Progressive-Multi-stage-Feature-Mix-for-Person-Re-Identification

pytorch code for paper Progressive Multi-stage Feature Mix for Person Re-Identification: https://arxiv.org/abs/2007.08779

This project is based on batch-drop-block: https://github.com/daizuozhuo/batch-dropblock-network

The proposed PMM(Progressive-Multi-stage-Feature-Mix) model can be found in models/progressive_networks.py

Setup running environment

This project requires python3, cython, torch, torchvision, scikit-learn, tensorboardX, fire.

Prepare dataset

Create a directory to store reid datasets under this repo via
```bash
cd reid
mkdir data
```

For market1501 dataset, 
1. Download Market1501 dataset to `data/` from http://www.liangzheng.org/Project/project_reid.html
2. Extract dataset and rename to `market1501`. The data structure would like:
```
market1501/
    bounding_box_test/
    bounding_box_train/
    query/
```

For CUHK03 dataset,
1. Download CUHK03-NP dataset from https://github.com/zhunzhong07/person-re-ranking/tree/master/CUHK03-NP 
2. Extract dataset and rename folers inside it to cuhk-detect and cuhk-label.
For DukeMTMC-reID dataset,
Dowload from https://github.com/layumi/DukeMTMC-reID_evaluation

Training

Traning Market1501

python main_reid.py train --save_dir='./pytorch-ckpt/market-bfe' --max_epoch=400 --eval_step=30 --dataset=market1501 --test_batch=128 --train_batch=128 --optim=adam --adjust_lr