Skip to content

fix nn.MHA scriptability (#58727) #59072

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.

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

Merged
merged 1 commit into from
May 27, 2021
Merged

Conversation

bhosmer
Copy link

@bhosmer bhosmer commented May 27, 2021

#58727 fixes a post-1.8 regression which causes a hard-to-troubleshoot TorchScript type error when scripting an incorrectly quantized transformer layer. It's a simple short-term fix, just reinstating the naming difference that avoided the error previously.

Summary: Pull Request resolved: pytorch#58727

Test Plan: Imported from OSS

Reviewed By: ngimel

Differential Revision: D28593830

Pulled By: bhosmer

fbshipit-source-id: 37dee9efededaea9985a2bf040df1ba4b46f6580
@facebook-github-bot
Copy link
Contributor

facebook-github-bot commented May 27, 2021

💊 CI failures summary and remediations

As of commit b2a101d (more details on the Dr. CI page):



🕵️ 1 new failure recognized by patterns

The following CI failures do not appear to be due to upstream breakages:

See GitHub Actions build Linux CI (pytorch-linux-xenial-py3.6-gcc5.4) / test (1/1)

Step: "Test PyTorch" (full log | diagnosis details | 🔁 rerun)

2021-05-27T08:29:39.7241082Z Build left local git repository checkout dirty
2021-05-27T08:29:39.7235773Z  M third_party/nccl/nccl
2021-05-27T08:29:39.7236431Z  M third_party/tensorpipe
2021-05-27T08:29:39.7237190Z ?? third_party/cudnn_frontend/'
2021-05-27T08:29:39.7238013Z + [[ -n  M third_party/fbgemm
2021-05-27T08:29:39.7238448Z  M third_party/gloo
2021-05-27T08:29:39.7238796Z  M third_party/kineto
2021-05-27T08:29:39.7239139Z  M third_party/nccl/nccl
2021-05-27T08:29:39.7239529Z  M third_party/tensorpipe
2021-05-27T08:29:39.7239925Z ?? third_party/cudnn_frontend/ ]]
2021-05-27T08:29:39.7240533Z + echo 'Build left local git repository checkout dirty'
2021-05-27T08:29:39.7241082Z Build left local git repository checkout dirty
2021-05-27T08:29:39.7241645Z + echo 'git status --porcelain:'
2021-05-27T08:29:39.7242330Z git status --porcelain:
2021-05-27T08:29:39.7243028Z + echo ' M third_party/fbgemm
2021-05-27T08:29:39.7243390Z  M third_party/gloo
2021-05-27T08:29:39.7243738Z  M third_party/kineto
2021-05-27T08:29:39.7244082Z  M third_party/nccl/nccl
2021-05-27T08:29:39.7244468Z  M third_party/tensorpipe
2021-05-27T08:29:39.7244969Z ?? third_party/cudnn_frontend/'
2021-05-27T08:29:39.7245337Z  M third_party/fbgemm
2021-05-27T08:29:39.7245682Z  M third_party/gloo

🚧 1 fixed upstream failure:

These were probably caused by upstream breakages that were already fixed.

Please rebase on the viable/strict branch (expand for instructions)

If your commit is older than viable/strict, run these commands:

git fetch https://github.com/pytorch/pytorch viable/strict
git rebase FETCH_HEAD

This comment was automatically generated by Dr. CI (expand for details).Follow this link to opt-out of these comments for your Pull Requests.

Please report bugs/suggestions to the (internal) Dr. CI Users group.

Click here to manually regenerate this comment.

@driazati driazati merged commit 0896c6b into pytorch:release/1.9 May 27, 2021
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants