Auto Label Using TinaFace

---

Docker.

docker pull itsthanhtung/tina-face:2

Data folder

• data
  • data/train2017
  • data/out_annotation

train2017: image folder

out_annotation: result folder

Mount data directory in local to container directory: /workspace/vedadet/data

Example:

docker run -it \
		--gpus all \
		--name tina \
		--mount type=bind,source="home/tung/data"/target,target=/workspace/vedadet/data \
  		itsthanhtung/tina-face:2

Run

CUDA_VISIBLE_DEVICES="0" python tools/infer.py --config ./configs/infer/tinaface/tinaface_r50_fpn_gn_dcn.py --path data/train2017 --out_path data/out_annotation

--out_img_path(optional) for result visualization