hkust-vgd/shellnet

FLOPs Problems

Hlxwk opened this issue · 1 comments

Hlxwk commented

Thanks for your sharing great work about efficient point cloud processing. I am curious about the way of computing FLOPs, especially for inference FLOPs. Looking forward to your reply.

Please refer to tf.profiler.profile to show the FLOPs. For inference, there are no ops for back propagation. So the FLOPs for inference are much less than those of training.