From 3e77ce06980d5c8a55fbd789c463cf517d8c834d Mon Sep 17 00:00:00 2001 From: Nexisato <978452096@qq.com> Date: Wed, 20 May 2026 21:31:03 +0800 Subject: [PATCH 1/6] fix: typo docs --- .vitepress/en.ts | 22 +++++++------- .vitepress/theme/custom.css | 7 ++--- .vitepress/zh.ts | 22 +++++++------- en/api/cli-swanlab-convert.md | 8 ++--- en/api/py-sync-mlflow.md | 2 +- en/examples/audio_classification.md | 2 +- en/examples/cats_dogs_classification.md | 10 +++---- en/examples/ner.md | 12 ++++---- en/examples/qwen3-medical.md | 8 ++--- en/examples/qwen_vl_coco.md | 12 ++++---- en/examples/stable_diffusion.md | 6 ++-- en/examples/unet-medical-segmentation.md | 4 +-- en/guide_cloud/general/changelog.md | 10 +++---- en/guide_cloud/general/what-is-swanlab.md | 28 ++++++++--------- en/guide_cloud/integration/index.md | 16 +++++----- .../integration-huggingface-trl.md | 2 +- .../integration/integration-mlflow.md | 4 +-- .../integration/integration-mmengine.md | 2 +- .../integration/integration-omegaconf.md | 2 +- .../integration-paddledetection.md | 2 +- .../integration/integration-paddleyolo.md | 10 +++---- en/guide_cloud/integration/integration-sb3.md | 2 +- .../integration/integration-specforge.md | 4 +-- .../integration/integration-swift.md | 4 +-- .../integration/integration-tensorboard.md | 30 +++++++++---------- .../integration/integration-wandb.md | 24 +++++++-------- .../integration/integration-xtuner.md | 8 ++--- en/index.md | 8 ++--- en/plugin/notification-bark.md | 2 +- en/plugin/notification-dingtalk.md | 2 +- en/plugin/notification-discord.md | 2 +- en/plugin/notification-email.md | 2 +- en/plugin/notification-lark.md | 2 +- en/plugin/notification-slack.md | 2 +- en/plugin/notification-telegram.md | 2 +- en/plugin/notification-wxwork.md | 2 +- en/plugin/writer-csv.md | 2 +- zh/api/cli-swanlab-convert.md | 22 +++++++------- zh/api/py-sync-mlflow.md | 2 +- .../01-traditionmodel/3.rnn/rnn_tutorial_2.md | 2 +- .../llm_train_course/03-sft/2.ner/README.md | 8 ++--- .../03-sft/4.qwen3-medical-finetune/README.md | 10 +++---- .../8.other_frameworks/paddlenlp_finetune.md | 2 +- .../06-multillm/1.qwen_vl_coco/README.md | 4 +-- .../06-multillm/3.stable_diffusion/README.md | 6 ++-- .../11-swanlab_rag/1.swanlab-rag.md | 8 ++--- zh/examples/agent/swanlab-rag.md | 8 ++--- zh/examples/audio_classification.md | 4 +-- zh/examples/cats_dogs_classification.md | 10 +++---- zh/examples/ner.md | 8 ++--- zh/examples/paddlenlp_finetune.md | 2 +- zh/examples/qwen3-medical.md | 10 +++---- zh/examples/qwen_vl_coco.md | 4 +-- zh/examples/rnn_tutorial_2.md | 2 +- zh/examples/stable_diffusion.md | 6 ++-- zh/examples/unet-medical-segmentation.md | 4 +-- .../experiment_track/view-result.md | 2 +- zh/guide_cloud/general/changelog.md | 2 +- zh/guide_cloud/general/what-is-swanlab.md | 24 +++++++-------- zh/guide_cloud/integration/index.md | 16 +++++----- .../integration-huggingface-trl.md | 2 +- .../integration/integration-mlflow.md | 4 +-- .../integration/integration-mmengine.md | 2 +- .../integration/integration-omegaconf.md | 2 +- .../integration-paddledetection.md | 2 +- .../integration/integration-paddleyolo.md | 10 +++---- zh/guide_cloud/integration/integration-sb3.md | 4 +-- .../integration/integration-specforge.md | 4 +-- .../integration/integration-swift.md | 4 +-- .../integration/integration-tensorboard.md | 30 +++++++++---------- .../integration/integration-wandb.md | 22 +++++++------- .../integration/integration-xtuner.md | 8 ++--- zh/index.md | 10 +++---- zh/plugin/notification-lark.md | 2 +- zh/plugin/notification-telegram.md | 2 +- 75 files changed, 281 insertions(+), 282 deletions(-) diff --git a/.vitepress/en.ts b/.vitepress/en.ts index d12bfba0..d541b239 100644 --- a/.vitepress/en.ts +++ b/.vitepress/en.ts @@ -43,7 +43,7 @@ export const en = defineConfig({ { text: 'Swift', link: base_path_guide_cloud + '/integration/integration-swift'}, { text: 'Ultralytics', link: base_path_guide_cloud + '/integration/integration-ultralytics'}, { text: 'veRL', link: base_path_guide_cloud + '/integration/integration-verl'}, - { text: 'Sb3', link: base_path_guide_cloud + '/integration/integration-sb3'}, + { text: 'SB3', link: base_path_guide_cloud + '/integration/integration-sb3'}, ] }, { @@ -62,7 +62,7 @@ export const en = defineConfig({ { text: 'Community', link: 'https://swanlab.cn/benchmarks' }, { text: 'Join Us', link: 'https://rcnpx636fedp.feishu.cn/wiki/BxtVwAc0siV0xrkCbPTcldBEnNP' }, { text: 'Feedback', link: 'https://geektechstudio.feishu.cn/share/base/form/shrcn8koDFRcH2mMcBYMh9tiKfI' }, - { text: 'Docs Github', link: 'https://github.com/SwanHubX/SwanLab-Docs' }, + { text: 'Docs GitHub', link: 'https://github.com/SwanHubX/SwanLab-Docs' }, ] }, { component: 'HeaderDocHelperButtonEN', @@ -206,7 +206,7 @@ function sidebarGuideCloud(): SidebarItemEx[] { // collapsed: false, items: [ { text: 'Online support', link: 'community/online-support'}, - { text: 'Github badge', link: 'community/github-badge'}, + { text: 'GitHub badge', link: 'community/github-badge'}, { text: 'Paper citation', link: 'community/paper-cite'}, { text: 'Contributing code', link: 'community/contributing-code'}, { text: 'Contributing docs', link: 'community/contributing-docs'}, @@ -225,7 +225,7 @@ function sidebarIntegration(): DefaultTheme.SidebarItem[] { { text: 'Argparse', link:'integration-argparse' }, { text: 'Areal', link: 'integration-areal' }, { text: 'Ascend NPU & MindSpore', link: 'integration-ascend' }, - { text: 'Catboost', link: 'integration-catboost'}, + { text: 'CatBoost', link: 'integration-catboost'}, { text: 'DiffSynth-Studio', link: 'integration-diffsynth-studio' }, { text: 'EasyR1', link: 'integration-easyr1' }, { text: 'EvalScope', link: 'integration-evalscope' }, @@ -238,20 +238,20 @@ function sidebarIntegration(): DefaultTheme.SidebarItem[] { items: [ { text: 'HuggingFace Accelerate', link: 'integration-huggingface-accelerate' }, { text: 'HuggingFace Transformers', link: 'integration-huggingface-transformers' }, - { text: 'HuggingFace Trl', link: 'integration-huggingface-trl' }, + { text: 'HuggingFace TRL', link: 'integration-huggingface-trl' }, { text: 'Hydra', link: 'integration-hydra' }, { text: 'Keras', link: 'integration-keras' }, { text: 'LightGBM', link: 'integration-lightgbm'}, { text: 'LLaMA Factory', link: 'integration-llama-factory'}, { text: 'LLaMA Factory Online', link: 'integration-llama-factory-online' }, { text: 'MindSpeed-RL', link: 'integration-mindspeed-rl' }, - { text: 'MLFlow', link: 'integration-mlflow'}, + { text: 'MLflow', link: 'integration-mlflow'}, { text: 'MLX LM', link: 'integration-mlx-lm' }, { text: 'MMEngine', link: 'integration-mmengine' }, { text: 'MMPretrain', link: 'integration-mmpretrain' }, { text: 'MMDetection', link: 'integration-mmdetection' }, { text: 'MMSegmentation', link: 'integration-mmsegmentation' }, - { text: 'Modelscope Swift', link: 'integration-swift' }, + { text: 'ModelScope Swift', link: 'integration-swift' }, { text: 'NVIDIA-NeMo RL', link: 'integration-nvidia-nemo-rl' }, ] }, @@ -260,7 +260,7 @@ function sidebarIntegration(): DefaultTheme.SidebarItem[] { // collapsed: false, items: [ { text: 'OpenAI', link: 'integration-openai' }, - { text: 'Omegaconf', link: 'integration-omegaconf' }, + { text: 'OmegaConf', link: 'integration-omegaconf' }, { text: 'PaddleDetection', link: 'integration-paddledetection' }, { text: 'PaddleNLP', link: 'integration-paddlenlp' }, { text: 'PaddleYOLO', link: 'integration-paddleyolo' }, @@ -271,9 +271,9 @@ function sidebarIntegration(): DefaultTheme.SidebarItem[] { { text: 'RLinf', link: 'integration-rlinf'}, { text: 'ROLL', link: 'integration-roll' }, { text: 'Sentence Transformers', link: 'integration-sentence-transformers'}, - { text: 'Specforge', link: 'integration-specforge'}, - { text: 'Stable Baseline3', link: 'integration-sb3' }, - { text: 'Tensorboard', link: 'integration-tensorboard'}, + { text: 'SpecForge', link: 'integration-specforge'}, + { text: 'Stable Baselines3', link: 'integration-sb3' }, + { text: 'TensorBoard', link: 'integration-tensorboard'}, ] }, { diff --git a/.vitepress/theme/custom.css b/.vitepress/theme/custom.css index f331f80a..9c2a5e25 100644 --- a/.vitepress/theme/custom.css +++ b/.vitepress/theme/custom.css @@ -1,10 +1,9 @@ :root { --vp-c-brand-1: #107d8b; - --vp-layout-max-width:1660px; + --vp-layout-max-width: 100%; --vp-sidebar-bg-color: var(--vp-c-bg-alt); --vp-sidebar-width: 