deep-high-resolution-net.pytorch icon indicating copy to clipboard operation
deep-high-resolution-net.pytorch copied to clipboard

Published 20 hours ago • verified publisher icon leoxiaobin

Bugs in `get_model_summary' and flops calculation

Open jin-s13 opened this issue 4 years ago • 1 comments

Dear all,

Thanks for releasing the codes! I noticed some problems in model flops calculation.

  1. The classname Conv2d and ConvTranspose2d both contain 'Conv', so their flops are counted in get_model_summary. https://github.com/leoxiaobin/deep-high-resolution-net.pytorch/blob/ba50a82dce412df97f088c572d86d7977753bf74/lib/utils/utils.py#L124

However, the flops-calculation for Conv2d and ConvTranspose2d should be different.

For Conv2d, it is

flops = (torch.prod(torch.LongTensor(list(module.weight.data.size()))) * torch.prod(torch.LongTensor(list(output.size())[2:]))).item()

But for ConvTranspose2d, it should be

flops = (torch.prod(torch.LongTensor(list(module.weight.data.size()))) * torch.prod(torch.LongTensor(list(input[0].size())[2:]))).item()

  1. Flops of many other ops (e.g. BN, ReLU) are not calculated.

jin-s13 avatar Jun 19 '21 15:06 jin-s13

Yes, indeed! @leoxiaobin @sunke123

zengwang430521 avatar Apr 12 '22 13:04 zengwang430521