Triton RMSNorm #1050
Merged
Triton RMSNorm #1050
Conversation
This file contains 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
dakinggg
reviewed
Mar 22, 2024
dakinggg
reviewed
Mar 22, 2024
dakinggg
reviewed
Mar 22, 2024
dakinggg
reviewed
Mar 22, 2024
dakinggg
reviewed
Mar 22, 2024
dakinggg
reviewed
Mar 22, 2024
dakinggg
reviewed
Mar 22, 2024
dakinggg
reviewed
Mar 22, 2024
dakinggg
approved these changes
Mar 22, 2024
Co-authored-by: Daniel King <43149077+dakinggg@users.noreply.github.com>
vchiley
approved these changes
Apr 2, 2024
KuuCi
pushed a commit
that referenced
this pull request
Apr 18, 2024
* Triton RMS Norm * fix * refactor * Fix test * no bias * Update tests/models/test_model.py Co-authored-by: Daniel King <43149077+dakinggg@users.noreply.github.com> * fixes * Unittest * fix * fix2 * fix3 * fix3 --------- Co-authored-by: Daniel King <43149077+dakinggg@users.noreply.github.com>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
This PR adds support for FlashAttention's Triton implementation of RMSnorm (commonly found in LLaMa-like models).
The flag is enabled by setting
parameters.model.norm_typetotriton_rmsnorm.Under a LLaMa-2-7b-like workload (displayed below) using this implementation produces a lift of ~1400 tok/s/gpu while preserving the loss. The improvement is so large because the current RMSnorm implementation is done in vanilla PyTorch (see here).