320px; --vp-sidebar-padding-left: 16px; - /* --vp-layout-max-width: 65%; */ } @@ -17,11 +16,11 @@ } */ .VPDoc.has-aside .content-container { - max-width: 800px !important; + max-width: none !important; } .VPDoc.has-aside .content { - max-width: 800px !important; + max-width: none !important; } .title { diff --git a/.vitepress/zh.ts b/.vitepress/zh.ts index 4482060b..85909835 100644 --- a/.vitepress/zh.ts +++ b/.vitepress/zh.ts @@ -51,7 +51,7 @@ export const zh = defineConfig({ { text: 'MS-Swift', link: base_path_guide_cloud + '/integration/integration-swift' }, { text: 'veRL', link: base_path_guide_cloud + '/integration/integration-verl' }, { text: 'Ultralytics', link: base_path_guide_cloud + '/integration/integration-ultralytics' }, - { text: 'Sb3', link: base_path_guide_cloud + '/integration/integration-sb3' }, + { text: 'SB3', link: base_path_guide_cloud + '/integration/integration-sb3' }, ] }, { @@ -255,7 +255,7 @@ function sidebarGuideCloud(): DefaultTheme.SidebarItem[] { // collapsed: false, items: [ { text: '在线支持', link: 'community/online-support' }, - { text: 'Github徽章', link: 'community/github-badge' }, + { text: 'GitHub徽章', link: 'community/github-badge' }, // { text: '论文引用', link: 'community/paper-cite'}, // { text: '贡献代码', link: 'community/contributing-code'}, // { text: '贡献官方文档', link: 'community/contributing-docs'}, @@ -274,7 +274,7 @@ function sidebarIntegration(): DefaultTheme.SidebarItem[] { { text: 'Argparse', link: 'integration-argparse' }, { text: 'Areal', link: 'integration-areal' }, { text: 'Ascend NPU & MindSpore', link: 'integration-ascend' }, - { text: 'Catboost', link: 'integration-catboost'}, + { text: 'CatBoost', link: 'integration-catboost'}, { text: 'DiffSynth-Studio', link: 'integration-diffsynth-studio' }, { text: 'EasyR1', link: 'integration-easyr1' }, { text: 'EvalScope', link: 'integration-evalscope' }, @@ -287,20 +287,20 @@ function sidebarIntegration(): DefaultTheme.SidebarItem[] { items: [ { text: 'HuggingFace Accelerate', link: 'integration-huggingface-accelerate' }, { text: 'HuggingFace Transformers', link: 'integration-huggingface-transformers' }, - { text: 'HuggingFace Trl', link: 'integration-huggingface-trl' }, + { text: 'HuggingFace TRL', link: 'integration-huggingface-trl' }, { text: 'Hydra', link: 'integration-hydra' }, { text: 'Keras', link: 'integration-keras' }, { text: 'LightGBM', link: 'integration-lightgbm' }, { text: 'LLaMA-Factory', link: 'integration-llama-factory' }, { text: 'LLaMA-Factory Online', link: 'integration-llama-factory-online' }, { text: 'MindSpeed-RL', link: 'integration-mindspeed-rl' }, - { text: 'MLFlow', link: 'integration-mlflow' }, + { text: 'MLflow', link: 'integration-mlflow' }, { text: 'MLX LM', link: 'integration-mlx-lm' }, { text: 'MMEngine', link: 'integration-mmengine' }, { text: 'MMPretrain', link: 'integration-mmpretrain' }, { text: 'MMDetection', link: 'integration-mmdetection' }, { text: 'MMSegmentation', link: 'integration-mmsegmentation' }, - { text: 'Modelscope Swift', link: 'integration-swift' }, + { text: 'ModelScope Swift', link: 'integration-swift' }, { text: 'NVIDIA-NeMo RL', link: 'integration-nvidia-nemo-rl' }, ] }, @@ -309,7 +309,7 @@ function sidebarIntegration(): DefaultTheme.SidebarItem[] { // collapsed: false, items: [ { text: 'OpenAI', link: 'integration-openai' }, - { text: 'Omegaconf', link: 'integration-omegaconf' }, + { text: 'OmegaConf', link: 'integration-omegaconf' }, { text: 'PaddleDetection', link: 'integration-paddledetection' }, { text: 'PaddleNLP', link: 'integration-paddlenlp' }, { text: 'PaddleYOLO', link: 'integration-paddleyolo' }, @@ -320,9 +320,9 @@ function sidebarIntegration(): DefaultTheme.SidebarItem[] { { text: 'RLinf', link: 'integration-rlinf' }, { text: 'ROLL', link: 'integration-roll' }, { text: 'Sentence Transformers', link: 'integration-sentence-transformers' }, - { text: 'Specforge', link: 'integration-specforge' }, - { text: 'Stable Baseline3', link: 'integration-sb3' }, - { text: 'Tensorboard', link: 'integration-tensorboard' }, + { text: 'SpecForge', link: 'integration-specforge' }, + { text: 'Stable Baselines3', link: 'integration-sb3' }, + { text: 'TensorBoard', link: 'integration-tensorboard' }, ] }, { @@ -349,7 +349,7 @@ function sidebarExamples(): DefaultTheme.SidebarItem[] { { text: 'Hello_World', link: 'hello_world' }, { text: 'MNIST手写体识别', link: 'mnist' }, { text: 'FashionMNIST', link: 'fashionmnist' }, - { text: 'Cifar10图像分类', link: 'cifar10' }, + { text: 'CIFAR10图像分类', link: 'cifar10' }, ] }, { diff --git a/en/api/cli-swanlab-convert.md b/en/api/cli-swanlab-convert.md index d8f7e894..a7a9bdab 100644 --- a/en/api/cli-swanlab-convert.md +++ b/en/api/cli-swanlab-convert.md @@ -11,7 +11,7 @@ swanlab convert [OPTIONS] | `-w`, `--workspace` | Set the workspace where the SwanLab project is located, default is None. | | `-l`, `--logdir` | Set the log file save path for the SwanLab project, default is None. | | `--cloud` | Set whether the SwanLab project logs are uploaded to the cloud, default is True. | -| `--tb-logdir` | Path to the Tensorboard log files (tfevent) to be converted. | +| `--tb-logdir` | Path to the TensorBoard log files (tfevent) to be converted. | | `--wb-project` | Name of the Wandb project to be converted. | | `--wb-entity` | Entity where the Wandb project to be converted is located. | | `--wb-runid` | ID of the Wandb Run to be converted. | @@ -19,13 +19,13 @@ swanlab convert [OPTIONS] ## Introduction Convert content from other logging tools into SwanLab projects. -Supported tools for conversion include: `Tensorboard`, `Weights & Biases`. +Supported tools for conversion include: `TensorBoard`, `Weights & Biases`. ## Usage Examples -### Tensorboard +### TensorBoard -[Integration - Tensorboard](/en/guide_cloud/integration/integration-tensorboard.md) +[Integration - TensorBoard](/en/guide_cloud/integration/integration-tensorboard.md) ### Weights & Biases diff --git a/en/api/py-sync-mlflow.md b/en/api/py-sync-mlflow.md index 543bf940..218705cf 100644 --- a/en/api/py-sync-mlflow.md +++ b/en/api/py-sync-mlflow.md @@ -1,3 +1,3 @@ # swanlab.sync_mlflow -Sync MLFlow projects to SwanLab, [Documentation](/en/guide_cloud/integration/integration-mlflow.md) +Sync MLflow projects to SwanLab, [Documentation](/en/guide_cloud/integration/integration-mlflow.md) diff --git a/en/examples/audio_classification.md b/en/examples/audio_classification.md index 585315ae..bf192888 100644 --- a/en/examples/audio_classification.md +++ b/en/examples/audio_classification.md @@ -12,7 +12,7 @@ In current audio classification applications, it is often used for audio annotat In this article, we will train a ResNet series model on the GTZAN dataset using the PyTorch framework, and use [SwanLab](https://swanlab.cn) to monitor the training process and evaluate the model's performance. -* Github: [https://github.com/Zeyi-Lin/PyTorch-Audio-Classification](https://github.com/Zeyi-Lin/PyTorch-Audio-Classification) +* GitHub: [https://github.com/Zeyi-Lin/PyTorch-Audio-Classification](https://github.com/Zeyi-Lin/PyTorch-Audio-Classification) * Dataset: [https://pan.baidu.com/s/14CTI_9MD1vXCqyVxmAbeMw?pwd=1a9e](https://pan.baidu.com/s/14CTI_9MD1vXCqyVxmAbeMw?pwd=1a9e) Extraction Code: 1a9e * SwanLab Experiment Logs: [https://swanlab.cn/@ZeyiLin/PyTorch\_Audio\_Classification-simple/charts](https://swanlab.cn/@ZeyiLin/PyTorch\_Audio\_Classification-simple/charts) * More Experiment Logs: [https://swanlab.cn/@ZeyiLin/PyTorch\_Audio\_Classification/charts](https://swanlab.cn/@ZeyiLin/PyTorch\_Audio\_Classification/charts) diff --git a/en/examples/cats_dogs_classification.md b/en/examples/cats_dogs_classification.md index 9c616eef..128fa483 100644 --- a/en/examples/cats_dogs_classification.md +++ b/en/examples/cats_dogs_classification.md @@ -11,7 +11,7 @@ Cat and dog classification is one of the most fundamental tasks in computer visi ![](https://swanlab-docs-1301372061.cos.ap-beijing.myqcloud.com/assets/en/examples/cats_dogs/01.png) - You can view the experiment process on this webpage: [Cat and Dog Classification | SwanLab](https://swanlab.cn/@ZeyiLin/Cats_Dogs_Classification/runs/jzo93k112f15pmx14vtxf/chart) -- Code: [Github](https://github.com/Zeyi-Lin/Resnet50-cats_vs_dogs) +- Code: [GitHub](https://github.com/Zeyi-Lin/Resnet50-cats_vs_dogs) - Online Demo: [HuggingFace](https://huggingface.co/spaces/TheEeeeLin/Resnet50-cats_vs_dogs) - Dataset: [Baidu Cloud](https://pan.baidu.com/s/1qYa13SxFM0AirzDyFMy0mQ) Extraction code: 1ybm - Three open-source libraries: [SwanLab](https://github.com/swanhubx/swanlab), [Gradio](https://github.com/gradio-app/gradio), [PyTorch](https://github.com/pytorch/pytorch) @@ -46,7 +46,7 @@ Their respective functions are: ### 1.3 Download the Cat and Dog Classification Dataset -The dataset source is the [Cat and Dog Classification Dataset](https://modelscope.cn/datasets/tany0699/cats_and_dogs/summary) on Modelscope, which contains 275 images for training and 70 images for testing, totaling less than 10MB. +The dataset source is the [Cat and Dog Classification Dataset](https://modelscope.cn/datasets/tany0699/cats_and_dogs/summary) on ModelScope, which contains 275 images for training and 70 images for testing, totaling less than 10MB. I have organized the data, so it is recommended to download it using the following Baidu Netdisk link: > Baidu Netdisk: Link: https://pan.baidu.com/s/1qYa13SxFM0AirzDyFMy0mQ Extraction code: 1ybm @@ -216,7 +216,7 @@ optimizer = torch.optim.Adam(model.parameters(), lr=lr) ### 2.6 Initialize SwanLab In training, we use the `swanlab` library as the experiment management and metric visualization tool. -[swanlab](https://github.com/SwanHubX/SwanLab) is an open-source training chart visualization library similar to Tensorboard, with a lighter volume and more friendly API. In addition to recording metrics, it can automatically record training logs, hardware environment, Python environment, training time, and other information. +[swanlab](https://github.com/SwanHubX/SwanLab) is an open-source training chart visualization library similar to TensorBoard, with a lighter volume and more friendly API. In addition to recording metrics, it can automatically record training logs, hardware environment, Python environment, training time, and other information. ![Insert image description here](https://swanlab-docs-1301372061.cos.ap-beijing.myqcloud.com/assets/en/examples/cats_dogs/07.png) @@ -544,8 +544,8 @@ If this was helpful, please give it a thumbs up and save it! ## 4. Related Links - View the experiment process online: [Cat and Dog Classification · SwanLab](https://swanlab.cn/@ZeyiLin/Cats_Dogs_Classification/runs/jzo93k112f15pmx14vtxf/chart) -- SwanLab: [Github](https://github.com/SwanHubX/SwanLab) -- Cat and Dog Classification Code: [Github](https://github.com/xiaolin199912/Resnet50-cats_vs_dogs) +- SwanLab: [GitHub](https://github.com/SwanHubX/SwanLab) +- Cat and Dog Classification Code: [GitHub](https://github.com/xiaolin199912/Resnet50-cats_vs_dogs) - Online Demo: [HuggingFace](https://huggingface.co/spaces/TheEeeeLin/Resnet50-cats_vs_dogs) - Cat and Dog Classification Dataset (300 images): [ModelScope](https://modelscope.cn/datasets/tany0699/cats_and_dogs/summary) - Baidu Cloud Download: [Link](https://pan.baidu.com/s/1qYa13SxFM0AirzDyFMy0mQ) Extraction code: 1ybm diff --git a/en/examples/ner.md b/en/examples/ner.md index f1f7c88b..f9d2b016 100644 --- a/en/examples/ner.md +++ b/en/examples/ner.md @@ -12,9 +12,9 @@ Using Qwen2 as the base model, we perform high-precision Named Entity Recognitio In this tutorial, we'll fine-tune the [Qwen2-1.5b-Instruct](https://modelscope.cn/models/qwen/Qwen2-1.5B-Instruct/summary) model on the [Chinese NER](https://huggingface.co/datasets/qgyd2021/chinese_ner_sft) dataset while monitoring the training process and evaluating model performance using [SwanLab](https://swanlab.cn). -• Code: See Section 5 or [Github](https://github.com/Zeyi-Lin/LLM-Finetune) +• Code: See Section 5 or [GitHub](https://github.com/Zeyi-Lin/LLM-Finetune) • Training logs: [Qwen2-1.5B-NER-Fintune - SwanLab](https://swanlab.cn/@ZeyiLin/Qwen2-NER-fintune/runs/9gdyrkna1rxjjmz0nks2c/chart) -• Model: [Modelscope](https://modelscope.cn/models/qwen/Qwen2-1.5B-Instruct/summary) +• Model: [ModelScope](https://modelscope.cn/models/qwen/Qwen2-1.5B-Instruct/summary) • Dataset: [chinese_ner_sft](https://huggingface.co/datasets/qgyd2021/chinese_ner_sft) • SwanLab: [https://swanlab.cn](https://swanlab.cn) @@ -114,7 +114,7 @@ Download the dataset from [chinese_ner_sft - huggingface](https://huggingface.co ## 3. Load the Model -We'll download the Qwen2-1.5B-Instruct model via Modelscope (which has stable domestic speeds in China) and load it into Transformers for training: +We'll download the Qwen2-1.5B-Instruct model via ModelScope (which has stable domestic speeds in China) and load it into Transformers for training: ```python from modelscope import snapshot_download, AutoTokenizer @@ -123,7 +123,7 @@ from transformers import AutoModelForCausalLM, TrainingArguments, Trainer, DataC model_id = "qwen/Qwen2-1.5B-Instruct" model_dir = "./qwen/Qwen2-1___5B-Instruct" -# Download Qwen model from Modelscope +# Download Qwen model from ModelScope model_dir = snapshot_download(model_id, cache_dir="./", revision="master") # Load model weights into Transformers @@ -443,8 +443,8 @@ Output: ## Related Links -- Code: See Section 5 or [Github](https://github.com/Zeyi-Lin/LLM-Finetune) +- Code: See Section 5 or [GitHub](https://github.com/Zeyi-Lin/LLM-Finetune) - Training logs: [Qwen2-1.5B-NER-Fintune - SwanLab](https://swanlab.cn/@ZeyiLin/Qwen2-NER-fintune/runs/9gdyrkna1rxjjmz0nks2c/chart) -- Model: [Modelscope](https://modelscope.cn/models/qwen/Qwen2-1.5B-Instruct/summary) +- Model: [ModelScope](https://modelscope.cn/models/qwen/Qwen2-1.5B-Instruct/summary) - Dataset: [chinese_ner_sft](https://huggingface.co/datasets/qgyd2021/chinese_ner_sft) - SwanLab: [https://swanlab.cn](https://swanlab.cn) \ No newline at end of file diff --git a/en/examples/qwen3-medical.md b/en/examples/qwen3-medical.md index a5cb2c6f..7f11b7c1 100644 --- a/en/examples/qwen3-medical.md +++ b/en/examples/qwen3-medical.md @@ -16,9 +16,9 @@ In this article, we will fine-tune the [Qwen3-1.7b](https://www.modelscope.cn/mo > Full-parameter fine-tuning requires approximately **32GB of GPU memory**. If your GPU memory is insufficient, consider using Qwen3-0.6b or LoRA fine-tuning. -- **Code**: [Github](https://github.com/Zeyi-Lin/Qwen3-Medical-SFT) (or see Section 5 below) +- **Code**: [GitHub](https://github.com/Zeyi-Lin/Qwen3-Medical-SFT) (or see Section 5 below) - **Training Logs**: [qwen3-1.7B-linear - SwanLab](https://swanlab.cn/@ZeyiLin/qwen3-sft-medical/runs/agps0dkifth5l1xytcdyk/chart) (or search "qwen3-sft-medical" in [SwanLab Benchmark Community](https://swanlab.cn/benchmarks)) -- **Model**: [Modelscope](https://modelscope.cn/models/Qwen/Qwen3-1.7B) +- **Model**: [ModelScope](https://modelscope.cn/models/Qwen/Qwen3-1.7B) - **Dataset**: [delicate_medical_r1_data](https://modelscope.cn/datasets/krisfu/delicate_medical_r1_data) - **SwanLab**: [https://swanlab.cn](https://swanlab.cn) @@ -155,7 +155,7 @@ model = AutoModelForCausalLM.from_pretrained("./Qwen/Qwen3-1.7B", device_map="au We use **SwanLab** to monitor training and evaluate model performance. -SwanLab is an open-source, lightweight AI training tracking and visualization tool, often called the "Chinese Weights & Biases + Tensorboard." It supports cloud/offline use and integrates with 40+ frameworks (PyTorch, Transformers, etc.). +SwanLab is an open-source, lightweight AI training tracking and visualization tool, often called the "Chinese Weights & Biases + TensorBoard." It supports cloud/offline use and integrates with 40+ frameworks (PyTorch, Transformers, etc.). ![09-05](https://swanlab-docs-1301372061.cos.ap-beijing.myqcloud.com/assets/en/examples/qwen3/05.png) ![09-06](https://swanlab-docs-1301372061.cos.ap-beijing.myqcloud.com/assets/en/examples/qwen3/06.png) @@ -348,7 +348,7 @@ print(predict(messages, model, tokenizer)) ## References -- **Code**: [Github](https://github.com/Zeyi-Lin/Qwen3-Medical-SFT) +- **Code**: [GitHub](https://github.com/Zeyi-Lin/Qwen3-Medical-SFT) - **Training Logs**: [SwanLab](https://swanlab.cn/@ZeyiLin/qwen3-sft-medical/runs/agps0dkifth5l1xytcdyk/chart) - **Model**: [ModelScope](https://modelscope.cn/models/Qwen/Qwen3-1.7B) - **Dataset**: [delicate_medical_r1_data](https://modelscope.cn/datasets/krisfu/delicate_medical_r1_data) diff --git a/en/examples/qwen_vl_coco.md b/en/examples/qwen_vl_coco.md index 50efcd75..38af6fdc 100644 --- a/en/examples/qwen_vl_coco.md +++ b/en/examples/qwen_vl_coco.md @@ -15,7 +15,7 @@ Qwen2-VL is a multimodal large model developed