# pytorch-toolkit-dev **Repository Path**: gaoshanwen/pytorch-toolkit-dev ## Basic Information - **Project Name**: pytorch-toolkit-dev - **Description**: building a multitask training platform based on Pytorch - **Primary Language**: Python - **License**: Apache-2.0 - **Default Branch**: main - **Homepage**: None - **GVP Project**: No ## Statistics - **Stars**: 2 - **Forks**: 0 - **Created**: 2023-12-04 - **Last Updated**: 2025-08-30 ## Categories & Tags **Categories**: Uncategorized **Tags**: timm, Python, PyTorch ## README # pytorch-toolkit-dev building a multitask training platform based on Pytorch ## Introduction [![](https://img.shields.io/badge/Python-3.8-3776AB?style=plastic&logo=python&logoColor=FFFFFF)](https://www.python.org) [![](https://img.shields.io/github/v/release/pytorch/pytorch?style=plastic&logo=pytorch&logoColor=FFFFFF&label=PyTorch&color=EE4C2C)](https://pytorch.org) [![](https://img.shields.io/github/v/release/huggingface/pytorch-image-models.svg?style=plastic&logo=huggingface&logoColor=FFFFFF&label=Timm&color=FFD21E)](https://github.com/huggingface/pytorch-image-models/tree/main) [![](https://img.shields.io/badge/ONNX-1.14.1-005CED?style=plastic&logo=ONNX&logoColor=FFFFFF&label=ONNX)](https://onnx.ai) [![](https://img.shields.io/github/v/release/PyMySQL/PyMySQL.svg?style=plastic&logo=mysql&logoColor=FFFFFF&label=PyMySQL&color=FFBE00)](https://zetcode.com/python/pymysql/) cls-dev branch for classfication or re-identification tasks.
### Supported Functions - [x] Classfication (Single-label, Multi-label) and Re-identification tasks. - [ ] Metric learning for Re-identification. - [x] Visualize training curve (wandb / tensorboard). - [x] Remove similarity or noise data, weighted k-nearest neighbor for reid. - [ ] Visualize Precision-Recall / Receiver-Operating-Characteristic curve. - [x] Mixed-Precision Training for faster speed. - [x] Visualize models' heatmaps and U-MAPs. - [x] Convert pth model to onnx / rknn / ncnn format. ### Major Motivation 1.👀 Use the State-of-the-Art image classfication toolkit. + Various backbones and pretrained models + Bag of training tricks + Large-scale training configs + High efficiency and extensibility + Powerful toolkits 2.🚀 Enhance codes' reusability. 3.🛠️ Minimize our project. ## Quick Start ### Install Environment For detailed installation guides, please refer to [INSTALL.md](docs/INSTALL.md). ### Command Guides + Train & Validate with Training Curve Refer to [README-Train/Val Models](tools/README.md) for details. + Convert Pth Model To ONNX/RKNN Refer to [README-Convert Models](tools/deploy/README.md) for details. ## License This project is released under the [Apache 2.0 license](LICENSE). ## Contribute Feel free to create a pull request if you want to contribute (e.g. networks or tricks).