Discover Cross-Modality Nuances for Visible-Infrared Person Re-Identification

[paper]

This repository is Pytorch code for our proposed joint Modality and Pattern Alignment Network (MPANet)

Environmental requirements:

PyTorch == 1.2.0

ignite == 0.2.1

torchvision == 0.4.0

apex == 0.1

Quick start

1. Clone this repository:

git clone https://github.com/MPANet/MPANet.git

1. Modify the path to datasets:

The path to datasets can be modified in the following file:

./configs/default/dataset.py

1. Training:

To train the model, you can use following command:

SYSU-MM01:

python train.py --cfg ./configs/SYSU.yml

RegDB:

python train.py --cfg ./configs/RegDB.yml

trained model:

The checkpoint can be found here:
https://pan.baidu.com/s/1TnjtfMFPnm5TEprgAhqz9A

Code: rfti

Reference:

LuckyDC/RGB-IR-ReID-Baseline