n", + "human.pedestrian.adult 0.0 4.578900e+04 \n", + "human.pedestrian.child 0.0 0.000000e+00 \n", + "human.pedestrian.construction_worker 0.0 2.390000e+02 \n", + "human.pedestrian.personal_mobility 0.0 0.000000e+00 \n", + "human.pedestrian.police_officer 0.0 0.000000e+00 \n", + "human.pedestrian.stroller 0.0 0.000000e+00 \n", + "human.pedestrian.wheelchair 0.0 0.000000e+00 \n", + "movable_object.barrier 0.0 3.130000e+02 \n", + "movable_object.debris 0.0 0.000000e+00 \n", + "movable_object.pushable_pullable 0.0 1.100000e+02 \n", + "movable_object.trafficcone 0.0 9.600000e+01 \n", + "static_object.bicycle_rack 0.0 0.000000e+00 \n", + "vehicle.bicycle 0.0 2.070000e+02 \n", + "vehicle.bus.bendy 0.0 0.000000e+00 \n", + "vehicle.bus.rigid 0.0 0.000000e+00 \n", + "vehicle.car 0.0 1.746000e+03 \n", + "vehicle.construction 0.0 3.890000e+02 \n", + "vehicle.emergency.ambulance 0.0 0.000000e+00 \n", + "vehicle.emergency.police 0.0 0.000000e+00 \n", + "vehicle.motorcycle 0.0 6.670000e+02 \n", + "vehicle.trailer 0.0 0.000000e+00 \n", + "vehicle.truck 0.0 5.660000e+02 \n", + "flat.driveable_surface 0.0 1.020000e+03 \n", + "\n", + " human.pedestrian.child \\\n", + "tp 0.0 \n", + "fp 0.0 \n", + "fn 0.0 \n", + "tn 3326172.0 \n", + "precision NaN \n", + "recall NaN \n", + "accuracy 1.0 \n", + "f1_score NaN \n", + "iou NaN \n", + "dice_score NaN \n", + "animal 0.0 \n", + "human.pedestrian.adult 0.0 \n", + "human.pedestrian.child 0.0 \n", + "human.pedestrian.construction_worker 0.0 \n", + "human.pedestrian.personal_mobility 0.0 \n", + "human.pedestrian.police_officer 0.0 \n", + "human.pedestrian.stroller 0.0 \n", + "human.pedestrian.wheelchair 0.0 \n", + "movable_object.barrier 0.0 \n", + "movable_object.debris 0.0 \n", + "movable_object.pushable_pullable 0.0 \n", + "movable_object.trafficcone 0.0 \n", + "static_object.bicycle_rack 0.0 \n", + "vehicle.bicycle 0.0 \n", + "vehicle.bus.bendy 0.0 \n", + "vehicle.bus.rigid 0.0 \n", + "vehicle.car 0.0 \n", + "vehicle.construction 0.0 \n", + "vehicle.emergency.ambulance 0.0 \n", + "vehicle.emergency.police 0.0 \n", + "vehicle.motorcycle 0.0 \n", + "vehicle.trailer 0.0 \n", + "vehicle.truck 0.0 \n", + "flat.driveable_surface 0.0 \n", + "\n", + " human.pedestrian.construction_worker \\\n", + "tp 2.280000e+02 \n", + "fp 2.690000e+02 \n", + "fn 1.494000e+03 \n", + "tn 3.324181e+06 \n", + "precision 4.587525e-01 \n", + "recall 1.324042e-01 \n", + "accuracy 9.994700e-01 \n", + "f1_score 2.054980e-01 \n", + "iou 1.145153e-01 \n", + "dice_score 2.054980e-01 \n", + "animal 0.000000e+00 \n", + "human.pedestrian.adult 9.000000e+01 \n", + "human.pedestrian.child 0.000000e+00 \n", + "human.pedestrian.construction_worker 2.280000e+02 \n", + "human.pedestrian.personal_mobility 0.000000e+00 \n", + "human.pedestrian.police_officer 0.000000e+00 \n", + "human.pedestrian.stroller 0.000000e+00 \n", + "human.pedestrian.wheelchair 0.000000e+00 \n", + "movable_object.barrier 1.500000e+01 \n", + "movable_object.debris 0.000000e+00 \n", + "movable_object.pushable_pullable 0.000000e+00 \n", + "movable_object.trafficcone 0.000000e+00 \n", + "static_object.bicycle_rack 0.000000e+00 \n", + "vehicle.bicycle 0.000000e+00 \n", + "vehicle.bus.bendy 0.000000e+00 \n", + "vehicle.bus.rigid 0.000000e+00 \n", + "vehicle.car 5.300000e+01 \n", + "vehicle.construction 6.820000e+02 \n", + "vehicle.emergency.ambulance 0.000000e+00 \n", + "vehicle.emergency.police 0.000000e+00 \n", + "vehicle.motorcycle 2.220000e+02 \n", + "vehicle.trailer 0.000000e+00 \n", + "vehicle.truck 3.350000e+02 \n", + "flat.driveable_surface 9.700000e+01 \n", + "\n", + " human.pedestrian.personal_mobility \\\n", + "tp 0.000000e+00 \n", + "fp 0.000000e+00 \n", + "fn 1.254000e+03 \n", + "tn 3.324918e+06 \n", + "precision NaN \n", + "recall 0.000000e+00 \n", + "accuracy 9.996230e-01 \n", + "f1_score NaN \n", + "iou 0.000000e+00 \n", + "dice_score 0.000000e+00 \n", + "animal 0.000000e+00 \n", + "human.pedestrian.adult 4.470000e+02 \n", + "human.pedestrian.child 0.000000e+00 \n", + "human.pedestrian.construction_worker 0.000000e+00 \n", + "human.pedestrian.personal_mobility 0.000000e+00 \n", + "human.pedestrian.police_officer 0.000000e+00 \n", + "human.pedestrian.stroller 0.000000e+00 \n", + "human.pedestrian.wheelchair 0.000000e+00 \n", + "movable_object.barrier 0.000000e+00 \n", + "movable_object.debris 0.000000e+00 \n", + "movable_object.pushable_pullable 0.000000e+00 \n", + "movable_object.trafficcone 0.000000e+00 \n", + "static_object.bicycle_rack 0.000000e+00 \n", + "vehicle.bicycle 0.000000e+00 \n", + "vehicle.bus.bendy 0.000000e+00 \n", + "vehicle.bus.rigid 0.000000e+00 \n", + "vehicle.car 0.000000e+00 \n", + "vehicle.construction 0.000000e+00 \n", + "vehicle.emergency.ambulance 0.000000e+00 \n", + "vehicle.emergency.police 0.000000e+00 \n", + "vehicle.motorcycle 7.190000e+02 \n", + "vehicle.trailer 0.000000e+00 \n", + "vehicle.truck 6.600000e+01 \n", + "flat.driveable_surface 2.200000e+01 \n", + "\n", + " human.pedestrian.police_officer \\\n", + "tp 0.0 \n", + "fp 0.0 \n", + "fn 0.0 \n", + "tn 3326172.0 \n", + "precision NaN \n", + "recall NaN \n", + "accuracy 1.0 \n", + "f1_score NaN \n", + "iou NaN \n", + "dice_score NaN \n", + "animal 0.0 \n", + "human.pedestrian.adult 0.0 \n", + "human.pedestrian.child 0.0 \n", + "human.pedestrian.construction_worker 0.0 \n", + "human.pedestrian.personal_mobility 0.0 \n", + "human.pedestrian.police_officer 0.0 \n", + "human.pedestrian.stroller 0.0 \n", + "human.pedestrian.wheelchair 0.0 \n", + "movable_object.barrier 0.0 \n", + "movable_object.debris 0.0 \n", + "movable_object.pushable_pullable 0.0 \n", + "movable_object.trafficcone 0.0 \n", + "static_object.bicycle_rack 0.0 \n", + "vehicle.bicycle 0.0 \n", + "vehicle.bus.bendy 0.0 \n", + "vehicle.bus.rigid 0.0 \n", + "vehicle.car 0.0 \n", + "vehicle.construction 0.0 \n", + "vehicle.emergency.ambulance 0.0 \n", + "vehicle.emergency.police 0.0 \n", + "vehicle.motorcycle 0.0 \n", + "vehicle.trailer 0.0 \n", + "vehicle.truck 0.0 \n", + "flat.driveable_surface 0.0 \n", + "\n", + " human.pedestrian.stroller \\\n", + "tp 0.0 \n", + "fp 0.0 \n", + "fn 0.0 \n", + "tn 3326172.0 \n", + "precision NaN \n", + "recall NaN \n", + "accuracy 1.0 \n", + "f1_score NaN \n", + "iou NaN \n", + "dice_score NaN \n", + "animal 0.0 \n", + "human.pedestrian.adult 0.0 \n", + "human.pedestrian.child 0.0 \n", + "human.pedestrian.construction_worker 0.0 \n", + "human.pedestrian.personal_mobility 0.0 \n", + "human.pedestrian.police_officer 0.0 \n", + "human.pedestrian.stroller 0.0 \n", + "human.pedestrian.wheelchair 0.0 \n", + "movable_object.barrier 0.0 \n", + "movable_object.debris 0.0 \n", + "movable_object.pushable_pullable 0.0 \n", + "movable_object.trafficcone 0.0 \n", + "static_object.bicycle_rack 0.0 \n", + "vehicle.bicycle 0.0 \n", + "vehicle.bus.bendy 0.0 \n", + "vehicle.bus.rigid 0.0 \n", + "vehicle.car 0.0 \n", + "vehicle.construction 0.0 \n", + "vehicle.emergency.ambulance 0.0 \n", + "vehicle.emergency.police 0.0 \n", + "vehicle.motorcycle 0.0 \n", + "vehicle.trailer 0.0 \n", + "vehicle.truck 0.0 \n", + "flat.driveable_surface 0.0 \n", + "\n", + " human.pedestrian.wheelchair \\\n", + "tp 0.0 \n", + "fp 0.0 \n", + "fn 0.0 \n", + "tn 3326172.0 \n", + "precision NaN \n", + "recall NaN \n", + "accuracy 1.0 \n", + "f1_score NaN \n", + "iou NaN \n", + "dice_score NaN \n", + "animal 0.0 \n", + "human.pedestrian.adult 0.0 \n", + "human.pedestrian.child 0.0 \n", + "human.pedestrian.construction_worker 0.0 \n", + "human.pedestrian.personal_mobility 0.0 \n", + "human.pedestrian.police_officer 0.0 \n", + "human.pedestrian.stroller 0.0 \n", + "human.pedestrian.wheelchair 0.0 \n", + "movable_object.barrier 0.0 \n", + "movable_object.debris 0.0 \n", + "movable_object.pushable_pullable 0.0 \n", + "movable_object.trafficcone 0.0 \n", + "static_object.bicycle_rack 0.0 \n", + "vehicle.bicycle 0.0 \n", + "vehicle.bus.bendy 0.0 \n", + "vehicle.bus.rigid 0.0 \n", + "vehicle.car 0.0 \n", + "vehicle.construction 0.0 \n", + "vehicle.emergency.ambulance 0.0 \n", + "vehicle.emergency.police 0.0 \n", + "vehicle.motorcycle 0.0 \n", + "vehicle.trailer 0.0 \n", + "vehicle.truck 0.0 \n", + "flat.driveable_surface 0.0 \n", + "\n", + " movable_object.barrier \\\n", + "tp 7.052100e+04 \n", + "fp 3.562000e+03 \n", + "fn 3.064000e+03 \n", + "tn 3.249025e+06 \n", + "precision 9.519188e-01 \n", + "recall 9.583611e-01 \n", + "accuracy 9.980079e-01 \n", + "f1_score 9.551291e-01 \n", + "iou 9.141120e-01 \n", + "dice_score 9.551291e-01 \n", + "animal 0.000000e+00 \n", + "human.pedestrian.adult 5.200000e+01 \n", + "human.pedestrian.child 0.000000e+00 \n", + "human.pedestrian.construction_worker 2.000000e+00 \n", + "human.pedestrian.personal_mobility 0.000000e+00 \n", + "human.pedestrian.police_officer 0.000000e+00 \n", + "human.pedestrian.stroller 0.000000e+00 \n", + "human.pedestrian.wheelchair 0.000000e+00 \n", + "movable_object.barrier 7.052100e+04 \n", + "movable_object.debris 0.000000e+00 \n", + "movable_object.pushable_pullable 0.000000e+00 \n", + "movable_object.trafficcone 2.320000e+02 \n", + "static_object.bicycle_rack 0.000000e+00 \n", + "vehicle.bicycle 0.000000e+00 \n", + "vehicle.bus.bendy 0.000000e+00 \n", + "vehicle.bus.rigid 0.000000e+00 \n", + "vehicle.car 2.890000e+02 \n", + "vehicle.construction 8.730000e+02 \n", + "vehicle.emergency.ambulance 0.000000e+00 \n", + "vehicle.emergency.police 0.000000e+00 \n", + "vehicle.motorcycle 1.760000e+02 \n", + "vehicle.trailer 6.000000e+01 \n", + "vehicle.truck 4.200000e+01 \n", + "flat.driveable_surface 1.338000e+03 \n", + "\n", + " movable_object.debris ... \\\n", + "tp 0.000000e+00 ... \n", + "fp 1.000000e+00 ... \n", + "fn 1.146000e+03 ... \n", + "tn 3.325025e+06 ... \n", + "precision 0.000000e+00 ... \n", + "recall 0.000000e+00 ... \n", + "accuracy 9.996552e-01 ... \n", + "f1_score NaN ... \n", + "iou 0.000000e+00 ... \n", + "dice_score 0.000000e+00 ... \n", + "animal 0.000000e+00 ... \n", + "human.pedestrian.adult 9.140000e+02 ... \n", + "human.pedestrian.child 0.000000e+00 ... \n", + "human.pedestrian.construction_worker 0.000000e+00 ... \n", + "human.pedestrian.personal_mobility 0.000000e+00 ... \n", + "human.pedestrian.police_officer 0.000000e+00 ... \n", + "human.pedestrian.stroller 0.000000e+00 ... \n", + "human.pedestrian.wheelchair 0.000000e+00 ... \n", + "movable_object.barrier 6.000000e+00 ... \n", + "movable_object.debris 0.000000e+00 ... \n", + "movable_object.pushable_pullable 0.000000e+00 ... \n", + "movable_object.trafficcone 0.000000e+00 ... \n", + "static_object.bicycle_rack 0.000000e+00 ... \n", + "vehicle.bicycle 0.000000e+00 ... \n", + "vehicle.bus.bendy 0.000000e+00 ... \n", + "vehicle.bus.rigid 0.000000e+00 ... \n", + "vehicle.car 0.000000e+00 ... \n", + "vehicle.construction 0.000000e+00 ... \n", + "vehicle.emergency.ambulance 0.000000e+00 ... \n", + "vehicle.emergency.police 0.000000e+00 ... \n", + "vehicle.motorcycle 0.000000e+00 ... \n", + "vehicle.trailer 0.000000e+00 ... \n", + "vehicle.truck 0.000000e+00 ... \n", + "flat.driveable_surface 2.260000e+02 ... \n", + "\n", + " vehicle.car vehicle.construction \\\n", + "tp 3.026940e+05 1.243870e+05 \n", + "fp 1.643800e+04 3.742000e+03 \n", + "fn 1.231400e+04 3.680000e+03 \n", + "tn 2.994726e+06 3.194363e+06 \n", + "precision 9.484915e-01 9.707951e-01 \n", + "recall 9.609089e-01 9.712650e-01 \n", + "accuracy 9.913558e-01 9.977686e-01 \n", + "f1_score 9.546599e-01 9.710300e-01 \n", + "iou 9.132528e-01 9.436913e-01 \n", + "dice_score 9.546599e-01 9.710300e-01 \n", + "animal 0.000000e+00 0.000000e+00 \n", + "human.pedestrian.adult 1.235000e+03 2.140000e+02 \n", + "human.pedestrian.child 0.000000e+00 0.000000e+00 \n", + "human.pedestrian.construction_worker 0.000000e+00 2.000000e+01 \n", + "human.pedestrian.personal_mobility 0.000000e+00 0.000000e+00 \n", + "human.pedestrian.police_officer 0.000000e+00 0.000000e+00 \n", + "human.pedestrian.stroller 0.000000e+00 0.000000e+00 \n", + "human.pedestrian.wheelchair 0.000000e+00 0.000000e+00 \n", + "movable_object.barrier 3.000000e+00 7.490000e+02 \n", + "movable_object.debris 0.000000e+00 0.000000e+00 \n", + "movable_object.pushable_pullable 0.000000e+00 0.000000e+00 \n", + "movable_object.trafficcone 2.000000e+01 9.700000e+01 \n", + "static_object.bicycle_rack 0.000000e+00 0.000000e+00 \n", + "vehicle.bicycle 0.000000e+00 0.000000e+00 \n", + "vehicle.bus.bendy 0.000000e+00 0.000000e+00 \n", + "vehicle.bus.rigid 1.400000e+02 8.200000e+01 \n", + "vehicle.car 3.026940e+05 5.070000e+02 \n", + "vehicle.construction 8.930000e+02 1.243870e+05 \n", + "vehicle.emergency.ambulance 0.000000e+00 0.000000e+00 \n", + "vehicle.emergency.police 0.000000e+00 0.000000e+00 \n", + "vehicle.motorcycle 1.369000e+03 1.490000e+02 \n", + "vehicle.trailer 0.000000e+00 2.900000e+01 \n", + "vehicle.truck 5.173000e+03 1.542000e+03 \n", + "flat.driveable_surface 3.481000e+03 2.910000e+02 \n", + "\n", + " vehicle.emergency.ambulance \\\n", + "tp 0.0 \n", + "fp 0.0 \n", + "fn 0.0 \n", + "tn 3326172.0 \n", + "precision NaN \n", + "recall NaN \n", + "accuracy 1.0 \n", + "f1_score NaN \n", + "iou NaN \n", + "dice_score NaN \n", + "animal 0.0 \n", + "human.pedestrian.adult 0.0 \n", + "human.pedestrian.child 0.0 \n", + "human.pedestrian.construction_worker 0.0 \n", + "human.pedestrian.personal_mobility 0.0 \n", + "human.pedestrian.police_officer 0.0 \n", + "human.pedestrian.stroller 0.0 \n", + "human.pedestrian.wheelchair 0.0 \n", + "movable_object.barrier 0.0 \n", + "movable_object.debris 0.0 \n", + "movable_object.pushable_pullable 0.0 \n", + "movable_object.trafficcone 0.0 \n", + "static_object.bicycle_rack 0.0 \n", + "vehicle.bicycle 0.0 \n", + "vehicle.bus.bendy 0.0 \n", + "vehicle.bus.rigid 0.0 \n", + "vehicle.car 0.0 \n", + "vehicle.construction 0.0 \n", + "vehicle.emergency.ambulance 0.0 \n", + "vehicle.emergency.police 0.0 \n", + "vehicle.motorcycle 0.0 \n", + "vehicle.trailer 0.0 \n", + "vehicle.truck 0.0 \n", + "flat.driveable_surface 0.0 \n", + "\n", + " vehicle.emergency.police \\\n", + "tp 0.0 \n", + "fp 0.0 \n", + "fn 0.0 \n", + "tn 3326172.0 \n", + "precision NaN \n", + "recall NaN \n", + "accuracy 1.0 \n", + "f1_score NaN \n", + "iou NaN \n", + "dice_score NaN \n", + "animal 0.0 \n", + "human.pedestrian.adult 0.0 \n", + "human.pedestrian.child 0.0 \n", + "human.pedestrian.construction_worker 0.0 \n", + "human.pedestrian.personal_mobility 0.0 \n", + "human.pedestrian.police_officer 0.0 \n", + "human.pedestrian.stroller 0.0 \n", + "human.pedestrian.wheelchair 0.0 \n", + "movable_object.barrier 0.0 \n", + "movable_object.debris 0.0 \n", + "movable_object.pushable_pullable 0.0 \n", + "movable_object.trafficcone 0.0 \n", + "static_object.bicycle_rack 0.0 \n", + "vehicle.bicycle 0.0 \n", + "vehicle.bus.bendy 0.0 \n", + "vehicle.bus.rigid 0.0 \n", + "vehicle.car 0.0 \n", + "vehicle.construction 0.0 \n", + "vehicle.emergency.ambulance 0.0 \n", + "vehicle.emergency.police 0.0 \n", + "vehicle.motorcycle 0.0 \n", + "vehicle.trailer 0.0 \n", + "vehicle.truck 0.0 \n", + "flat.driveable_surface 0.0 \n", + "\n", + " vehicle.motorcycle vehicle.trailer \\\n", + "tp 2.299100e+04 6.436000e+03 \n", + "fp 3.728000e+03 5.490000e+02 \n", + "fn 1.579000e+03 1.489800e+04 \n", + "tn 3.297874e+06 3.304289e+06 \n", + "precision 8.604738e-01 9.214030e-01 \n", + "recall 9.357346e-01 3.016781e-01 \n", + "accuracy 9.984045e-01 9.953559e-01 \n", + "f1_score 8.965275e-01 4.545358e-01 \n", + "iou 8.124602e-01 2.941096e-01 \n", + "dice_score 8.965275e-01 4.545358e-01 \n", + "animal 0.000000e+00 0.000000e+00 \n", + "human.pedestrian.adult 6.370000e+02 0.000000e+00 \n", + "human.pedestrian.child 0.000000e+00 0.000000e+00 \n", + "human.pedestrian.construction_worker 0.000000e+00 0.000000e+00 \n", + "human.pedestrian.personal_mobility 0.000000e+00 0.000000e+00 \n", + "human.pedestrian.police_officer 