search

facebookresearch / ConvNeXt

Code release for ConvNeXt model
MIT License
5.79k stars 696 forks source link

Why do you need global sync? #106

Open collinmccarthy opened 2 years ago

collinmccarthy commented 2 years ago

Hi,

Why do you need to call torch.cuda.synchronize() during train_one_epoch?

Thanks, -Collin