Text Image Matching

Data preparation

Download the CUHK-PEDES dataset from here

Organize them in your dataset root dir folder as follows:

|-- your dataset root dir/
|   |-- <CUHK-PEDES>/
|       |-- imgs
|            |-- cam_a
|            |-- cam_b
|            |-- ...
|       |-- reid_raw.json

Training

bash run/clip_multiview.sh

train code in: tools/train