by Alibaba's Tongyi Lab. This art LoRA is an efficient fine-tuning method. For a deeper understanding of its principles, refer to the blog: [Zhihu | An In-Depth Explanation of LoRA](https://zhuanlan.zhihu.com/p/650197598). • Training Process: [Qwen2-VL-finetune](https://swanlab.cn/@ZeyiLin/Qwen2-VL-finetune/runs/pkgest5xhdn3ukpdy6kv5/chart) -• Github: [Code Repository](https://github.com/Zeyi-Lin/LLM-Finetune/tree/main/qwen2_vl), [self-llm](https://github.com/datawhalechina/self-llm) +• GitHub: [Code Repository](https://github.com/Zeyi-Lin/LLM-Finetune/tree/main/qwen2_vl), [self-llm](https://github.com/datawhalechina/self-llm) • Dataset: [coco_2014_caption](https://modelscope.cn/datasets/modelscope/coco_2014_caption/summary) • Model: [Qwen2-VL-2B-Instruct](https://modelscope.cn/models/Qwen/Qwen2-VL-2B-Instruct) @@ -80,7 +80,7 @@ Here, "from" represents the role (`user` for human, `assistant` for the model), **Dataset Download and Processing Steps** 1. **We need to do four things:** - ◦ Download the coco_2014_caption dataset via Modelscope. + ◦ Download the coco_2014_caption dataset via ModelScope. ◦ Load the dataset and save the images locally. ◦ Convert the image paths and captions into a CSV file. ◦ Convert the CSV file into a JSON file. @@ -99,7 +99,7 @@ MAX_DATA_NUMBER = 500 # Check if the directory already exists if not os.path.exists('coco_2014_caption'): - # Download the COCO 2014 image caption dataset from Modelscope + # Download the COCO 2014 image caption dataset from ModelScope ds = MsDataset.load('modelscope/coco_2014_caption', subset_name='coco_2014_caption', split='train') print(len(ds)) # Set the maximum number of images to process @@ -187,14 +187,14 @@ With this, the dataset preparation is complete. ## 3. Model Download and Loading -Here, we use Modelscope to download the Qwen2-VL-2B-Instruct model and load it into Transformers for training: +Here, we use ModelScope to download the Qwen2-VL-2B-Instruct model and load it into Transformers for training: ```python from modelscope import snapshot_download, AutoTokenizer from transformers import TrainingArguments, Trainer, DataCollatorForSeq2Seq, Qwen2VLForConditionalGeneration, AutoProcessor import torch -# Download the Qwen2-VL model from Modelscope to a local directory +# Download the Qwen2-VL model from ModelScope to a local directory model_dir = snapshot_download("Qwen/Qwen2-VL-2B-Instruct", cache_dir="./", revision="master") # Load the model weights using Transformers @@ -365,7 +365,7 @@ def predict(messages, model): return output_text[0] -# Download the Qwen2-VL model from Modelscope to a local directory +# Download the Qwen2-VL model from ModelScope to a local directory model_dir = snapshot_download("Qwen/Qwen2-VL-2B-Instruct", cache_dir="./", revision="master") # Load the model weights using Transformers diff --git a/en/examples/stable_diffusion.md b/en/examples/stable_diffusion.md index 5e04bd46..2759e472 100644 --- a/en/examples/stable_diffusion.md +++ b/en/examples/stable_diffusion.md @@ -12,7 +12,7 @@ Using SD1.5 as a pre-trained model, fine-tuning a Naruto-style text-to-image mod In this article, we will use the [SD-1.5](https://huggingface.co/runwayml/stable-diffusion-v1-5) model to train on the [Naruto](https://huggingface.co/datasets/lambdalabs/naruto-blip-captions) dataset, while monitoring the training process and evaluating model performance using [SwanLab](https://swanlab.cn). -- Code: [Github](https://github.com/Zeyi-Lin/Stable-Diffusion-Example) +- Code: [GitHub](https://github.com/Zeyi-Lin/Stable-Diffusion-Example) - Experiment Log: [SD-naruto - SwanLab](https://swanlab.cn/@ZeyiLin/SD-Naruto/runs/21flglg1lbnqo67a6f1kr/environment/requirements) - Model: [runwayml/stable-diffusion-v1-5](https://huggingface.co/runwayml/stable-diffusion-v1-5) - Dataset: [lambdalabs/naruto-blip-captions](https://huggingface.co/datasets/lambdalabs/naruto-blip-captions) @@ -93,7 +93,7 @@ If this is your first time using SwanLab, you need to register an account at htt ## 5. Start Training -Since the training code is relatively long, I have placed it on [Github](https://github.com/Zeyi-Lin/Stable-Diffusion-Example/tree/main). Please clone the code: +Since the training code is relatively long, I have placed it on [GitHub](https://github.com/Zeyi-Lin/Stable-Diffusion-Example/tree/main). Please clone the code: ```bash git clone https://github.com/Zeyi-Lin/Stable-Diffusion-Example.git @@ -228,7 +228,7 @@ image.save("result.png") ## Related Links -- Code: [Github](https://github.com/Zeyi-Lin/Stable-Diffusion-Example) +- Code: [GitHub](https://github.com/Zeyi-Lin/Stable-Diffusion-Example) - Experiment Log: [SD-naruto - SwanLab](https://swanlab.cn/@ZeyiLin/SD-Naruto/runs/21flglg1lbnqo67a6f1kr/environment/requirements) - Model: [runwayml/stable-diffusion-v1-5](https://huggingface.co/runwayml/stable-diffusion-v1-5) - Dataset: [lambdalabs/naruto-blip-captions](https://huggingface.co/datasets/lambdalabs/naruto-blip-captions) diff --git a/en/examples/unet-medical-segmentation.md b/en/examples/unet-medical-segmentation.md index ed40dfa8..bdcf908f 100644 --- a/en/examples/unet-medical-segmentation.md +++ b/en/examples/unet-medical-segmentation.md @@ -12,7 +12,7 @@ UNet is a convolutional neural network (CNN)-based model for medical image segme ![](https://swanlab-docs-1301372061.cos.ap-beijing.myqcloud.com/assets/en/examples/unet-medical-segmentation/train_image.png) -• **Code**: Full code available in Section 5 or on [Github](https://github.com/Zeyi-Lin/UNet-Medical) +• **Code**: Full code available in Section 5 or on [GitHub](https://github.com/Zeyi-Lin/UNet-Medical) • **Training Logs**: [Unet-Medical-Segmentation - SwanLab](https://swanlab.cn/@ZeyiLin/Unet-Medical-Segmentation/runs/67konj7kdqhnfdmusy2u6/chart) • **Model**: UNet (implemented directly in PyTorch) • **Dataset**: [brain-tumor-image-dataset-semantic-segmentation - Kaggle](https://www.kaggle.com/datasets/pkdarabi/brain-tumor-image-dataset-semantic-segmentation) @@ -603,7 +603,7 @@ GPU memory usage was **6.124 GB**, meaning any GPU with ≥6GB VRAM can run this ## References -• **Code**: Full code in Section 5 or on [Github](https://github.com/Zeyi-Lin/UNet-Medical) +• **Code**: Full code in Section 5 or on [GitHub](https://github.com/Zeyi-Lin/UNet-Medical) • **Training Logs**: [Unet-Medical-Segmentation - SwanLab](https://swanlab.cn/@ZeyiLin/Unet-Medical-Segmentation/runs/67konj7kdqhnfdmusy2u6/chart) • **Model**: UNet (PyTorch implementation) • **Dataset**: [brain-tumor-image-dataset-semantic-segmentation - Kaggle](https://www.kaggle.com/datasets/pkdarabi/brain-tumor-image-dataset-semantic-segmentation) diff --git a/en/guide_cloud/general/changelog.md b/en/guide_cloud/general/changelog.md index e8b8e1dd..39fe56a9 100644 --- a/en/guide_cloud/general/changelog.md +++ b/en/guide_cloud/general/changelog.md @@ -314,7 +314,7 @@ Released SwanLab Kubernetes version, deployment instructions see [this document] - Introduced `swanlab.register_callback()`, enabling the registration of callback functions outside of `init`. [Documentation](/api/py-register-callback.html) - Upgraded `swanlab.login()` with new parameters `host`, `web_host`, and `save`, adapting to the characteristics of self-hosted deployment services and supporting the option to not write user login credentials locally for shared server scenarios. [Documentation](/zh/api/py-login.md) - Upgraded `swanlab login` with new parameters `host`, `web_host`, and `api-key`. [Documentation](/zh/api/cli-swanlab-login.md) -- Added support for using `swanlab.sync_mlflow()` to synchronize MLFlow projects to SwanLab. [Documentation](/guide_cloud/integration/integration-mlflow.md) +- Added support for using `swanlab.sync_mlflow()` to synchronize MLflow projects to SwanLab. [Documentation](/guide_cloud/integration/integration-mlflow.md) **🤔 Optimizations** - We have significantly optimized the SDK architecture, improving its performance in scenarios with a large number of metrics. @@ -343,7 +343,7 @@ Released SwanLab Kubernetes version, deployment instructions see [this document] **🚀 New Features** • Added integration with [DiffSynth-Studio](https://github.com/modelscope/DiffSynth-Studio), [Documentation](/en/guide_cloud/integration/integration-diffsynth-studio.md). -• Added support for converting **MLFlow** experiments to SwanLab. [Documentation](/en/guide_cloud/integration/integration-mlflow.md). +• Added