0.000000e+00 0.000000e+00 \n", + "human.pedestrian.stroller 0.000000e+00 0.000000e+00 \n", + "human.pedestrian.wheelchair 0.000000e+00 0.000000e+00 \n", + "movable_object.barrier 0.000000e+00 0.000000e+00 \n", + "movable_object.debris 0.000000e+00 0.000000e+00 \n", + "movable_object.pushable_pullable 0.000000e+00 0.000000e+00 \n", + "movable_object.trafficcone 0.000000e+00 4.600000e+01 \n", + "static_object.bicycle_rack 4.000000e+01 0.000000e+00 \n", + "vehicle.bicycle 0.000000e+00 0.000000e+00 \n", + "vehicle.bus.bendy 0.000000e+00 0.000000e+00 \n", + "vehicle.bus.rigid 0.000000e+00 0.000000e+00 \n", + "vehicle.car 7.370000e+02 1.821000e+03 \n", + "vehicle.construction 0.000000e+00 0.000000e+00 \n", + "vehicle.emergency.ambulance 0.000000e+00 0.000000e+00 \n", + "vehicle.emergency.police 0.000000e+00 0.000000e+00 \n", + "vehicle.motorcycle 2.299100e+04 0.000000e+00 \n", + "vehicle.trailer 0.000000e+00 6.436000e+03 \n", + "vehicle.truck 4.000000e+00 1.282800e+04 \n", + "flat.driveable_surface 1.610000e+02 2.030000e+02 \n", + "\n", + " vehicle.truck flat.driveable_surface \\\n", + "tp 7.031500e+04 2.559201e+06 \n", + "fp 2.202200e+04 8.568000e+03 \n", + "fn 9.345000e+03 9.131000e+03 \n", + "tn 3.224490e+06 7.492720e+05 \n", + "precision 7.615041e-01 9.966633e-01 \n", + "recall 8.826889e-01 9.964448e-01 \n", + "accuracy 9.905696e-01 9.946789e-01 \n", + "f1_score 8.176305e-01 9.965540e-01 \n", + "iou 6.915187e-01 9.931317e-01 \n", + "dice_score 8.176305e-01 9.965540e-01 \n", + "animal 0.000000e+00 0.000000e+00 \n", + "human.pedestrian.adult 6.580000e+02 1.288000e+03 \n", + "human.pedestrian.child 0.000000e+00 0.000000e+00 \n", + "human.pedestrian.construction_worker 8.000000e+00 0.000000e+00 \n", + "human.pedestrian.personal_mobility 0.000000e+00 0.000000e+00 \n", + "human.pedestrian.police_officer 0.000000e+00 0.000000e+00 \n", + "human.pedestrian.stroller 0.000000e+00 0.000000e+00 \n", + "human.pedestrian.wheelchair 0.000000e+00 0.000000e+00 \n", + "movable_object.barrier 3.480000e+02 1.260000e+03 \n", + "movable_object.debris 0.000000e+00 0.000000e+00 \n", + "movable_object.pushable_pullable 0.000000e+00 0.000000e+00 \n", + "movable_object.trafficcone 1.360000e+02 4.700000e+02 \n", + "static_object.bicycle_rack 0.000000e+00 0.000000e+00 \n", + "vehicle.bicycle 0.000000e+00 0.000000e+00 \n", + "vehicle.bus.bendy 0.000000e+00 0.000000e+00 \n", + "vehicle.bus.rigid 1.290000e+02 1.900000e+02 \n", + "vehicle.car 6.235000e+03 4.619000e+03 \n", + "vehicle.construction 8.140000e+02 9.100000e+01 \n", + "vehicle.emergency.ambulance 0.000000e+00 0.000000e+00 \n", + "vehicle.emergency.police 0.000000e+00 0.000000e+00 \n", + "vehicle.motorcycle 0.000000e+00 4.260000e+02 \n", + "vehicle.trailer 4.100000e+02 0.000000e+00 \n", + "vehicle.truck 7.031500e+04 7.870000e+02 \n", + "flat.driveable_surface 6.070000e+02 2.559201e+06 \n", + "\n", + " macro micro \n", + "tp NaN NaN \n", + "fp NaN NaN \n", + "fn NaN NaN \n", + "tn NaN NaN \n", + "precision 0.789279 0.978781 \n", + "recall 0.676058 0.978781 \n", + "accuracy 0.998232 0.998232 \n", + "f1_score 0.789115 0.978781 \n", + "iou 0.617749 0.958443 \n", + "dice_score 0.690476 0.978781 \n", + "animal NaN NaN \n", + "human.pedestrian.adult NaN NaN \n", + "human.pedestrian.child NaN NaN \n", + "human.pedestrian.construction_worker NaN NaN \n", + "human.pedestrian.personal_mobility NaN NaN \n", + "human.pedestrian.police_officer NaN NaN \n", + "human.pedestrian.stroller NaN NaN \n", + "human.pedestrian.wheelchair NaN NaN \n", + "movable_object.barrier NaN NaN \n", + "movable_object.debris NaN NaN \n", + "movable_object.pushable_pullable NaN NaN \n", + "movable_object.trafficcone NaN NaN \n", + "static_object.bicycle_rack NaN NaN \n", + "vehicle.bicycle NaN NaN \n", + "vehicle.bus.bendy NaN NaN \n", + "vehicle.bus.rigid NaN NaN \n", + "vehicle.car NaN NaN \n", + "vehicle.construction NaN NaN \n", + "vehicle.emergency.ambulance NaN NaN \n", + "vehicle.emergency.police NaN NaN \n", + "vehicle.motorcycle NaN NaN \n", + "vehicle.trailer NaN NaN \n", + "vehicle.truck NaN NaN \n", + "flat.driveable_surface NaN NaN \n", + "\n", + "[34 rows x 26 columns]" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], "source": [ - "# Inspect overlap first\n", - "model_classes = set(model.ontology.keys())\n", - "dataset_classes = set(dataset.ontology.keys())\n", - "print(\"Model classes:\", len(model_classes))\n", - "print(\"Dataset classes:\", len(dataset_classes))\n", - "print(\n", - " \"Overlap:\",\n", - " len(model_classes & dataset_classes),\n", - " sorted(model_classes & dataset_classes)[:20],\n", - ")\n", - "\n", - "# Build a safe translation map\n", - "# old_ontology = dataset.ontology, new_ontology = model.ontology\n", - "# dataset ontology -> model ontology\n", - "fallback = list(model.ontology.keys())[0]\n", - "ontology_translation = {\n", - " class_name: (class_name if class_name in model.ontology else fallback)\n", + "# Translate original nuImages dataset labels into the model ontology.\n", + "# Classes ignored by the model config are masked before this translation is applied.\n", + "fallback_class = next(iter(pm_model.ontology.keys()))\n", + "dataset_to_model_translation = {\n", + " class_name: (class_name if class_name in pm_model.ontology else fallback_class)\n", " for class_name in dataset.ontology.keys()\n", "}\n", "\n", - "# TorchImageSegmentationModel.eval expects a JSON filename, not a Python dict.\n", "translation_path = Path(\"local/data/nuimages_to_model_ontology_translation.json\")\n", "translation_path.parent.mkdir(parents=True, exist_ok=True)\n", "with open(translation_path, \"w\", encoding=\"utf-8\") as f:\n", - " json.dump(ontology_translation, f, indent=2)\n", + " json.dump(dataset_to_model_translation, f, indent=2)\n", "\n", - "results = model.eval(\n", + "results = pm_model.eval(\n", " dataset,\n", - " split=\"train\",\n", + " split=\"val\",\n", " ontology_translation=str(translation_path),\n", + " translation_direction=\"dataset_to_model\",\n", ")\n", - "display(results)" + "display(results)\n" ] } ], From d9cc24d26d1a56d4de61fd5ae595373c2669617a Mon Sep 17 00:00:00 2001 From: tejass Date: Mon, 6 Jul 2026 16:19:01 +0200 Subject: [PATCH 09/10] Add GUI support for image segmentation --- tabs/dataset_viewer.py | 8 +- tabs/evaluator.py | 4 +- tabs/inference.py | 2 +- tabs/tasks/image_detection/dataset_viewer.py | 1 + tabs/tasks/image_detection/evaluator.py | 5 +- tabs/tasks/image_detection/inference.py | 7 +- tabs/tasks/image_detection/sidebar.py | 62 ++++++- .../image_segmentation/dataset_viewer.py | 172 ++++++++++++++---- tabs/tasks/image_segmentation/evaluator.py | 118 +++++++----- tabs/tasks/image_segmentation/inference.py | 7 + tabs/tasks/image_segmentation/sidebar.py | 70 ++++++- 11 files changed, 354 insertions(+), 102 deletions(-) diff --git a/tabs/dataset_viewer.py b/tabs/dataset_viewer.py index f5338e12..8ded02a9 100644 --- a/tabs/dataset_viewer.py +++ b/tabs/dataset_viewer.py @@ -1,7 +1,11 @@ import streamlit as st from tabs.tasks.image_detection.dataset_viewer import render_image_detection_viewer -from tabs.tasks.image_segmentation.dataset_viewer import render_image_segmentation_viewer -from tabs.tasks.lidar_segmentation.dataset_viewer import render_lidar_segmentation_viewer +from tabs.tasks.image_segmentation.dataset_viewer import ( + render_image_segmentation_viewer, +) +from tabs.tasks.lidar_segmentation.dataset_viewer import ( + render_lidar_segmentation_viewer, +) def dataset_viewer_tab(): diff --git a/tabs/evaluator.py b/tabs/evaluator.py index e3565d7f..6d49e31f 100644 --- a/tabs/evaluator.py +++ b/tabs/evaluator.py @@ -3,18 +3,18 @@ from tabs.tasks.image_segmentation.evaluator import render_image_segmentation_evaluator from tabs.tasks.lidar_segmentation.evaluator import render_lidar_segmentation_evaluator + def evaluator_tab(): task = st.session_state.get("task", "Image Detection") if task == "Image Detection": render_image_detection_evaluator() return - if task == "Image Segmentation": render_image_segmentation_evaluator() return - + if task == "Lidar Segmentation": render_lidar_segmentation_evaluator() return diff --git a/tabs/inference.py b/tabs/inference.py index 8e406f16..8213aa30 100644 --- a/tabs/inference.py +++ b/tabs/inference.py @@ -3,6 +3,7 @@ from tabs.tasks.image_segmentation.inference import render_image_segmentation_inference from tabs.tasks.lidar_segmentation.inference import render_lidar_segmentation_inference + def inference_tab(): task = st.session_state.get("task", "Image Detection") @@ -10,7 +11,6 @@ def inference_tab(): render_image_detection_inference() return - if task == "Image Segmentation": render_image_segmentation_inference() return diff --git a/tabs/tasks/image_detection/dataset_viewer.py b/tabs/tasks/image_detection/dataset_viewer.py index 52b04bb6..2eb3f72a 100644 --- a/tabs/tasks/image_detection/dataset_viewer.py +++ b/tabs/tasks/image_detection/dataset_viewer.py @@ -6,6 +6,7 @@ def render_image_detection_viewer(): + """Render the image detection dataset viewer tab in Streamlit.""" import tempfile from perceptionmetrics.datasets.coco import CocoDataset from perceptionmetrics.datasets.yolo import YOLODataset diff --git a/tabs/tasks/image_detection/evaluator.py b/tabs/tasks/image_detection/evaluator.py index e4e4fdf0..a0e788d0 100644 --- a/tabs/tasks/image_detection/evaluator.py +++ b/tabs/tasks/image_detection/evaluator.py @@ -16,6 +16,7 @@ def browse_predictions_outdir(): def render_image_detection_evaluator(): + """Render the image detection evaluator tab in Streamlit.""" st.header("Evaluator") st.markdown("Evaluate your model on the loaded dataset using PerceptionMetrics.") @@ -109,7 +110,9 @@ def render_image_detection_evaluator(): "
", unsafe_allow_html=True ) st.button( - "Browse", on_click=browse_predictions_outdir, key="browse_preds_outdir" + "Browse", + on_click=browse_predictions_outdir, + key="browse_preds_outdir", ) predictions_outdir_input = st.session_state.get("predictions_outdir") diff --git a/tabs/tasks/image_detection/inference.py b/tabs/tasks/image_detection/inference.py index 4be364c4..a884d23c 100644 --- a/tabs/tasks/image_detection/inference.py +++ b/tabs/tasks/image_detection/inference.py @@ -3,12 +3,11 @@ import streamlit as st import json from PIL import Image + try: import torch except ImportError: - raise ImportError( - "PyTorch is required for GUI-based inference and evaluation. " - ) + raise ImportError("PyTorch is required for GUI-based inference and evaluation. ") def draw_detections(image: Image, predictions: dict, label_map: Optional[dict] = None): @@ -49,6 +48,8 @@ def draw_detections(image: Image, predictions: dict, label_map: Optional[dict] = def render_image_detection_inference(): + """Render the image detection inference tab in Streamlit.""" + st.header("Model Inference") st.markdown("Select an image and run inference using the loaded model.") diff --git a/tabs/tasks/image_detection/sidebar.py b/tabs/tasks/image_detection/sidebar.py index d20e5124..a68845a6 100644 --- a/tabs/tasks/image_detection/sidebar.py +++ b/tabs/tasks/image_detection/sidebar.py @@ -9,14 +9,23 @@ def browse_dataset_path(): - st.session_state.dataset_path = browse_folder() + folder = browse_folder() + if folder: + st.session_state.dataset_path = folder def render_image_detection_sidebar(available_devices): + """Render the sidebar for the image detection task in Streamlit. + :param available_devices: List of available devices for model inference + :type available_devices: list + """ + with st.expander("Image Detection Dataset", expanded=True): col1, col2 = st.columns(2) with col1: - st.selectbox("Type", ["COCO", "YOLO"], key="dataset_type") # split into two columns + st.selectbox( + "Type", ["COCO", "YOLO"], key="dataset_type" + ) # split into two columns with col2: st.selectbox("Split", ["train", "val", "test"], key="split") @@ -25,10 +34,14 @@ def render_image_detection_sidebar(available_devices): st.text_input("Dataset Folder", key="dataset_path") with col2: st.markdown( - "
", # add some spacing to align with the text input + "
", # add some spacing to align with the text input unsafe_allow_html=True, ) - st.button("Browse", on_click=browse_dataset_path) # add a button to browse for the dataset folder + st.button( + "Browse", + on_click=browse_dataset_path, + key="browse_detection_dataset_path", + ) if st.session_state.get("dataset_type", "COCO") == "YOLO": st.file_uploader( @@ -57,7 +70,7 @@ def render_image_detection_sidebar(available_devices): ["Manual Configuration", "Upload Config File"], key="config_option", horizontal=True, - ) # radio button to select between manual configuration and uploading a config file + ) # radio button to select between manual configuration and uploading a config file if ( st.session_state.get("config_option", "Manual Configuration") == "Upload Config File" @@ -80,7 +93,12 @@ def render_image_detection_sidebar(available_devices): ): load_image_detection_model() + def _render_manual_detection_model_config(available_devices): + """Render the manual configuration options for the image detection model in the sidebar. + :param available_devices: List of available devices for model inference + :type available_devices: list + """ col1, col2 = st.columns(2) with col1: st.slider( @@ -113,8 +131,7 @@ def _render_manual_detection_model_config(available_devices): ["torchvision", "YOLO"], index=( 0 - if st.session_state.get("model_format", "torchvision") - == "torchvision" + if st.session_state.get("model_format", "torchvision") == "torchvision" else 1 ), key="model_format", @@ -207,7 +224,9 @@ def _render_manual_detection_model_config(available_devices): help="Center crop width", ) + def load_image_detection_model(): + """Load the image detection model based on the provided configuration and ontology files.""" from perceptionmetrics.models.torch_detection import TorchImageDetectionModel model_file = st.session_state.get("model_file") @@ -241,7 +260,12 @@ def load_image_detection_model(): st.session_state.detection_model_loaded = False st.error(f"Failed to load model: {e}") + def _write_detection_config() -> Optional[str]: + """Write the detection configuration to a temporary JSON file based on the selected configuration method. + :return: Path to the temporary JSON configuration file or None if an error occurred + :rtype: Optional[str] + """ config_option = st.session_state.get("config_option", "Manual Configuration") config_file = ( st.session_state.get("config_file") @@ -267,7 +291,12 @@ def _write_detection_config() -> Optional[str]: st.error(f"Failed to prepare configuration: {e}") return None + def _manual_detection_config() -> dict: + """Generate a configuration dictionary based on the manual configuration options in the sidebar. + :return: Configuration dictionary + :rtype: dict + """ resize_cfg = None if st.session_state.get("enable_resize", True): resize_strategy = st.session_state.get("resize_strategy", "Fixed Dimensions") @@ -284,9 +313,7 @@ def _manual_detection_config() -> dict: st.session_state.get("confidence_threshold", 0.5) ), "nms_threshold": float(st.session_state.get("nms_threshold", 0.5)), - "max_detections_per_image": int( - st.session_state.get("max_detections", 100) - ), + "max_detections_per_image": int(st.session_state.get("max_detections", 100)), "device": st.session_state.get("device", "cpu"), "batch_size": int(st.session_state.get("batch_size", 1)), "evaluation_step": int(st.session_state.get("evaluation_step", 5)), @@ -303,7 +330,15 @@ def _manual_detection_config() -> dict: return config_data + def _uploaded_json_to_tempfile(uploaded_file) -> Optional[str]: + """Save the uploaded JSON file to a temporary file and return its path. + :param uploaded_file: Uploaded JSON file + :type uploaded_file: UploadedFile + :return: Path to the