PatchKD

Code for ACM MM 2022 paper Patch-based Knowledge Distillation for Lifelong Person Re-Identification.

Installation. I recommend using conda environment for creating environment and installing the packages.

pip install -r requirements.txt

Install Pytorch (Prefferably with CUDA). Example of installing with conda:

conda install pytorch torchvision torchaudio pytorch-cuda=12.4 -c pytorch -c nvidia

Please follow Torchreid_Datasets_Doc to download datasets and unzip them to your data path (we refer to 'machine_dataset_path' in train_test.py). Alternatively, you could download some datasets from light-reid and DualNorm.

Quick Start

Training + evaluation on Market1501 dataset. Make sure the visdom server is listening.

python train_test.py

Evaluation from checkpoint:

python train_test.py --mode test --resume_test_model /path/to/pretrained/model

Visualization from checkpoint:

python train_test.py --mode visualize --resume_visualize_model /path/to/pretrained/model

Citation

@inproceedings{sun2022patch,
    author = {Sun, Zhicheng and Mu, Yadong},
    title = {Patch-based Knowledge Distillation for Lifelong Person Re-Identification},
    booktitle = {Proceedings of the 30th ACM International Conference on Multimedia},
    pages = {696--707},
    year = {2022}
}

Acknowledgement

Code is based on the implementation of PatchKD.