search

tonylins / pytorch-mobilenet-v2

A PyTorch implementation of MobileNet V2 architecture and pretrained model.
Apache License 2.0
1.37k stars 328 forks source link

calculate flops #37

Closed young-fire closed 4 years ago

young-fire commented 4 years ago

bn layer is not involved when calulate flops?

tonylins commented 4 years ago

Hi, during deployment, the BN layer is usually fused into convolutions. Therefore, we do not need to consider the computation.

Even without fusing, the computation of BN is negligible.