temporary JSON file or None if an error occurred + :rtype: Optional[str] + """ + try: data = json.load(uploaded_file) with tempfile.NamedTemporaryFile( @@ -315,7 +350,14 @@ def _uploaded_json_to_tempfile(uploaded_file) -> Optional[str]: st.error(f"Failed to load JSON file: {e}") return None + def _uploaded_model_to_tempfile(uploaded_file) -> Optional[str]: + """Save the uploaded model file to a temporary file and return its path. + :param uploaded_file: Uploaded model file + :type uploaded_file: UploadedFile + :return: Path to the temporary model file or None if an error occurred + :rtype: Optional[str] + """ try: suffix = os.path.splitext(uploaded_file.name)[1] or ".pt" with tempfile.NamedTemporaryFile( diff --git a/tabs/tasks/image_segmentation/dataset_viewer.py b/tabs/tasks/image_segmentation/dataset_viewer.py index 23f0e93f..721fd9c0 100644 --- a/tabs/tasks/image_segmentation/dataset_viewer.py +++ b/tabs/tasks/image_segmentation/dataset_viewer.py @@ -6,10 +6,41 @@ from PIL import Image from perceptionmetrics.datasets.cityscapes import CityscapesImageSegmentationDataset +from perceptionmetrics.datasets.nuimages import NuImagesSegmentationDataset from tabs.tasks.utils import render_image_grid +def _overlay_mask(image, label, ontology, opacity): + """Overlay a segmentation mask on an image. + param image: PIL Image object of the original image. + param label: 2D numpy array of the segmentation mask. + param ontology: Dictionary mapping class names to their properties, including 'idx' and 'rgb'. + param opacity: Float value between 0 and 1 indicating the opacity of the overlay. + return: PIL Image object of the image with the overlay applied. + + """ + image_np = np.array(image) + color_mask = np.zeros((*label.shape, 3), dtype=np.uint8) + for class_data in ontology.values(): + class_idx = int(class_data["idx"]) + rgb = class_data.get("rgb") + if rgb is None: + rng = np.random.default_rng(abs(class_idx)) + rgb = tuple(int(value) for value in rng.integers(0, 255, size=3)) + color_mask[label == class_idx] = rgb + + resampling = getattr(Image, "Resampling", Image) + color_mask_image = Image.fromarray(color_mask).resize( + image.size, resampling.NEAREST + ) + color_mask_np = np.array(color_mask_image) + + overlay = ((1.0 - opacity) * image_np + opacity * color_mask_np).astype(np.uint8) + return Image.fromarray(overlay) + + def render_image_segmentation_viewer(): + """Render the image segmentation dataset viewer tab in Streamlit.""" dataset_type = st.session_state.get("segmentation_dataset_type", "Cityscapes") st.header("Dataset Viewer") @@ -18,10 +49,16 @@ def render_image_segmentation_viewer(): _render_cityscapes_viewer() return + if dataset_type == "NuImages": + _render_nuimages_viewer() + return + st.info(f"{dataset_type} image segmentation viewer is not wired yet.") def _render_cityscapes_viewer(): + """Render the Cityscapes dataset viewer in Streamlit.""" + dataset_path = st.session_state.get("dataset_path", "") split = st.session_state.get("split", "val") @@ -69,7 +106,15 @@ def _render_cityscapes_viewer(): try: image = Image.open(image_fname).convert("RGB") - label = _read_cityscapes_label(dataset, label_fname) + if label_fname.endswith("_gtFine_color.png"): + label_rgb = np.array(Image.open(label_fname).convert("RGB")) + label = np.zeros(label_rgb.shape[:2], dtype=np.uint8) + for class_data in dataset.ontology.values(): + class_idx = int(class_data["idx"]) + rgb = np.array(class_data["rgb"], dtype=np.uint8) + label[(label_rgb == rgb).all(axis=2)] = class_idx + else: + label = dataset.read_label(label_fname) except Exception as exc: st.error(f"Failed to read sample '{sample_name}': {exc}") return @@ -87,6 +132,11 @@ def _render_cityscapes_viewer(): def _load_cityscapes_dataset(dataset_path, split): + """Load the Cityscapes dataset based on the provided path and split. + param dataset_path: Path to the Cityscapes dataset directory. + param split: Dataset split to load (e.g., "train", "val", "test"). + return: Instance of CityscapesImageSegmentationDataset as a session state variable. + """ roots = {"train": None, "val": None, "test": None} roots[split] = dataset_path @@ -118,56 +168,112 @@ def _load_cityscapes_dataset(dataset_path, split): return st.session_state[dataset_key] -def _overlay_mask(image, label, ontology, opacity): - image_np = np.array(image) - color_mask = _colorize_mask(label, ontology) +def _render_nuimages_viewer(): + """Render the NuImages dataset viewer in Streamlit.""" + dataset_path = st.session_state.get("dataset_path", "") + split = st.session_state.get("split", "val") - resampling = getattr(Image, "Resampling", Image) - color_mask_image = Image.fromarray(color_mask).resize( - image.size, resampling.NEAREST - ) - color_mask_np = np.array(color_mask_image) + if split == "test": + st.warning("NuImages segmentation viewer supports train and val splits.") + return - overlay = ((1.0 - opacity) * image_np + opacity * color_mask_np).astype(np.uint8) - return Image.fromarray(overlay) + if not dataset_path or not os.path.isdir(dataset_path): + st.warning("Please select a valid NuImages dataset folder.") + return + try: + dataset = _load_nuimages_dataset(dataset_path, split) + except Exception as exc: + st.error(f"Failed to load NuImages dataset: {exc}") + return -def _read_cityscapes_label(dataset, label_fname): - if label_fname.endswith("_gtFine_color.png"): - return _read_color_label(label_fname, dataset.ontology) + split_df = dataset.dataset[dataset.dataset["split"] == split] + if split_df.empty: + st.warning(f"No NuImages samples found for split '{split}'.") + return - return dataset.read_label(label_fname) + sample_names = split_df.index.astype(str).tolist() + image_paths = split_df["image"].tolist() + selected_img_path, sample_name = render_image_grid( + item_names=sample_names, + image_paths=image_paths, + state_prefix="nuimages_segmentation", + context=f"{dataset_path}_{split}_{st.session_state.get('nuimages_segmentation_version', 'v1.0-mini')}", + search_label="sample", + ) + if not selected_img_path: + st.info("Select an image to view the ground truth mask.") + return -def _read_color_label(label_fname, ontology): - label_rgb = np.array(Image.open(label_fname).convert("RGB")) - label = np.zeros(label_rgb.shape[:2], dtype=np.uint8) + mask_opacity = st.slider( + "Mask Opacity", + min_value=0.0, + max_value=1.0, + value=0.45, + step=0.05, + key="nuimages_segmentation_mask_opacity", + ) - for class_data in ontology.values(): - class_idx = int(class_data["idx"]) - rgb = np.array(class_data["rgb"], dtype=np.uint8) - label[(label_rgb == rgb).all(axis=2)] = class_idx + row = split_df.loc[int(sample_name) if sample_name.isdigit() else sample_name] + image_fname = row["image"] + label_fname = row["label"] - return label + try: + image = Image.open(image_fname).convert("RGB") + label = dataset.read_label(label_fname) + except Exception as exc: + st.error(f"Failed to read sample '{sample_name}': {exc}") + return + overlay = _overlay_mask(image, label, dataset.ontology, mask_opacity) -def _colorize_mask(label, ontology): - color_mask = np.zeros((*label.shape, 3), dtype=np.uint8) + image_col, overlay_col = st.columns(2) + with image_col: + st.image(image, caption="Image", use_container_width=True) + with overlay_col: + st.image(overlay, caption="Ground Truth Overlay", use_container_width=True) - for class_data in ontology.values(): - class_idx = int(class_data["idx"]) - rgb = class_data.get("rgb", _fallback_color(class_idx)) - color_mask[label == class_idx] = rgb + with st.expander("Classes", expanded=False): + st.dataframe(_classes_dataframe(dataset.ontology), use_container_width=True) - return color_mask +def _load_nuimages_dataset(dataset_path, split): + """Load the NuImages dataset based on the provided path and split. + param dataset_path: Path to the NuImages dataset directory. + param split: Dataset split to load (e.g., "train", "val"). + return: