pytorch-lars
pytorch-lars copied to clipboard
"Layer-wise Adaptive Rate Scaling" in PyTorch
Results
3
pytorch-lars issues
Sort by
recently updated
recently updated
newest added
Hi! I have come into a `ZeroDivisionError: float division by zero` error when using your LARS code, which is reported to happen at line 87 of `lars.py`. To fix this,...
Thanks for your sharing. I've noticed that the generalization gap between small batch size and large batch size is not the same as described as in the original paper. I'm...