Is it possible to change activation functions all at once?

Asked Jul 17 '23 at 18:38

Active Jul 18 '23 at 09:33

Viewed 17 times

I'd like to load a pretrained Inception-V3 model. I have two question.

How can I access to activation functions? Is it possible to change all Relu activation functions to tanh all at once?

import torch
model = torch.hub.load("pytorch/vision", "inception_v3", weights="IMAGENET1K_V1")

edited Jul 18 '23 at 09:33

asked Jul 17 '23 at 18:38

linkho

0 Answers0