Skip to content

Conversation

@t-vi
Copy link
Collaborator

@t-vi t-vi commented Oct 7, 2019

Fixes: #27442

Thank you Mohamed Yousef (@ASDen) for the report with minimal
reproducing example and detailed analysis!

Fixes: pytorch#27442

Thank you Mohamed Yousef (@ASDen) for the report with minimal
reproducing example and detailed analysis!
@pytorchbot pytorchbot added module: cuda Related to torch.cuda, and CUDA support in general module: nn Related to torch.nn module: operators labels Oct 7, 2019
@t-vi t-vi mentioned this pull request Oct 7, 2019
Copy link
Contributor

@facebook-github-bot facebook-github-bot left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@soumith is landing this pull request. If you are a Facebook employee, you can view this diff on Phabricator.

@t-vi
Copy link
Collaborator Author

t-vi commented Oct 9, 2019

@soumith Something seems to have gone wrong with landing this(?).

zdevito pushed a commit to zdevito/ATen that referenced this pull request Oct 10, 2019
…460)

Summary:
Fixes: pytorch/pytorch#27442

Thank you Mohamed Yousef (ASDen) for the report with minimal
reproducing example and detailed analysis!
Pull Request resolved: pytorch/pytorch#27460

Differential Revision: D17789378

Pulled By: soumith

fbshipit-source-id: dc01a31b998cced4462e933d4b32e09b331f7e41
@facebook-github-bot
Copy link
Contributor

@soumith merged this pull request in e66e00c.

thiagocrepaldi pushed a commit to thiagocrepaldi/pytorch that referenced this pull request Feb 4, 2020
…orch#27460)

Summary:
Fixes: pytorch#27442

Thank you Mohamed Yousef (ASDen) for the report with minimal
reproducing example and detailed analysis!
Pull Request resolved: pytorch#27460

Differential Revision: D17789378

Pulled By: soumith

fbshipit-source-id: dc01a31b998cced4462e933d4b32e09b331f7e41
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

Merged module: cuda Related to torch.cuda, and CUDA support in general module: nn Related to torch.nn

Projects

None yet

Development

Successfully merging this pull request may close these issues.

CTCLoss cuda backend computes wrong gradient when target (i.e. label) length is greater than 896 for double inputs or 1024 for float inputs

5 participants