94bafa8
1
2
export CUDA_VISIBLE_DEVICES=4,5 torchrun --nnodes=1 --nproc_per_node=2 --master_port=29509 train.py --config ./configs/ucf101/ucf101_train.yaml