support for converting **MLflow** experiments to SwanLab. [Documentation](/en/guide_cloud/integration/integration-mlflow.md). • Introduced **Project Descriptions**, allowing you to add short notes to your projects. **Improvements** @@ -364,7 +364,7 @@ Released SwanLab Kubernetes version, deployment instructions see [this document] ## v0.4.8 - 2025.2.16 **🚀 New Features** -- Added integration with Modelscope Swift, [Docs](/en/guide_cloud/integration/integration-swift.md) +- Added integration with ModelScope Swift, [Docs](/en/guide_cloud/integration/integration-swift.md) - Added `Add Group` and `Move Chart to Another Group` functions **Optimizations** @@ -389,7 +389,7 @@ Released SwanLab Kubernetes version, deployment instructions see [this document] ## v0.4.5 - 2025.1.22 **🚀New Features** -- Added `swanlab.sync_tensorboardX()` and `swanlab.sync_tensorboard_torch()`: Supports synchronizing metrics to SwanLab when using TensorboardX or PyTorch.utils.tensorboard for experiment tracking. [Docs](/en/guide_cloud/integration/integration-tensorboard.md) +- Added `swanlab.sync_tensorboardX()` and `swanlab.sync_tensorboard_torch()`: Supports synchronizing metrics to SwanLab when using TensorBoardX or PyTorch.utils.tensorboard for experiment tracking. [Docs](/en/guide_cloud/integration/integration-tensorboard.md) **Optimizations** - Optimized the code compatibility of `sync_wandb()` @@ -398,7 +398,7 @@ Released SwanLab Kubernetes version, deployment instructions see [this document] ## v0.4.3 - 2025.1.17 **🚀 New Features** -- Added `swanlab.sync_wandb()`: Supports synchronizing metrics to SwanLab when using Weights&Biases for experiment tracking. [Docs](/en/guide_cloud/integration/integration-wandb.md) +- Added `swanlab.sync_wandb()`: Supports synchronizing metrics to SwanLab when using Weights & Biases for experiment tracking. [Docs](/en/guide_cloud/integration/integration-wandb.md) - Added framework integration: Configuration items will now record the framework being used. **Optimizations** diff --git a/en/guide_cloud/general/what-is-swanlab.md b/en/guide_cloud/general/what-is-swanlab.md index 2d48b63c..f2a4bc14 100644 --- a/en/guide_cloud/general/what-is-swanlab.md +++ b/en/guide_cloud/general/what-is-swanlab.md @@ -1,6 +1,6 @@ # Welcome to SwanLab -[Official Website](https://swanlab.cn) · [Framework Integration](/guide_cloud/integration/integration-huggingface-transformers.html) · [Github](https://github.com/swanhubx/swanlab) · [Quick Start](/guide_cloud/general/quick-start.md) · [Sync WandB](/guide_cloud/integration/integration-wandb.md#_1-sync-tracking) · [Benchmark Community](https://swanlab.cn/benchmarks) +[Official Website](https://swanlab.cn) · [Framework Integration](/guide_cloud/integration/integration-huggingface-transformers.html) · [GitHub](https://github.com/swanhubx/swanlab) · [Quick Start](/guide_cloud/general/quick-start.md) · [Sync WandB](/guide_cloud/integration/integration-wandb.md#_1-sync-tracking) · [Benchmark Community](https://swanlab.cn/benchmarks) ::: warning 🎉 Self-Hosted Kubernetes Version Officially Released! The self-hosted Kubernetes version supports local use with features comparable to the public cloud edition. For deployment instructions, see [this document](/en/guide_cloud/self_host/kubernetes-deploy.md). @@ -62,7 +62,7 @@ Video Demo: - **Automatic Background Logging**: Logging, hardware environment, Git repo, Python environment, Python library list, project directory. - **Resume Training Logging**: Supports adding new metric data to the same experiment after training completes/interrupts. -**2. ⚡️ Comprehensive Framework Integration**: Supports **40+** frameworks including PyTorch, 🤗HuggingFace Transformers, PyTorch Lightning, 🦙LLaMA Factory, MMDetection, Ultralytics, PaddleDetection, LightGBM, XGBoost, Keras, Tensorboard, Weights&Biases, OpenAI, Swift, XTuner, Stable Baseline3, and Hydra. +**2. ⚡️ Comprehensive Framework Integration**: Supports **40+** frameworks including PyTorch, 🤗HuggingFace Transformers, PyTorch Lightning, 🦙LLaMA Factory, MMDetection, Ultralytics, PaddleDetection, LightGBM, XGBoost, Keras, TensorBoard, Weights & Biases, OpenAI, Swift, XTuner, Stable Baselines3, and Hydra. ![](https://swanlab-docs-1301372061.cos.ap-beijing.myqcloud.com/assets/assets/integrations.png) @@ -110,12 +110,12 @@ We hope this guide helps you understand SwanLab—we believe it can assist you. ## Comparison with Familiar Tools -### Tensorboard vs. SwanLab +### TensorBoard vs. SwanLab -- **☁️ Online Usage**: SwanLab syncs experiments to the cloud for remote monitoring, sharing, and collaboration. Tensorboard is offline-only. -- **👥 Collaboration**: SwanLab simplifies team training management, while Tensorboard is designed for individual use. -- **💻 Centralized Dashboard**: SwanLab aggregates results from any machine; Tensorboard requires manual TFEvent file management. -- **💪 Powerful Tables**: SwanLab tables support searching/filtering thousands of model versions. Tensorboard struggles with large projects. +- **☁️ Online Usage**: SwanLab syncs experiments to the cloud for remote monitoring, sharing, and collaboration. TensorBoard is offline-only. +- **👥 Collaboration**: SwanLab simplifies team training management, while TensorBoard is designed for individual use. +- **💻 Centralized Dashboard**: SwanLab aggregates results from any machine; TensorBoard requires manual TFEvent file management. +- **💪 Powerful Tables**: SwanLab tables support searching/filtering thousands of model versions. TensorBoard struggles with large projects. ### W&B vs. SwanLab @@ -135,7 +135,7 @@ Use SwanLab with your favorite frameworks! Below is our integration list. Submit - [PyTorch Lightning](/guide_cloud/integration/integration-pytorch-lightning.html) - [HuggingFace Transformers](/guide_cloud/integration/integration-huggingface-transformers.html) - [LLaMA Factory](/guide_cloud/integration/integration-llama-factory.html) -- [Modelscope Swift](/guide_cloud/integration/integration-swift.html) +- [ModelScope Swift](/guide_cloud/integration/integration-swift.html) - [DiffSynth-Studio](/guide_cloud/integration/integration-diffsynth-studio.html) - [Sentence Transformers](/guide_cloud/integration/integration-sentence-transformers.html) - [OpenMind](https://modelers.cn/docs/zh/openmind-library/1.0.0/basic_tutorial/finetune/finetune_pt.html#%E8%AE%AD%E7%BB%83%E7%9B%91%E6%8E%A7) @@ -157,9 +157,9 @@ Use SwanLab with your favorite frameworks! Below is our integration list. Submit - [PaddleYOLO](/guide_cloud/integration/integration-paddleyolo.html) **Reinforcement Learning** -- [Stable Baseline3](/guide_cloud/integration/integration-sb3.html) +- [Stable Baselines3](/guide_cloud/integration/integration-sb3.html) - [veRL](/guide_cloud/integration/integration-verl.html) -- [HuggingFace trl](/guide_cloud/integration/integration-huggingface-trl.html) +- [HuggingFace TRL](/guide_cloud/integration/integration-huggingface-trl.html) - [EasyR1](/guide_cloud/integration/integration-easyr1.html) - [AReaL](/guide_cloud/integration/integration-areal.html) - [ROLL](/guide_cloud/integration/integration-roll.html) @@ -167,12 +167,12 @@ Use SwanLab with your favorite frameworks! Below is our integration list. Submit - [MindSpeed-RL](/guide_cloud/integration/integration-mindspeed-rl.html) **Other Frameworks**: -- [Tensorboard](/guide_cloud/integration/integration-tensorboard.html) -- [Weights&Biases](/guide_cloud/integration/integration-wandb.html) -- [MLFlow](/guide_cloud/integration/integration-mlflow.html) +- [TensorBoard](/guide_cloud/integration/integration-tensorboard.html) +- [Weights & Biases](/guide_cloud/integration/integration-wandb.html) +- [MLflow](/guide_cloud/integration/integration-mlflow.html) - [HuggingFace Accelerate](/guide_cloud/integration/integration-huggingface-accelerate.html) - [Hydra](/guide_cloud/integration/integration-hydra.html) -- [Omegaconf](/guide_cloud/integration/integration-omegaconf.html) +- [OmegaConf](/guide_cloud/integration/integration-omegaconf.html) - [OpenAI](/guide_cloud/integration/integration-openai.html) - [ZhipuAI](/guide_cloud/integration/integration-zhipuai.html) diff --git a/en/guide_cloud/integration/index.md b/en/guide_cloud/integration/index.md index 0916e8a9..188b818e 100644 --- a/en/guide_cloud/integration/index.md +++ b/en/guide_cloud/integration/index.md @@ -12,7 +12,7 @@ Below is a list of frameworks