Tutorial Video link
Code based on End-to-end Object Detection with Transformer
Download wider face dataset link and unzip the files in dataset folder
Move to the dataset folder, and convert downloaded wider face dataset into COCO format
$ python face_to_coco.py
Move to detr folder and run main.py
$ python -m torch.distributed.launch --nproc_per_node=[num_of_gpus] --use_env main.py --data_path ../dataset/
Run test.py
$ python test.py --data_path ../dataset/WIDER_test/images/ --resume [path_to_checkpoint.pth]