Inceptionv3 failing to work with different Loss functions

Hello,
I am trying to retrain the Inception v3 on some sign language alphabets, but am running into an issue when I train it:

The issue appears to be with inception v3 not reconizing log_softmax.
I tried using other loss functions such as NLLLoss, but that gives a different error .
I tried BCEWithLogitsLoss and MSELoss and got this issue:
Attribute Error: InceptionOutpurs object has no arrtibute size

What loss function should I use that will help me avoid these issues? Or is there a problem with inception i should focus on resolving?

Have a look at this post, which explains how to use the InceptionOutputs.

PS: you can post code snippets by wrapping them into three backticks ```, which makes debugging easier. :wink:

1 Like