Example for EncNet:
CUDA_VISIBLE_DEVICES=1,2,3 python3 train_fold.py --model encnet --backbone resnet50 --imagelist_path <path_to_imagelist> --image_path <path_to_images> --masks_path <path_to_masks> --lr=0.001 --wd=0.0 --epochs 50 --batch-size 24 --size=512 --se-loss --aux --ckpt_name /tmp/tmp.pth --use-dice --logger-dir /tmp | tee /tmp/log
Example for Unet:
CUDA_VISIBLE_DEVICES=1,2,3 python3 train_fold.py --model unet --backbone seresnext50 --imagelist_path <path_to_imagelist> --image_path <path_to_images> --masks_path <path_to_masks> --lr=0.1 --wd=0.0 --epochs 50 --batch-size 24 --size=512 --ckpt_name /tmp/tmp.pth --use-dice --logger-dir /tmp | tee /tmp/log
Requirements: pytorch-encoding