search

ltkong218 / IFRNet

IFRNet: Intermediate Feature Refine Network for Efficient Frame Interpolation (CVPR 2022)
MIT License
259 stars 23 forks source link

How to train using a single GPU？ #36

Closed Brightlcz closed 2 months ago

Brightlcz commented 9 months ago

Hello author, thank you for your excellent work.

Your code seems to be for multi-card distributed training. I only have one GPU. Could you please tell me where to change it for single-card training, so that it can be debugged with a single card.

Looking forward to your reply.

ltkong218 commented 9 months ago

Thanks for your interest!

For training on a machine equipped with single GPU, please try to run the following command:

$ python -m torch.distributed.launch --nproc_per_node=1 train_vimeo90k.py --world_size 1 --model_name 'IFRNet' --epochs 300 --batch_size 6 --lr_start 1e-4 --lr_end 1e-5

You can increase batch size if your GPU memory is enough.