Instance of NuImagesSegmentationDataset as a session state variable. + """ -def _fallback_color(class_idx): - rng = np.random.default_rng(abs(int(class_idx))) - return tuple(int(value) for value in rng.integers(0, 255, size=3)) + version = st.session_state.get("nuimages_segmentation_version", "v1.0-mini") + labels_rel_dir = st.session_state.get( + "nuimages_segmentation_labels_dir", + "generated/nuimages_segmentation_labels", + ) + dataset_key = ( + "nuimages_segmentation_dataset", + os.path.abspath(dataset_path), + version, + split, + labels_rel_dir, + ) + + if dataset_key not in st.session_state: + st.session_state[dataset_key] = NuImagesSegmentationDataset( + dataset_dir=dataset_path, + version=version, + split=split, + labels_rel_dir=labels_rel_dir, + ) + + return st.session_state[dataset_key] def _classes_dataframe(ontology): + """Convert the ontology dictionary to a pandas DataFrame for display. + param ontology: Dictionary mapping class names to their properties, including 'idx', 'train_id', 'category', and 'rgb'. + return: pandas DataFrame with columns ['class', 'id', 'train_id', 'category', 'rgb']. + """ rows = [] for class_name, class_data in ontology.items(): rows.append( diff --git a/tabs/tasks/image_segmentation/evaluator.py b/tabs/tasks/image_segmentation/evaluator.py index 6221dcab..15e38f60 100644 --- a/tabs/tasks/image_segmentation/evaluator.py +++ b/tabs/tasks/image_segmentation/evaluator.py @@ -5,6 +5,7 @@ import streamlit as st from perceptionmetrics.datasets.cityscapes import CityscapesImageSegmentationDataset +from perceptionmetrics.datasets.nuimages import NuImagesSegmentationDataset from tabs.tasks.utils import browse_folder @@ -15,11 +16,12 @@ def browse_segmentation_predictions_outdir(): def render_image_segmentation_evaluator(): + """Render the image segmentation evaluator tab in the Streamlit app.""" st.header("Evaluator") st.markdown("Evaluate your model on the loaded dataset using PerceptionMetrics.") dataset_type = st.session_state.get("segmentation_dataset_type", "Cityscapes") - if dataset_type != "Cityscapes": + if dataset_type not in ["Cityscapes", "NuImages"]: st.info(f"{dataset_type} image segmentation evaluation is not wired yet.") return @@ -32,39 +34,65 @@ def render_image_segmentation_evaluator(): st.warning( "No dataset path provided. Please set the dataset path in the sidebar." ) + elif dataset_type == "NuImages" and split == "test": + st.warning("NuImages segmentation evaluation supports train and val splits.") else: try: - roots = {"train": None, "val": None, "test": None} - roots[split] = dataset_path - dataset_key = ( - "cityscapes_segmentation_evaluation_dataset", - os.path.abspath(dataset_path), - split, - st.session_state.get( - "segmentation_image_dir", - "leftImg8bit_trainvaltest/leftImg8bit", - ), - st.session_state.get("segmentation_label_dir", "gtFine"), - st.session_state.get( - "segmentation_image_suffix", "_leftImg8bit.png" - ), - st.session_state.get( - "segmentation_label_suffix", "_gtFine_labelIds.png" - ), - st.session_state.get("segmentation_use_train_id", False), - ) + if dataset_type == "Cityscapes": + roots = {"train": None, "val": None, "test": None} + roots[split] = dataset_path + dataset_key = ( + "cityscapes_segmentation_evaluation_dataset", + os.path.abspath(dataset_path), + split, + st.session_state.get( + "segmentation_image_dir", + "leftImg8bit_trainvaltest/leftImg8bit", + ), + st.session_state.get("segmentation_label_dir", "gtFine"), + st.session_state.get( + "segmentation_image_suffix", "_leftImg8bit.png" + ), + st.session_state.get( + "segmentation_label_suffix", "_gtFine_labelIds.png" + ), + st.session_state.get("segmentation_use_train_id", False), + ) - if dataset_key not in st.session_state: - st.session_state[dataset_key] = CityscapesImageSegmentationDataset( - train_dataset_root=roots["train"], - val_dataset_root=roots["val"], - test_dataset_root=roots["test"], - image_dir=dataset_key[3], - label_dir=dataset_key[4], - image_suffix=dataset_key[5], - label_suffix=dataset_key[6], - use_train_id=dataset_key[7], + if dataset_key not in st.session_state: + st.session_state[dataset_key] = CityscapesImageSegmentationDataset( + train_dataset_root=roots["train"], + val_dataset_root=roots["val"], + test_dataset_root=roots["test"], + image_dir=dataset_key[3], + label_dir=dataset_key[4], + image_suffix=dataset_key[5], + label_suffix=dataset_key[6], + use_train_id=dataset_key[7], + ) + else: + version = st.session_state.get( + "nuimages_segmentation_version", "v1.0-mini" + ) + labels_rel_dir = st.session_state.get( + "nuimages_segmentation_labels_dir", + "generated/nuimages_segmentation_labels", ) + dataset_key = ( + "nuimages_segmentation_evaluation_dataset", + os.path.abspath(dataset_path), + version, + split, + labels_rel_dir, + ) + + if dataset_key not in st.session_state: + st.session_state[dataset_key] = NuImagesSegmentationDataset( + dataset_dir=dataset_path, + version=version, + split=split, + labels_rel_dir=labels_rel_dir, + ) dataset = st.session_state[dataset_key] st.success( @@ -116,10 +144,16 @@ def render_image_segmentation_evaluator(): key="segmentation_ontology_translation", help="JSON file for translating between dataset and model ontologies.", ) - st.info( - "For Cityscapes SegFormer models, use train-ID labels or provide a " - "label-ID to train-ID ontology translation." - ) + if dataset_type == "Cityscapes": + st.info( + "For Cityscapes SegFormer models, use train-ID labels or provide a " + "label-ID to train-ID ontology translation." + ) + elif dataset_type == "NuImages": + st.info( + "For NuImages SegFormer models, upload " + "nuimages_to_model_ontology_translation.json and use dataset_to_model." + ) translation_direction = st.selectbox( "Translation Direction", @@ -153,9 +187,11 @@ def render_image_segmentation_evaluator(): json.dump(json.load(ontology_translation), tmp_trans) ontology_translation_path = tmp_trans.name - predictions_outdir = predictions_outdir.strip() if ( - save_predictions and predictions_outdir - ) else None + predictions_outdir = ( + predictions_outdir.strip() + if (save_predictions and predictions_outdir) + else None + ) if predictions_outdir is not None: os.makedirs(predictions_outdir, exist_ok=True) @@ -172,9 +208,7 @@ def progress_callback(processed, total): def metrics_callback(metrics_df, processed, total): with intermediate_metrics_placeholder.container(): - st.markdown( - f"#### Results (after {processed}/{total} images)" - ) + st.markdown(f"#### Results (after {processed}/{total} images)") display_segmentation_evaluation_results( metrics_df, show_download=False ) @@ -206,11 +240,13 @@ def metrics_callback(metrics_df, processed, total): def display_segmentation_evaluation_results(results, show_download=True): + """Display the evaluation results in a Streamlit dataframe and provide a download button. + Param results: pd.DataFrame, the evaluation results to display + Param show_download: bool, whether to show the download button for the results""" if results is None or results.empty: st.warning("No evaluation results to display.") return - st.markdown("#### Metrics") display_df = results.copy() numeric_columns = display_df.select_dtypes(include=["float64", "int64"]).columns diff --git a/tabs/tasks/image_segmentation/inference.py b/tabs/tasks/image_segmentation/inference.py index ebdad937..8fbbb5cc 100644 --- a/tabs/tasks/image_segmentation/inference.py +++ b/tabs/tasks/image_segmentation/inference.py @@ -6,6 +6,7 @@ def render_image_segmentation_inference(): + """Render the image segmentation inference tab in the Streamlit app.""" st.header("Model Inference") st.markdown("Select an image and run inference using the loaded model.") @@ -50,6 +51,12 @@ def render_image_segmentation_inference(): def _overlay_mask(image, mask_rgb, opacity): + """Overlay a segmentation mask on an image with a given opacity. + Param image: