cifar100 reproducing problem

Open YeonsungJung opened this issue 5 years ago • 1 comments

Hi I found a reproducing problem in cifar100&resnet50 with your parameters. Accuracy with batch size 1024 is less than it of 512.

Simclr: 70.5 | 68.84 (512, 1024) SupCon: 76.51 | 75.27 (512, 1024)

Is there something more to reproduce..?

Sep 30 '20 11:09 YeonsungJung

Hi, @Ys-Jung77 ,

I don't know the difference between your and my settings, but the numbers reported in README are with batch size of 1024. E.g., SimCLR get 70.7 while SupCon get 76.5, with bsz=1204.

Oct 01 '20 22:10 HobbitLong