This repository provides the MS-COCO training code for the Visual Transformers with Primal Object Queries for Multi-Label Image Classification paper which will be published in ICPR2022.
- python 3.7
- pytorch 1.6.0
python train.py -image_path <image_path> -save_path <save_path> -mix_up
python train.py -image_path <image_path> -snapshot -test_model
Transformer encoder-decoder models in this repository are based on the implementation in here.