PIL.Image, the original image + Param mask_rgb: PIL.Image, the segmentation mask in RGB format + Param opacity: float, the opacity of the mask overlay (0.0 to 1.0) + Return: PIL.Image, the image with the mask overlay + """ image_np = np.array(image) mask_np = np.array(mask_rgb) overlay = ((1.0 - opacity) * image_np + opacity * mask_np).astype(np.uint8) diff --git a/tabs/tasks/image_segmentation/sidebar.py b/tabs/tasks/image_segmentation/sidebar.py index 141bea56..98ee7cd8 100644 --- a/tabs/tasks/image_segmentation/sidebar.py +++ b/tabs/tasks/image_segmentation/sidebar.py @@ -14,27 +14,41 @@ "GOOSE", ] +MODEL_INPUT_DATASETS = ["Cityscapes", "NuImages"] + def browse_dataset_path(): - st.session_state.dataset_path = browse_folder() + folder = browse_folder() + if folder: + st.session_state.dataset_path = folder def browse_segmentation_model_path(): if st.session_state.get("segmentation_model_type") == "Hugging Face SegFormer": - st.session_state.segmentation_model_path = browse_folder() + path = browse_folder() else: - st.session_state.segmentation_model_path = browse_file() + path = browse_file() + + if path: + st.session_state.segmentation_model_path = path def browse_segmentation_config_path(): - st.session_state.segmentation_config_path = browse_file() + path = browse_file() + if path: + st.session_state.segmentation_config_path = path def browse_segmentation_ontology_path(): - st.session_state.segmentation_ontology_path = browse_file() + path = browse_file() + if path: + st.session_state.segmentation_ontology_path = path def render_image_segmentation_sidebar(_available_devices): + """Render the sidebar for the image segmentation task in Streamlit. + param _available_devices: List of available devices for model inference. + """ with st.expander("Image Segmentation Dataset", expanded=True): col1, col2 = st.columns(2) with col1: @@ -54,16 +68,25 @@ def render_image_segmentation_sidebar(_available_devices): "
", unsafe_allow_html=True, ) - st.button("Browse", on_click=browse_dataset_path) + st.button( + "Browse", + on_click=browse_dataset_path, + key="browse_segmentation_dataset_path", + ) if dataset_type == "Cityscapes": render_cityscapes_dataset_inputs() + elif dataset_type == "NuImages": + render_nuimages_dataset_inputs() else: st.info(f"{dataset_type} image segmentation inputs are not wired yet.") with st.expander("Image Segmentation Model", expanded=False): - if st.session_state.get("segmentation_dataset_type", "Cityscapes") != "Cityscapes": - st.info("Image segmentation model loading is wired for Cityscapes first.") + dataset_type = st.session_state.get("segmentation_dataset_type", "Cityscapes") + if dataset_type not in MODEL_INPUT_DATASETS: + st.info( + f"Image segmentation model loading is not wired for {dataset_type} yet." + ) return render_segmentation_model_inputs() @@ -78,6 +101,7 @@ def render_image_segmentation_sidebar(_available_devices): def render_cityscapes_dataset_inputs(): + """Render the input fields for the Cityscapes dataset in the sidebar.""" st.text_input( "Image Directory", value="leftImg8bit_trainvaltest/leftImg8bit", @@ -106,7 +130,26 @@ def render_cityscapes_dataset_inputs(): ) +def render_nuimages_dataset_inputs(): + """Render the input fields for the NuImages dataset in the sidebar.""" + st.text_input( + "Version", + value="v1.0-mini", + key="nuimages_segmentation_version", + help="nuImages version, for example v1.0-mini, v1.0-train, or v1.0-val.", + ) + st.text_input( + "Generated Labels Directory", + value="generated/nuimages_segmentation_labels", + key="nuimages_segmentation_labels_dir", + help="Relative directory where generated segmentation masks are stored.", + ) + if st.session_state.get("split") == "test": + st.warning("NuImages segmentation supports train/val-style splits, not test.") + + def render_segmentation_model_inputs(): + """Render the input fields for the image segmentation model in the sidebar.""" model_type = st.selectbox( "Model Type", ["Torch Model File", "Hugging Face SegFormer"], @@ -116,7 +159,11 @@ def render_segmentation_model_inputs(): col1, col2 = st.columns([3, 1]) with col1: st.text_input( - "Model Path" if model_type == "Torch Model File" else "Model Name or Folder", + ( + "Model Path" + if model_type == "Torch Model File" + else "Model Name or Folder" + ), key="segmentation_model_path", help=( "Path to a TorchScript model or saved PyTorch model file." @@ -173,6 +220,7 @@ def render_segmentation_model_inputs(): def load_image_segmentation_model(): + """Render the image segmentation model in the sidebar.""" from perceptionmetrics.models.torch_segmentation import TorchImageSegmentationModel model_type = st.session_state.get("segmentation_model_type", "Torch Model File") @@ -208,6 +256,10 @@ def load_image_segmentation_model(): def load_model_for_type(model_type, model_path): + """Load a model based on the specified type and path. + param model_type: Type of the model to load (e.g., "Torch Model File", "Hugging Face SegFormer"). + param model_path: Path to the model file or directory. + """ if model_type == "Torch Model File": if not os.path.isfile(model_path): raise ValueError( From 603570fe49ce4fe8e159ae67c83bad612d791eef Mon Sep 17 00:00:00 2001 From: tejass Date: Wed, 15 Jul 2026 11:20:55 +0200 Subject: [PATCH 10/10] Refactor image segmentation dataset viewer --- .../image_segmentation/dataset_viewer.py | 153 ++++++------------ tabs/tasks/image_segmentation/evaluator.py | 63 +------- tabs/tasks/image_segmentation/sidebar.py | 5 +- 3 files changed, 52 insertions(+), 169 deletions(-) diff --git a/tabs/tasks/image_segmentation/dataset_viewer.py b/tabs/tasks/image_segmentation/dataset_viewer.py index 721fd9c0..911b11ea 100644 --- a/tabs/tasks/image_segmentation/dataset_viewer.py +++ b/tabs/tasks/image_segmentation/dataset_viewer.py @@ -10,6 +10,8 @@ from tabs.tasks.utils import render_image_grid + + def _overlay_mask(image, label, ontology, opacity): """Overlay a segmentation mask on an image. param image: PIL Image object of the original image. @@ -42,48 +44,51 @@ def _overlay_mask(image, label, ontology, opacity): def render_image_segmentation_viewer(): """Render the image segmentation dataset viewer tab in Streamlit.""" dataset_type = st.session_state.get("segmentation_dataset_type", "Cityscapes") - - st.header("Dataset Viewer") - - if dataset_type == "Cityscapes": - _render_cityscapes_viewer() - return - - if dataset_type == "NuImages": - _render_nuimages_viewer() - return - - st.info(f"{dataset_type} image segmentation viewer is not wired yet.") - - -def _render_cityscapes_viewer(): - """Render the Cityscapes dataset viewer in Streamlit.""" - dataset_path = st.session_state.get("dataset_path", "") split = st.session_state.get("split", "val") + st.header("Dataset Viewer") + if not dataset_path or not os.path.isdir(dataset_path): - st.warning("Please select a valid Cityscapes dataset folder.") + st.warning("Please select a valid image segmentation dataset folder.") return try: - dataset = _load_cityscapes_dataset(dataset_path, split) + dataset = load_image_segmentation_dataset(dataset_type, dataset_path, split) except Exception as exc: - st.error(f"Failed to load Cityscapes dataset: {exc}") + st.error(f"Failed to load image segmentation dataset: {exc}") return + render_segmentation_dataset_viewer( + dataset=dataset, + dataset_type=dataset_type, + split=split, + state_prefix="image_segmentation", + context=f"{dataset_path}_{split}", + ) + + + +def render_segmentation_dataset_viewer( + dataset, + dataset_type, + split, + state_prefix, + context, +): + """Render a loaded image segmentation dataset.""" split_df = dataset.dataset[dataset.dataset["split"] == split] if split_df.empty: - st.warning(f"No Cityscapes samples found for split '{split}'.") + st.warning(f"No {dataset_type} samples found for split '{split}'.") return - sample_names = split_df.index.tolist() + sample_names = split_df.index.astype(str).tolist() image_paths = split_df["image"].tolist() selected_img_path, sample_name = render_image_grid( item_names=sample_names, image_paths=image_paths, - state_prefix="cityscapes_segmentation", - context=f"{dataset_path}_{split}", + state_prefix=state_prefix, + context=context, search_label="sample", ) @@ -97,24 +102,20 @@ def _render_cityscapes_viewer(): max_value=1.0, value=0.45, step=0.05, - key="cityscapes_segmentation_mask_opacity", + key=f"{state_prefix}_mask_opacity", ) - row = split_df.loc[sample_name] + row_key = sample_name + if row_key not in split_df.index and sample_name.isdigit(): + row_key = int(sample_name) + + row = split_df.loc[row_key] image_fname = row["image"] label_fname = row["label"] try: image = Image.open(image_fname).convert("RGB") - if label_fname.endswith("_gtFine_color.png"): - label_rgb = np.array(Image.open(label_fname).convert("RGB")) - label = np.zeros(label_rgb.shape[:2], dtype=np.uint8) - for class_data in dataset.ontology.values(): - class_idx = int(class_data["idx"]) - rgb = np.array(class_data["rgb"], dtype=np.uint8) - label[(label_rgb == rgb).all(axis=2)] = class_idx - else: - label = dataset.read_label(label_fname) + label = dataset.read_label(label_fname) except Exception as exc: st.error(f"Failed to read sample '{sample_name}': {exc}") return @@ -131,7 +132,15 @@ def _render_cityscapes_viewer(): st.dataframe(_classes_dataframe(dataset.ontology), use_container_width=True) -def _load_cityscapes_dataset(dataset_path, split): +def load_image_segmentation_dataset(dataset_type, dataset_path, split): + if dataset_type == "Cityscapes": + return load_cityscapes_dataset(dataset_path, split) + if dataset_type == "NuImages": + return load_nuimages_dataset(dataset_path, split) + raise ValueError(f"{dataset_type} image segmentation dataset is not wired yet.") + + +def load_cityscapes_dataset(dataset_path, split): """Load the Cityscapes dataset based on the provided path and split. param dataset_path: Path to the Cityscapes dataset directory. param split: Dataset split to load (e.g., "train", "val", "test"). @@ -168,77 +177,7 @@ def _load_cityscapes_dataset(dataset_path, split): return st.session_state[dataset_key] -def _render_nuimages_viewer(): - """Render the NuImages dataset viewer in Streamlit.""" - dataset_path = st.session_state.get("dataset_path", "") - split = st.session_state.get("split", "val") - - if split == "test": - st.warning("NuImages segmentation viewer supports train and val splits.") - return - - if not dataset_path or not os.path.isdir(dataset_path): - st.warning("Please select a valid NuImages dataset folder.") - return - - try: - dataset = _load_nuimages_dataset(dataset_path, split) - except Exception as exc: - st.error(f"Failed to load NuImages dataset: {exc}") - return - - split_df = dataset.dataset[dataset.dataset["split"] == split] - if split_df.empty: - st.warning(f"No NuImages samples found for split '{split}'.") - return - - sample_names = split_df.index.astype(str).tolist() - image_paths = split_df["image"].tolist() - selected_img_path, sample_name = render_image_grid( - item_names=sample_names, - image_paths=image_paths, - state_prefix="nuimages_segmentation", - context=f"{dataset_path}_{split}_{st.session_state.get('nuimages_segmentation_version', 'v1.0-mini')}", - search_label="sample", - ) - - if not selected_img_path: - st.info("Select an image to view the ground truth mask.") - return - - mask_opacity = st.slider( - "Mask Opacity", - min_value=0.0, - max_value=1.0, - value=0.45, - step=0.05, - key="nuimages_segmentation_mask_opacity", - ) - - row = split_df.loc[int(sample_name) if sample_name.isdigit() else sample_name] - image_fname = row["image"] - label_fname = row["label"] - - try: - image = Image.open(image_fname).convert("RGB") - label = dataset.read_label(label_fname) - except Exception as exc: - st.error(f"Failed to read sample '{sample_name}': {exc}") - return - - overlay = _overlay_mask(image, label, dataset.ontology, mask_opacity) - - image_col, overlay_col = st.columns(2) - with image_col: - st.image(image, caption="Image", use_container_width=True) - with overlay_col: - st.image(overlay, caption="Ground Truth Overlay", use_container_width=True) - - with st.expander("Classes", expanded=False): - st.dataframe(_classes_dataframe(dataset.ontology), use_container_width=True) - - -def _load_nuimages_dataset(dataset_path, split): +def load_nuimages_dataset(dataset_path, split): """Load the NuImages dataset based on the provided path and split. param dataset_path: Path to the NuImages dataset directory. param split: Dataset split to load (e.g., "train", "val"). diff --git a/tabs/tasks/image_segmentation/evaluator.py b/tabs/tasks/image_segmentation/evaluator.py index 15e38f60..fd05f78f 100644 --- a/tabs/tasks/image_segmentation/evaluator.py +++ b/tabs/tasks/image_segmentation/evaluator.py @@ -4,8 +4,9 @@ import streamlit as st -from perceptionmetrics.datasets.cityscapes import CityscapesImageSegmentationDataset -from perceptionmetrics.datasets.nuimages import NuImagesSegmentationDataset +from tabs.tasks.image_segmentation.dataset_viewer import ( + load_image_segmentation_dataset, +) from tabs.tasks.utils import browse_folder @@ -38,63 +39,7 @@ def render_image_segmentation_evaluator(): st.warning("NuImages segmentation evaluation supports train and val splits.") else: try: - if dataset_type == "Cityscapes": - roots = {"train": None, "val": None, "test": None} - roots[split] = dataset_path - dataset_key = ( - "cityscapes_segmentation_evaluation_dataset", - os.path.abspath(dataset_path), - split, - st.session_state.get( - "segmentation_image_dir", - "leftImg8bit_trainvaltest/leftImg8bit", - ), - st.session_state.get("segmentation_label_dir", "gtFine"), - st.session_state.get( - "segmentation_image_suffix", "_leftImg8bit.png" - ), - st.session_state.get( - "segmentation_label_suffix", "_gtFine_labelIds.png" - ), - st.session_state.get("segmentation_use_train_id", False), - ) - - if dataset_key not in st.session_state: - st.session_state[dataset_key] = CityscapesImageSegmentationDataset( - train_dataset_root=roots["train"], - val_dataset_root=roots["val"], - test_dataset_root=roots["test"], - image_dir=dataset_key[3], - label_dir=dataset_key[4], - image_suffix=dataset_key[5], - label_suffix=dataset_key[6], - use_train_id=dataset_key[7], - ) - else: - version = st.session_state.get( - "nuimages_segmentation_version", "v1.0-mini" - ) - labels_rel_dir = st.session_state.get( - "nuimages_segmentation_labels_dir", - "generated/nuimages_segmentation_labels", - ) - dataset_key = ( - "nuimages_segmentation_evaluation_dataset", - os.path.abspath(dataset_path), - version, - split, - labels_rel_dir, - ) - - if dataset_key not in st.session_state: - st.session_state[dataset_key] = NuImagesSegmentationDataset( - dataset_dir=dataset_path, - version=version, - split=split, - labels_rel_dir=labels_rel_dir, - ) - - dataset = st.session_state[dataset_key] + dataset = load_image_segmentation_dataset(dataset_type, dataset_path, split) st.success( f"✅ Dataset loaded: {dataset_path} ({split} split) - {len(dataset.dataset)} samples" ) diff --git a/tabs/tasks/image_segmentation/sidebar.py b/tabs/tasks/image_segmentation/sidebar.py index 98ee7cd8..6679abf3 100644 --- a/tabs/tasks/image_segmentation/sidebar.py +++ b/tabs/tasks/image_segmentation/sidebar.py @@ -8,6 +8,8 @@ IMAGE_SEGMENTATION_DATASETS = [ "Cityscapes", "NuImages", + "GAIA", + "Generic", "Wildscenes", "RUGD", "Rellis3D", @@ -144,9 +146,6 @@ def render_nuimages_dataset_inputs(): key="nuimages_segmentation_labels_dir", help="Relative directory where generated segmentation masks are stored.", ) - if st.session_state.get("split") == "test": - st.warning("NuImages segmentation supports train/val-style splits, not test.") - def render_segmentation_model_inputs(): """Render the input fields for the image segmentation model in the sidebar."""