GSCNN icon indicating copy to clipboard operation
GSCNN copied to clipboard

convTri function in custom_functional.py doesn't do anything

Open kampelmuehler opened this issue 5 years ago • 5 comments

There is a return before actual execution of the function code, is this expected bahavior?

https://github.com/nv-tlabs/GSCNN/blob/3648b86d822f3e73ab1d8e827cf47466bf8f01a7/my_functionals/custom_functional.py#L93

kampelmuehler avatar Nov 28 '19 11:11 kampelmuehler

You can just remove this "return input". And there are a few other "return input, input", remove them as well.

SHMCU avatar Nov 29 '19 00:11 SHMCU

I realized it's at some other points as well, just wanted to let other people know in case they're just executing rather than reading the code

kampelmuehler avatar Nov 29 '19 09:11 kampelmuehler

I realized it's at some other points as well, just wanted to let other people know in case they're just executing rather than reading the code

Hello, when I deleted "return input "and"return input,input", a mistake was reported here:"RuntimeError: Expected object of backend CUDA but got backend CPU for argument #2 'weight'".No error if not deleted.Can you help me with that?Thank you!!!!!!

ayinaaaaaa avatar May 16 '20 08:05 ayinaaaaaa

@ayinaaaaaa I have met the same question, how do you solve this problem?

XGBoost avatar Jul 16 '21 10:07 XGBoost

Did you train your model with these returns or not?

You can just remove this "return input". And there are a few other "return input, input", remove them as well.

Nagapetyan avatar Jan 30 '23 17:01 Nagapetyan