Hi, I try to get the per layer latency of the pytorch model.
For this, I use the profile output of the TVM debugger. The problem is that I need to map the layers of the original network with the layers of the debugger. Is there a way to map the layers?