#205205th of 206 in ML frameworks and libraries
deep-person-reid
PyTorch library for training and evaluating person re-identification models
- Stars
- 4.9k
- License
- MIT
- Last commit
- Jan 2026
- Last release
- Oct 2019
- Language
- Python
4,793 on 22 Apr 20264,944 on 11 Nov 2026
From GitHub's daily star counts.
Checked on GitHub.
Overview
Torchreid is a PyTorch library for person re-identification, covering both image and video re-ID. It provides data managers for datasets such as Market1501, CUHK03, MSMT17 and MARS, plus implementations of ReID models including OSNet, OSNet-AIN, PCB, HACNN and MLFN, and ImageNet backbones like ResNet. Training runs from Python code or a config-driven script, with multi-GPU, cross-dataset evaluation and export to ONNX, OpenVINO and TFLite.
Who it is for: Computer vision researchers and engineers building person re-ID models
Strengths
- Many standard re-ID datasets, image and video, with multi-dataset training and cross-dataset evaluation
- Includes OSNet and OSNet-AIN plus other ReID and ImageNet backbones, with a pretrained model zoo
- Feature extractor API and export script for ONNX, OpenVINO and TFLite
- YAML configs and scripts/main.py give a reproducible train and test interface
Weaknesses
- Last release is v1.4.0 from 2019-10-23 per tracker data; README install targets Python 3.7
- Install guide pins cudatoolkit=9.0 with conda, which is outdated for current GPUs
- Research-oriented: no inference server or tracking pipeline, only training, evaluation and feature extraction
- Datasets must be downloaded and prepared separately; some model weights are hosted on Google Drive
What it needs
- GPU optional
- Docker
- Needs PyTorch, torchvision, conda, Cython
- Models: OSNet, OSNet-AIN, ResNet, PCB, HACNN
What changed
From the release notes. No flag means the note does not say so, not that the upgrade is safe.
Also in ML frameworks and libraries
See all 206| Rank | Project | Score |
|---|---|---|
| 1 | transformersPython library defining open models for text, vision, audio and multimodal tasks | 94 out of 100 |
| 2 | ultralyticsPython library and CLI for training and running YOLO vision models | 94 out of 100 |
| 3 | pytorchPython tensor library with GPU support and tape-based autograd | 93 out of 100 |
| #4 | pytorch-image-modelsPyTorch library of image models, pretrained weights and training scripts | 93 out of 100 |
| #5 | supervisionPython toolkit for loading, annotating and counting computer vision detections | 92 out of 100 |