first review of onnx LP

Author: pareenaverma

content/learning-paths/mobile-graphics-and-gaming/onnx/02_setup.md

@@ -19,28 +19,55 @@ You can choose a variety of hardware, including:
 The nice thing about ONNX is that the **same model file** can run across all of these, so your setup is flexible.
 
 ## Install Python
-Depending on the hardware you use you follow different installation paths
 
-1. Linux (Arm64). In the console type:
+{{% notice Note %}}
+ONNX Runtime provides prebuilt wheels only for specific Python versions. At the time of writing, Python 3.12 is not yet supported by ONNX Runtime on macOS or Arm platforms. If you see an error like:
+
+```output
+ERROR: No matching distribution found for onnxruntime
+```
+
+it usually means your Python version is too new. Python 3.10 is tested and recommended for this Learning Path.
+{{% /notice %}}
+
+Depending on the hardware you use, follow different installation paths:
+
+1. Linux (Arm64). Install Python 3.10 by typing in the console:
 ```console
 sudo apt update
-sudo apt install -y python3 python3-venv python3-pip build-essential libopenblas-dev libgl1 libglib2.0-0
+sudo apt install -y python3.10 python3.10-venv python3.10-dev python3-pip build-essential libopenblas-dev libgl1 libglib2.0-0
 ```
 
-2. macOS (Apple Sillicon):
+If Python 3.10 is not available in your default repositories, you can use the deadsnakes PPA:
 ```console
-brew install python
+sudo add-apt-repository ppa:deadsnakes/ppa
+sudo apt update
+sudo apt install -y python3.10 python3.10-venv python3.10-dev
 ```
 
+2. macOS (Apple Silicon). Install Python 3.10 using Homebrew:
+```console
+brew install python@3.10
+```
+
+After installation, use `python3.10` explicitly when creating virtual environments.
+
 3. Windows on Arm:
-* Install Python 3.10+ from python.org (Arm64 build).
+* Download and install Python 3.10 from [python.org](https://www.python.org/downloads/) (select the Arm64 build).
 * Ensure pip is on PATH.
 
-After installing Python, open a terminal or console, create a clean virtual environment, and update pip and wheel:
+After installing Python 3.10, open a terminal or console, create a clean virtual environment using Python 3.10 explicitly, and update pip and wheel:
+
+```console
+python3.10 -m venv .venv
+source .venv/bin/activate
+python -m pip install --upgrade pip wheel
+```
 
+On macOS, if `python3.10` is not found, use the full Homebrew path:
 ```console
-python3 -m venv .venv
-source .venv/bin/activate                   # on Windows use: .venv\Scripts\activate
+/opt/homebrew/bin/python3.10 -m venv .venv
+source .venv/bin/activate
 python -m pip install --upgrade pip wheel
 ```
 

content/learning-paths/mobile-graphics-and-gaming/onnx/03_preparingdata.md

@@ -241,9 +241,18 @@ data/
 pip install pandas pyarrow opencv-python tqdm numpy
 ```
 
-2. Place the Parquet file (e.g., train_1.parquet) next to the script or update PARQUET_PATH accordingly. Here we used the file from [this location](https://huggingface.co/datasets/Ritvik19/Sudoku-Dataset/blob/main/train_1.parquet).
+2. Download the Sudoku dataset from Hugging Face. This Learning Path uses the `train_1.parquet` file from the [Ritvik19/Sudoku-Dataset](https://huggingface.co/datasets/Ritvik19/Sudoku-Dataset) repository.
 
-3. Run the generator
+Download the dataset file:
+```console
+wget https://huggingface.co/datasets/Ritvik19/Sudoku-Dataset/resolve/main/train_1.parquet
+```
+
+Alternatively, you can download it manually from the [direct link](https://huggingface.co/datasets/Ritvik19/Sudoku-Dataset/blob/main/train_1.parquet) and save it as `train_1.parquet` in your working directory.
+
+The Parquet file should be placed next to the script, or you can update `PARQUET_PATH` in the code to point to its location.
+
+3. Run the generator:
 ```console
 python3 02_PrepareData.py
 ```

content/learning-paths/mobile-graphics-and-gaming/onnx/04_training.md

@@ -61,7 +61,6 @@ import torch.nn.functional as F
 from torch.utils.data import DataLoader
 from torchvision import datasets, transforms
 from tqdm import tqdm
-from torch.onnx import dynamo_export
 from torch.export import Dim
 import onnxruntime as ort  
 
@@ -189,6 +188,15 @@ The file will create two artifacts:
 Right after export, the script runs a parity test: it feeds the same randomly generated batch through both the PyTorch model and the ONNX model (executed by ONNX Runtime) and prints the mean absolute error between their logits. A tiny value confirms the exported graph faithfully matches your trained network.
 
 ## Running the script
+
+{{% notice Note %}}
+The Dynamo-based ONNX exporter requires PyTorch 2.1 or later. If you encounter errors related to `torch.export.Dim` or the `dynamo` parameter, ensure you have an up-to-date PyTorch installation:
+
+```console
+pip install --upgrade torch torchvision
+```
+{{% /notice %}}
+
 To run the training script, type:
 
 ```console

content/learning-paths/mobile-graphics-and-gaming/onnx/05_inference.md

@@ -213,6 +213,13 @@ These artifacts provide both quantitative and qualitative insight into model per
 In the sample grid, each tile shows one crop together with its True label (T:) and Predicted label (P:), with correct predictions highlighted in green and mistakes highlighted in red. This makes it easy to quickly verify that the classifier behaves sensibly and to spot remaining failure modes.
 
 ## Running the script
+
+Before running the evaluation script, install matplotlib for visualization:
+
+```console
+pip install matplotlib
+```
+
 Run the evaluation script from the project root:
 
 ```console

content/learning-paths/mobile-graphics-and-gaming/onnx/_index.md

@@ -17,7 +17,7 @@ learning_objectives:
   - Deploy an optimized ONNX model inside an Android app.
 
 prerequisites:
-  - A development machine with Python 3.10+ installed.
+  - A development machine with Python 3.10 installed (Python 3.11 also works; Python 3.12 is not yet supported by ONNX Runtime on Arm platforms).
   - Basic familiarity with PyTorch or TensorFlow.
   - An Arm64 device (e.g., Raspberry Pi or Android smartphone).
   - "[Android Studio](https://developer.android.com/studio) installed for deployment testing."