Add torch.no_grad() guard on export #2373
Closed
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Summary: For some models where the grad is enabled (like LLava model), we need to disable grad before exporting. This PR is to add this guard when using export script. It would not affect the models that grad is not enabled. Reviewed By: cccclai Differential Revision: D54812718
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/2373
Note: Links to docs will display an error until the docs builds have been completed. ❌ 1 New FailureAs of commit dba4a9e with merge base 75284d2 ( NEW FAILURE - The following job has failed:
This comment was automatically generated by Dr. CI and updates every 15 minutes. |
facebook-github-bot
added
the
CLA Signed
Contributor
|
This pull request was exported from Phabricator. Differential Revision: D54812718 |
cccclai
approved these changes
Mar 12, 2024
Contributor
|
This pull request has been merged in 35a847e. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Summary: For some models where the grad is enabled (like LLava model), we need to disable grad before exporting. This PR is to add this guard when using export script. It would not affect the models that grad is not enabled.
Reviewed By: cccclai
Differential Revision: D54812718