Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Accuracy = 0 #6

Open
robd2 opened this issue Sep 13, 2018 · 2 comments
Open

Accuracy = 0 #6

robd2 opened this issue Sep 13, 2018 · 2 comments

Comments

@robd2
Copy link
Contributor

robd2 commented Sep 13, 2018

Thank you so much @giannisnik for publishing this code. Is a very interesting use of a CNN. I'm trying to reproduce some of your results but keep getting Accuracy = 0 measurements. Any thoughts on where to focus my debugging? Maybe related to the missing "best model"?

image
@robd2
Copy link
Contributor Author

robd2 commented Sep 14, 2018 •
edited
Loading

just FYI, I think I've narrowed it down to the logic in the lambda function:
evaluation = lambda output, target: torch.sum(output.eq(target)) / target.size()[0]

Seems to be an issue with floor division if both operands have integer types. Maybe due to some recent PyTorch updates. Pull request submitted for proposed fix.

robd2 added a commit to robd2/cnn-graph-classification that referenced this issue Sep 14, 2018
@robd2
fix for giannisnik#6
e77488f 
Accuracy = 0 as a result of recent PyTorch changes. See here for info: pytorch/pytorch#5411
robd2 added a commit to robd2/cnn-graph-classification that referenced this issue Sep 14, 2018
@robd2
Revert "fix for giannisnik#6"
18deed2 
This reverts commit e77488f.
robd2 added a commit to robd2/cnn-graph-classification that referenced this issue Sep 14, 2018
@robd2
fix for giannisnik#6
05acfc8 
Accuracy = 0 as a result of recent PyTorch changes. See here for info: pytorch/pytorch#5411
@robd2 robd2 mentioned this issue Sep 14, 2018
Merged
@Wangshuaiia
Copy link

thank you for your issue@robd2. I have met this problem just now.

giannisnik added a commit that referenced this issue Feb 15, 2019
@giannisnik
Merge pull request #7 from robd2/master
740c682 
fix for #6 Accuracy = 0 as a result of recent PyTorch changes. See here for info: pytorch/pytorch#5411
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
2 participants
@robd2 @Wangshuaiia