[cpu][flash attention] fix nan issue #133598
Merged
+39
−17
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
Fixes #127055. NaNs are generated in flash attention because the computation of `std::exp((-inf) - (-inf))` and `+/-inf * 0` in lazy softmax. We fix the issue by avoiding the related calculation. Pull Request resolved: #130014 Approved by: https://github.com/jgong5, https://github.com/drisspg (cherry picked from commit 868d9a4)
This was referenced Aug 15, 2024
pytorch-bot
bot
added
the
module: cpu
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/133598
Note: Links to docs will display an error until the docs builds have been completed. ❌ 4 New FailuresAs of commit c506b12 with merge base b66e3f0 ( NEW FAILURES - The following jobs have failed:
This comment was automatically generated by Dr. CI and updates every 15 minutes. |
atalman
approved these changes
Aug 21, 2024
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Outdated suggestions cannot be applied.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.
Fixes #127055.
NaNs are generated in flash attention because the computation of
std::exp((-inf) - (-inf))and+/-inf * 0in lazy softmax. We fix the issue by avoiding the related calculation.cc @jgong5 @mingfeima @XiaobingSuper @sanchitintel @ashokei @jingxu10