we have integrated, please submit [Issue](https:// - [PyTorch Lightning](/en/guide_cloud/integration/integration-pytorch-lightning.html) - [HuggingFace Transformers](/en/guide_cloud/integration/integration-huggingface-transformers.html) - [LLaMA Factory](/en/guide_cloud/integration/integration-llama-factory.html) -- [Modelscope Swift](/en/guide_cloud/integration/integration-swift.html) +- [ModelScope Swift](/en/guide_cloud/integration/integration-swift.html) - [DiffSynth-Studio](/en/guide_cloud/integration/integration-diffsynth-studio.html) - [Sentence Transformers](/en/guide_cloud/integration/integration-sentence-transformers.html) - [PaddleNLP](/en/guide_cloud/integration/integration-paddlenlp.md) @@ -36,9 +36,9 @@ Below is a list of frameworks we have integrated, please submit [Issue](https:// - [PaddleYOLO](/en/guide_cloud/integration/integration-paddleyolo.html) ## Reinforcement Learning -- [Stable Baseline3](/en/guide_cloud/integration/integration-sb3.html) +- [Stable Baselines3](/en/guide_cloud/integration/integration-sb3.html) - [veRL](/en/guide_cloud/integration/integration-verl.html) -- [HuggingFace trl](/en/guide_cloud/integration/integration-huggingface-trl.html) +- [HuggingFace TRL](/en/guide_cloud/integration/integration-huggingface-trl.html) - [EasyR1](/en/guide_cloud/integration/integration-easyr1.html) - [AReaL](/en/guide_cloud/integration/integration-areal.html) - [ROLL](/en/guide_cloud/integration/integration-roll.html) @@ -47,13 +47,13 @@ Below is a list of frameworks we have integrated, please submit [Issue](https:// - [MindSpeed-RL](/en/guide_cloud/integration/integration-mindspeed-rl.html) ## Others: -- [Tensorboard](/en/guide_cloud/integration/integration-tensorboard.html) -- [Weights&Biases](/en/guide_cloud/integration/integration-wandb.html) -- [MLFlow](/en/guide_cloud/integration/integration-mlflow.html) +- [TensorBoard](/en/guide_cloud/integration/integration-tensorboard.html) +- [Weights & Biases](/en/guide_cloud/integration/integration-wandb.html) +- [MLflow](/en/guide_cloud/integration/integration-mlflow.html) - [HuggingFace Accelerate](/en/guide_cloud/integration/integration-huggingface-accelerate.html) - [Ray](/en/guide_cloud/integration/integration-ray.html) - [Hydra](/en/guide_cloud/integration/integration-hydra.html) -- [Omegaconf](/en/guide_cloud/integration/integration-omegaconf.html) +- [OmegaConf](/en/guide_cloud/integration/integration-omegaconf.html) - [OpenAI](/en/guide_cloud/integration/integration-openai.html) - [ZhipuAI](/en/guide_cloud/integration/integration-zhipuai.html) -- [Specforge](/en/guide_cloud/integration/integration-specforge.html) \ No newline at end of file +- [SpecForge](/en/guide_cloud/integration/integration-specforge.html) \ No newline at end of file diff --git a/en/guide_cloud/integration/integration-huggingface-trl.md b/en/guide_cloud/integration/integration-huggingface-trl.md index 605ad90a..6f454552 100644 --- a/en/guide_cloud/integration/integration-huggingface-trl.md +++ b/en/guide_cloud/integration/integration-huggingface-trl.md @@ -1,4 +1,4 @@ -# 🤗HuggingFace Trl +# 🤗HuggingFace TRL [TRL](https://github.com/huggingface/trl) (Transformers Reinforcement Learning) is a leading Python library designed to optimize foundational models through advanced techniques such as Supervised Fine-Tuning (SFT), Proximal Policy Optimization (PPO), and Direct Preference Optimization (DPO). Built on top of the 🤗 Transformers ecosystem, TRL supports multiple model architectures and modalities, and can scale across various hardware configurations. diff --git a/en/guide_cloud/integration/integration-mlflow.md b/en/guide_cloud/integration/integration-mlflow.md index c8abede9..011edca1 100644 --- a/en/guide_cloud/integration/integration-mlflow.md +++ b/en/guide_cloud/integration/integration-mlflow.md @@ -1,6 +1,6 @@ -# MLFlow +# MLflow -[MLFlow](https://github.com/mlflow/mlflow) is an open-source platform for managing the machine learning lifecycle, created and maintained by Databricks. It aims to help data scientists and machine learning engineers manage the entire lifecycle of machine learning projects more efficiently, including experiment tracking, model management, model deployment, and collaboration. MLflow is modular and can integrate with any machine learning library, framework, or tool. +[MLflow](https://github.com/mlflow/mlflow) is an open-source platform for managing the machine learning lifecycle, created and maintained by Databricks. It aims to help data scientists and machine learning engineers manage the entire lifecycle of machine learning projects more efficiently, including experiment tracking, model management, model deployment, and collaboration. MLflow is modular and can integrate with any machine learning library, framework, or tool. ![mlflow](https://swanlab-docs-1301372061.cos.ap-beijing.myqcloud.com/assets/en/guide_cloud/integration/mlflow/logo.png) diff --git a/en/guide_cloud/integration/integration-mmengine.md b/en/guide_cloud/integration/integration-mmengine.md index efcaf42a..34df9df6 100644 --- a/en/guide_cloud/integration/integration-mmengine.md +++ b/en/guide_cloud/integration/integration-mmengine.md @@ -25,7 +25,7 @@ Frameworks using MMEngine can all use the following methods to introduce SwanLab > You can check out which excellent frameworks are available under the [OpenMMLab official GitHub account](https://github.com/open-mmlab). -Some frameworks, such as [Xtuner](https://github.com/InternLM/xtuner), are not fully compatible with MMEngine and require some simple modifications. You can refer to [SwanLab's Xtuner Integration](https://docs.swanlab.cn/zh/guide_cloud/integration/integration-xtuner.html) to see how to use SwanLab in Xtuner. +Some frameworks, such as [XTuner](https://github.com/InternLM/xtuner), are not fully compatible with MMEngine and require some simple modifications. You can refer to [SwanLab's XTuner Integration](https://docs.swanlab.cn/zh/guide_cloud/integration/integration-xtuner.html) to see how to use SwanLab in XTuner. There are two methods to introduce SwanLab for experiment visualization tracking using MMEngine: diff --git a/en/guide_cloud/integration/integration-omegaconf.md b/en/guide_cloud/integration/integration-omegaconf.md index 7f273f95..6ae2ec65 100644 --- a/en/guide_cloud/integration/integration-omegaconf.md +++ b/en/guide_cloud/integration/integration-omegaconf.md @@ -1,4 +1,4 @@ -# Omegaconf +# OmegaConf OmegaConf is a Python library for handling configurations, especially useful in scenarios that require flexible configurations and configuration merging. Integrating OmegaConf with swanlab is very simple; just pass the `omegaconf` object to `swanlab.config` to record it as hyperparameters: diff --git a/en/guide_cloud/integration/integration-paddledetection.md b/en/guide_cloud/integration/integration-paddledetection.md index 7fd6fd40..33a1de9a 100644 --- a/en/guide_cloud/integration/integration-paddledetection.md +++ b/en/guide_cloud/integration/integration-paddledetection.md @@ -123,7 +123,7 @@ if self.cfg.get('use_swanlab', False) or 'swanlab' in self.cfg: self._callbacks.append(SwanLabCallback(self)) ``` -With this, you have completed the integration of SwanLab with PaddleYolo! Next, simply add `use_swanlab: True` to the training configuration file to start visualizing and tracking the training. +With this, you have completed the integration of SwanLab with PaddleYOLO! Next, simply add `use_swanlab: True` to the training configuration file to start visualizing and tracking the training. ## 3. Modify the Configuration File diff --git a/en/guide_cloud/integration/integration-paddleyolo.md b/en/guide_cloud/integration/integration-paddleyolo.md index 5582d339..7be7f37d 100644 --- a/en/guide_cloud/integration/integration-paddleyolo.md +++ b/en/guide_cloud/integration/integration-paddleyolo.md @@ -1,14 +1,14 @@ -# PaddleYolo +# PaddleYOLO -[PaddleYolo](https://github.com/PaddlePaddle/PaddleYOLO) is an object detection library under the PaddlePaddle framework, primarily used for object detection in images and videos. PaddleYOLO contains code related to the YOLO series models, supporting models such as YOLOv3, PP-YOLO, PP-YOLOv2, PP-YOLOE, PP-YOLOE+, RT-DETR, YOLOX, YOLOv5, YOLOv6, YOLOv7, YOLOv8, YOLOv5u, YOLOv7u, YOLOv6Lite, RTMDet, etc. +[PaddleYOLO](https://github.com/PaddlePaddle/PaddleYOLO) is an object detection library