Open chenyuanTW opened 1 year ago
Hi, thanks you for the great work! I retrain for high resolution with
python train.py --model dior \ --name $NAME --dataroot $DATAROOT \ --batch_size 8 --lr 1e-4 --init_type orthogonal \ --loss_coe_seg 0 \ --netG $NET_G --ngf $NGF \ --netD gfla --ndf 32 --n_layers_D 4 \ --n_epochs 22 --n_epochs_decay 0 --lr_update_unit 4 \ --print_freq 20 --display_freq 10 --save_epoch_freq 10 --save_latest_freq 2 \ --n_cpus 8 --gpu_ids 0 \ --flownet_path $PRETRAINED_FLOWNET_PATH --frozen_flownet \ --crop_size 512 --gpu_ids 0,1,2 --random_rate 0 --warmup --perturb
then it show that RuntimeError: CUDA out of memory. ......
but there are 3 RTX 2080 ti, and I only saw the first gpu was used.
Hi, thanks you for the great work! I retrain for high resolution with
python train.py --model dior \ --name $NAME --dataroot $DATAROOT \ --batch_size 8 --lr 1e-4 --init_type orthogonal \ --loss_coe_seg 0 \ --netG $NET_G --ngf $NGF \ --netD gfla --ndf 32 --n_layers_D 4 \ --n_epochs 22 --n_epochs_decay 0 --lr_update_unit 4 \ --print_freq 20 --display_freq 10 --save_epoch_freq 10 --save_latest_freq 2 \ --n_cpus 8 --gpu_ids 0 \ --flownet_path $PRETRAINED_FLOWNET_PATH --frozen_flownet \ --crop_size 512 --gpu_ids 0,1,2 --random_rate 0 --warmup --perturb
then it show that RuntimeError: CUDA out of memory. ......
but there are 3 RTX 2080 ti, and I only saw the first gpu was used.