paper : https://arxiv.org/abs/1905.00641
review : https://khyeyoon.github.io/face%20detection%20paper/RetinaFace/
- This code was written by modifying the code in the link below
- Python3
- PyTorch >= 1.1.0
- Torchvision >= 0.3.0
- NVIDIA GPU + CUDA
Use it by modifying it appropriately for your path
CUDA_VISIBLE_DEVICES=0,1,2,3 python train.py --network resnet50 or
CUDA_VISIBLE_DEVICES=0 python train.py --network mobile0.25
Use it by modifying it appropriately for your path
- single image
python evaluation-singleimg.py --trained_model weight_file --network mobile0.25 or resnet50
- multiple image
evaluation-multipleimg.py --trained_model weight_file --network mobile0.25 or resnet50