Skip to content

Add approximate gelu replacement to opt level 2 #10129

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
Apr 14, 2025

Conversation

mcremon-meta
Copy link
Contributor

Summary:
As titled. Gelu is prohibitively expensive to run on DSPs, due to the std::erf call in the function. The PT approximate version using an approximation based on tanh, which is faster on the ASR encoder 27M model for example.

Seems like BUCK files (even with just on_call commands, the linter is complaining).

Differential Revision: D72935935

@mcremon-meta mcremon-meta requested a review from tarun292 as a code owner April 14, 2025 02:18
Copy link

pytorch-bot bot commented Apr 14, 2025

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/10129

Note: Links to docs will display an error until the docs builds have been completed.

❌ 1 New Failure

As of commit a23d0c1 with merge base 4022ff1 (image):

NEW FAILURE - The following job has failed:

This comment was automatically generated by Dr. CI and updates every 15 minutes.

@facebook-github-bot facebook-github-bot added CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. fb-exported labels Apr 14, 2025
@facebook-github-bot
Copy link
Contributor

This pull request was exported from Phabricator. Differential Revision: D72935935

mcremon-meta added a commit that referenced this pull request Apr 14, 2025
Summary:

As titled. Gelu is prohibitively expensive to run on DSPs, due to the std::erf call in the function. The PT approximate version using an approximation based on `tanh`, which is  faster on the ASR encoder 27M model for example.

Seems like BUCK files (even with just on_call commands, the linter is complaining).

Differential Revision: D72935935
@facebook-github-bot
Copy link
Contributor

This pull request was exported from Phabricator. Differential Revision: D72935935

mcremon-meta added a commit that referenced this pull request Apr 14, 2025
Summary:
Pull Request resolved: #10129

As titled. Gelu is prohibitively expensive to run on DSPs, due to the std::erf call in the function. The PT approximate version using an approximation based on `tanh`, which is  faster on the ASR encoder 27M model for example.

Seems like BUCK files (even with just on_call commands, the linter is complaining).

Differential Revision: D72935935
mcremon-meta added a commit that referenced this pull request Apr 14, 2025
Summary:

As titled. Gelu is prohibitively expensive to run on DSPs, due to the std::erf call in the function. The PT approximate version using an approximation based on `tanh`, which is  faster on the ASR encoder 27M model for example.

Seems like BUCK files (even with just on_call commands, the linter is complaining).

Differential Revision: D72935935
Summary:
Pull Request resolved: #10129

As titled. Gelu is prohibitively expensive to run on DSPs, due to the std::erf call in the function. The PT approximate version using an approximation based on `tanh`, which is  faster on the ASR encoder 27M model for example.

Seems like BUCK files (even with just on_call commands, the linter is complaining).

Differential Revision: D72935935
@facebook-github-bot
Copy link
Contributor

This pull request was exported from Phabricator. Differential Revision: D72935935

@facebook-github-bot facebook-github-bot merged commit c178637 into main Apr 14, 2025
81 of 84 checks passed
@facebook-github-bot facebook-github-bot deleted the export-D72935935 branch April 14, 2025 23:36
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. fb-exported topic: not user facing
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants