Closed kamiLight closed 2 months ago
ptflops can take into account only ops that are derived from nn.Module. Support of torch.bmm and similar operations requires complete redesign of the flops counting mechanism. I've started thinking of that but barely will do this soon.
Supported by both pytorch and aten backend, see https://github.com/sovrasov/flops-counter.pytorch/blob/master/ptflops/pytorch_ops.py#L443
One part of my model performs the matrix multiplication, i.e torch.bmm() / torch.matmul(), but results show the GMac of this part is 0.0, i want to know if this is because the GMac is too small or it is not support torch.bmm() or torch.matmul()?