under the PaddlePaddle framework, primarily used for object detection in images and videos. PaddleYOLO contains code related to the YOLO series models, supporting models such as YOLOv3, PP-YOLO, PP-YOLOv2, PP-YOLOE, PP-YOLOE+, RT-DETR, YOLOX, YOLOv5, YOLOv6, YOLOv7, YOLOv8, YOLOv5u, YOLOv7u, YOLOv6Lite, RTMDet, etc. -You can use PaddleYolo to quickly train object detection models while using SwanLab for experiment tracking and visualization. +You can use PaddleYOLO to quickly train object detection models while using SwanLab for experiment tracking and visualization. [Demo](https://swanlab.cn/@ZeyiLin/PaddleYOLO/runs/10zy8zickn2062kubch34/chart) ## 1. Import SwanLabCallback -First, in your cloned PaddleYolo project, find the `ppdet/engine/callbacks.py` file and add the following code at the bottom: +First, in your cloned PaddleYOLO project, find the `ppdet/engine/callbacks.py` file and add the following code at the bottom: ```python class SwanLabCallback(Callback): @@ -123,7 +123,7 @@ if self.cfg.get('use_swanlab', False) or 'swanlab' in self.cfg: self._callbacks.append(SwanLabCallback(self)) ``` -With this, you have completed the integration of SwanLab with PaddleYolo! Next, simply add `use_swanlab: True` to the training configuration file to start visualizing and tracking the training. +With this, you have completed the integration of SwanLab with PaddleYOLO! Next, simply add `use_swanlab: True` to the training configuration file to start visualizing and tracking the training. ## 3. Modify the Configuration File diff --git a/en/guide_cloud/integration/integration-sb3.md b/en/guide_cloud/integration/integration-sb3.md index 12af81ad..7981b060 100644 --- a/en/guide_cloud/integration/integration-sb3.md +++ b/en/guide_cloud/integration/integration-sb3.md @@ -1,4 +1,4 @@ -# Stable-Baseline3 +# Stable Baselines3 [![](https://swanlab-docs-1301372061.cos.ap-beijing.myqcloud.com/assets/assets/colab.svg)](https://colab.research.google.com/drive/1JfU4oCKCS7FQE_AXqZ3k9Bt1vmK-6pMO?usp=sharing) diff --git a/en/guide_cloud/integration/integration-specforge.md b/en/guide_cloud/integration/integration-specforge.md index adda89f2..c39e998d 100644 --- a/en/guide_cloud/integration/integration-specforge.md +++ b/en/guide_cloud/integration/integration-specforge.md @@ -1,4 +1,4 @@ -# Specforge +# SpecForge [SpecForge](https://github.com/sgl-project/SpecForge) is an ecosystem project developed by the SGLang team. It is a framework for training speculative decoding models, enabling developers to seamlessly integrate them into the SGLang service framework to accelerate inference speed. @@ -8,7 +8,7 @@ You can use SpecForge for rapid model training while employing SwanLab for experiment tracking and visualization. -## Integrating Specforge with SwanLab +## Integrating SpecForge with SwanLab > Reference Documentation: https://docs.sglang.io/SpecForge/basic_usage/training.html#experiment-tracking diff --git a/en/guide_cloud/integration/integration-swift.md b/en/guide_cloud/integration/integration-swift.md index 46b77dc7..252542ad 100644 --- a/en/guide_cloud/integration/integration-swift.md +++ b/en/guide_cloud/integration/integration-swift.md @@ -1,9 +1,9 @@ -# Modelscope Swift +# ModelScope Swift > SwanLab has been officially integrated with Swift, see: [#3142](https://github.com/modelscope/ms-swift/pull/3142) > Online Demo: [swift-robot](https://swanlab.cn/@ZeyiLin/swift-robot/runs/9lc9rmmwm4hh7ay1vkzd7/chart) -[Modelscope](https://modelscope.cn/)'s [Swift](https://github.com/modelscope/swift) is a framework that integrates model training, fine-tuning, inference, and deployment. +[ModelScope](https://modelscope.cn/)'s [Swift](https://github.com/modelscope/swift) is a framework that integrates model training, fine-tuning, inference, and deployment. ![logo](https://swanlab-docs-1301372061.cos.ap-beijing.myqcloud.com/assets/en/guide_cloud/integration/swift/logo.png) diff --git a/en/guide_cloud/integration/integration-tensorboard.md b/en/guide_cloud/integration/integration-tensorboard.md index 9053077b..fb506b6a 100644 --- a/en/guide_cloud/integration/integration-tensorboard.md +++ b/en/guide_cloud/integration/integration-tensorboard.md @@ -1,13 +1,13 @@ -# Tensorboard +# TensorBoard [TensorBoard](https://github.com/tensorflow/tensorboard) is a visualization tool provided by Google TensorFlow, designed to help understand, debug, and optimize machine learning models. It displays various metrics and data during the training process through a graphical interface, allowing developers to intuitively understand the performance and behavior of their models. ![TensorBoard](https://swanlab-docs-1301372061.cos.ap-beijing.myqcloud.com/assets/assets/ig-tensorboard.png) -**You can synchronize projects tracked with Tensorboard to SwanLab in two ways:** +**You can synchronize projects tracked with TensorBoard to SwanLab in two ways:** -- **Synchronous Tracking**: If your current project uses Tensorboard for experiment tracking, you can use the `swanlab.sync_tensorboardX()` or `swanlab.sync_tensorboard_torch()` commands to synchronize metrics to SwanLab while running your training script. -- **Convert Existing Projects**: If you want to copy a project from Tensorboard to SwanLab, you can use `swanlab convert` to convert a directory containing TFevent files into a SwanLab project. +- **Synchronous Tracking**: If your current project uses TensorBoard for experiment tracking, you can use the `swanlab.sync_tensorboardX()` or `swanlab.sync_tensorboard_torch()` commands to synchronize metrics to SwanLab while running your training script. +- **Convert Existing Projects**: If you want to copy a project from TensorBoard to SwanLab, you can use `swanlab convert` to convert a directory containing TFevent files into a SwanLab project. ::: info The current version only supports converting scalar and image charts. @@ -17,9 +17,9 @@ The current version only supports converting scalar and image charts. ## 1. Synchronous Tracking -### 1.1 TensorboardX: Add the `sync_tensorboardX` Command +### 1.1 TensorBoardX: Add the `sync_tensorboardX` Command -If you are using TensorboardX, you can add the `swanlab.sync_tensorboardX()` command anywhere before executing `tensorboardX.SummaryWriter()` to synchronize metrics to SwanLab during training. +If you are using TensorBoardX, you can add the `swanlab.sync_tensorboardX()` command anywhere before executing `tensorboardX.SummaryWriter()` to synchronize metrics to SwanLab during training. ```python import swanlab @@ -32,7 +32,7 @@ writer = SummaryWriter(log_dir='./runs') ### 1.2 PyTorch: Add the `sync_tensorboard_torch` Command -If you are using PyTorch's built-in Tensorboard, you can add the `swanlab.sync_tensorboard_torch()` command anywhere before executing `torch.utils.tensorboard.SummaryWriter()` to synchronize metrics to SwanLab during training. +If you are using PyTorch's built-in TensorBoard, you can add the `swanlab.sync_tensorboard_torch()` command anywhere before executing `torch.utils.tensorboard.SummaryWriter()` to synchronize metrics to SwanLab during training. ```python import swanlab @@ -45,11 +45,11 @@ writer = torch.utils.tensorboard.SummaryWriter(log_dir='./runs') ### 1.3 Alternative Approach -You can also manually initialize SwanLab first and then run the Tensorboard code. +You can also manually initialize SwanLab first and then run the TensorBoard code. ::: code-group -```python [TensorboardX] +```python [TensorBoardX] import swanlab from tensorboardX import SummaryWriter @@ -78,7 +78,7 @@ writer = SummaryWriter(log_dir='./runs') ::: code-group -```python [TensorboardX] +```python [TensorBoardX] import swanlab from tensorboardX import SummaryWriter import random @@ -134,7 +134,7 @@ Use the following command to synchronize tensorboard logs: swanlab convert -t tensorboard --tb_logdir [TFEVENT_LOGDIR] ``` -Here, `[TFEVENT_LOGDIR]` refers to the path of the log files generated when you previously recorded experiments with Tensorboard. +Here, `[TFEVENT_LOGDIR]` refers to the path of the log files generated when you previously recorded experiments with TensorBoard. The SwanLab Converter will automatically detect `tfevent` files in the specified directory and its subdirectories (default depth is 3) and create a SwanLab experiment for each `tfevent` file. @@ -166,7 +166,7 @@ from swanlab.converter import TFBConverter tfb_converter = TFBConverter( convert_dir="./runs", - project="Tensorboard-Converter", + project="TensorBoard-Converter", workspace="SwanLab", logdir="./logs", ) @@ -175,14 +175,14 @@ tfb_converter.run() The equivalent CLI command: ```bash -swanlab convert -t tensorboard --tb_logdir ./runs -p Tensorboard-Converter -w SwanLab -l ./logs +swanlab convert -t tensorboard --tb_logdir ./runs -p TensorBoard-Converter -w SwanLab -l ./logs ``` -Executing the above script will create a project named `Tensorboard-Converter` in the `SwanLab` workspace, convert the `tfevent` files in the `./runs` directory into SwanLab experiments, and save the logs generated by SwanLab in the `./logs` directory. +Executing the above script will create a project named `TensorBoard-Converter` in the `SwanLab` workspace, convert the `tfevent` files in the `./runs` directory into SwanLab experiments, and save the logs generated by SwanLab in the `./logs` directory. ## 3. API Mapping Table -| Function | Tensorboard | SwanLab | +| Function | TensorBoard | SwanLab | | ---- | ---------- | --------------------- | | Create Experiment | writer = SummaryWriter(logdir="./runs") | swanlab.init(logdir="./runs") | | Record Scalar Metrics | writer.add_scalar(key, value, step) | swanlab.log({key, value}, step=step) | diff --git a/en/guide_cloud/integration/integration-wandb.md b/en/guide_cloud/integration/integration-wandb.md index d179e317..7227cb2e 100644 --- a/en/guide_cloud/integration/integration-wandb.md +++ b/en/guide_cloud/integration/integration-wandb.md @@ -1,19 +1,19 @@ # Weights & Biases -Weights & Biases (Wandb) is a platform for experiment tracking, model optimization, and collaboration in machine learning and deep learning projects. W&B provides powerful tools to log and visualize experimental results, helping data scientists and researchers better manage and share their work. +Weights & Biases (W&B) is a platform for experiment tracking, model optimization, and collaboration in machine learning and deep learning projects. W&B provides powerful tools to log and visualize experimental results, helping data scientists and researchers better manage and share their work. ![wandb](https://swanlab-docs-1301372061.cos.ap-beijing.myqcloud.com/assets/assets/ig-wandb.png) :::warning Synchronization Tutorials for Other Tools - [TensorBoard](/guide_cloud/integration/integration-tensorboard.md) -- [MLFlow](/guide_cloud/integration/integration-mlflow.md) +- [MLflow](/guide_cloud/integration/integration-mlflow.md) ::: -**You can sync projects from Wandb to SwanLab in three ways:** +**You can sync projects from W&B to SwanLab in three ways:** 1. **Real-time Syncing**: If your current project uses wandb for experiment tracking, you can use the `swanlab.sync_wandb()` command to simultaneously log metrics to SwanLab while running your training script. -2. **Convert existing projects from the wandb website**: If you want to copy projects from the wandb server (wandb.ai or privately deployed wandb) to SwanLab, you can use `swanlab convert` to transform existing Wandb projects into SwanLab projects. +2. **Convert existing projects from the W&B website**: If you want to copy projects from the wandb server (wandb.ai or privately deployed wandb) to SwanLab, you can use `swanlab convert` to transform existing W&B projects into SwanLab projects. 3. **Convert existing projects from local wandb log files**: If you want to upload local wandb log files to SwanLab, you can use `swanlab convert` to transform local wandb log files into SwanLab projects. ::: info @@ -26,7 +26,7 @@ The current version only supports converting scalar charts. ### 1.1 Add the `sync_wandb` Command -Add the `swanlab.sync_wandb()` command anywhere in your code before `wandb.init()` to synchronize Wandb metrics to SwanLab during training. +Add the `swanlab.sync_wandb()` command anywhere in your code before `wandb.init()` to synchronize W&B metrics to SwanLab during training. ```python import swanlab @@ -45,7 +45,7 @@ With this implementation, `wandb.init()` will simultaneously initialize SwanLab, **`sync_wandb` supports two parameters:** - `mode`: SwanLab logging mode, supporting `cloud`, `local`, and `disabled`. -- `wandb_run`: If set to **False**, data will not be uploaded to Wandb (equivalent to `wandb.init(mode="offline")`). +- `wandb_run`: If set to **False**, data will not be uploaded to W&B (equivalent to `wandb.init(mode="offline")`). ::: @@ -106,7 +106,7 @@ Location of `runid`: ### 2.2 Method 1: Command-Line Conversion -First, ensure you are logged into Wandb and have access to the target project. +First, ensure you are logged into W&B and have access to the target project. Conversion command: @@ -121,9 +121,9 @@ Supported parameters: - `-w`: SwanLab workspace name. - `--mode`: (str) Logging mode (default: `"cloud"`), options: `["cloud", "local", "offline", "disabled"]`. - `-l`: Log directory path. -- `--wb-project`: Wandb project name to convert. -- `--wb-entity`: Wandb entity (username/team) where the project resides. -- `--wb-runid`: Wandb Run ID (specific experiment under the project). +- `--wb-project`: W&B project name to convert. +- `--wb-entity`: W&B entity (username/team) where the project resides. +- `--wb-runid`: W&B Run ID (specific experiment under the project). If `--wb-runid` is omitted, all Runs under the project will be converted. If specified, only the selected Run will be converted. @@ -167,8 +167,8 @@ This achieves the same result as command-line conversion. `WandbConverter.run` parameters: - `wb_project`: Wandb project name. -- `wb_entity`: Wandb entity (username/team). -- `wb_runid`: Wandb Run ID (specific experiment). +- `wb_entity`: W&B entity (username/team). +- `wb_runid`: W&B Run ID (specific experiment). **Asynchronous Conversion (Download Data Locally First, Then Upload to SwanLab)** diff --git a/en/guide_cloud/integration/integration-xtuner.md b/en/guide_cloud/integration/integration-xtuner.md index d136aa34..2f0dd512 100644 --- a/en/guide_cloud/integration/integration-xtuner.md +++ b/en/guide_cloud/integration/integration-xtuner.md @@ -1,4 +1,4 @@ -# Xtuner +# XTuner [XTuner](https://github.com/InternLM/xtuner) is a highly efficient, flexible, and versatile tool library for fine-tuning large models. @@ -6,15 +6,15 @@ -Xtuner supports adaptation with multiple open-source large models such as InternLM and Llama, and can perform tasks such as incremental pre-training, instruction fine-tuning, and tool-based instruction fine-tuning. In terms of hardware requirements, developers can train with the lowest consumer-grade graphics cards, such as Tesla T4 and A100, to achieve specific demand capabilities of large models. +XTuner supports adaptation with multiple open-source large models such as InternLM and Llama, and can perform tasks such as incremental pre-training, instruction fine-tuning, and tool-based instruction fine-tuning. In terms of hardware requirements, developers can train with the lowest consumer-grade graphics cards, such as Tesla T4 and A100, to achieve specific demand capabilities of large models.
-Xtuner supports online tracking using SwanLab through MMEngine. By adding a few lines of code to the configuration file, you can track and visualize metrics such as loss and memory usage. +XTuner supports online tracking using SwanLab through MMEngine. By adding a few lines of code to the configuration file, you can track and visualize metrics such as loss and memory usage. -## Visualizing and Tracking Xtuner Fine-Tuning Progress with SwanLab +## Visualizing and Tracking XTuner Fine-Tuning Progress with SwanLab Open the configuration file you want to train (for example, [qwen1_5_7b_chat_full_alpaca_e3.py](https://github.com/InternLM/xtuner/blob/main/xtuner/configs/qwen/qwen1_5/qwen1_5_7b_chat/qwen1_5_7b_chat_full_alpaca_e3.py)), find the `visualizer` parameter, and replace it with: diff --git a/en/index.md b/en/index.md index 9059b589..756d90c1 100644 --- a/en/index.md +++ b/en/index.md @@ -187,10 +187,10 @@ features: -
-
-

◎ Emotion Machine (Beijing) Technology Co., Ltd.

- +
+
+

◎ Emotion Machine (Beijing) Technology Co., Ltd.

+
diff --git a/zh/index.md b/zh/index.md index 4bef6abc..e2a7663e 100644 --- a/zh/index.md +++ b/zh/index.md @@ -359,11 +359,10 @@ features: -
-
-

◎ 情感机器（北京）科技有限公司

-

京公网安备11010502056567号 · 京ICP备2024101706号-1

- +
+
+

◎ 情感机器（北京）科技有限公司 · 京公网安备11010502056567号 · 京ICP备2024101706号-1

+