bityangke/semi-reid

Self-Paced Co-training Person-ReID

Pytorch implementaion based on Open-Reid

Run

bash spaco.sh

alternatively, you can run with

python examples/spaco.py --arch1 resnet50 --arch2 densenet121 --iter-step 5 --dataset market1501std --logs-dir logs/

Please cite spaco in your publications if it helps your research:
@inproceedings{ma2017self,
title={Self-Paced Co-training},
author={Ma, Fan and Meng, Deyu and Xie, Qi and Li, Zina and Dong, Xuanyi},
booktitle={International Conference on Machine Learning},
pages={2275--2284},
year={2017}
}

Open-ReID

Open-ReID is a lightweight library of person re-identification for research purpose. It aims to provide a uniform interface for different datasets, a full set of models and evaluation metrics, as well as examples to reproduce (near) state-of-the-art results.

Installation

Install PyTorch (version >= 0.2.0). Although we support both python2 and python3, we recommend python3 for better performance.

git clone https://github.com/Cysu/open-reid.git
cd open-reid
python setup.py install

Examples

python examples/softmax_loss.py -d viper -b 64 -j 2 -a resnet50 --logs-dir logs/softmax-loss/viper-resnet50

This is just a quick example. VIPeR dataset may not be large enough to train a deep neural network.

Check about more examples and benchmarks.