Bumping flash attention version to 2.6.3 and adding option for softcap in attention and lm_head logits. #1374
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
ShashankMosaicML
changed the title
ShashankMosaicML
changed the title
…ry into soft_cap_attn
ShashankMosaicML
changed the title
ShashankMosaicML
changed the title
dakinggg
approved these changes
Sep 3, 2024
dakinggg
approved these changes
Sep 22, 2024
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Outdated suggestions cannot be applied.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 bumps flash attention version to 2.6.3.
It also adds the option for softcap in attention and lm_head logits, to allow Gemma-like models. The config names are same as the huggingface names here: https://github.com/huggingface/transformers/blob/96a074fa7e2c04b904f72d9e827398d4c5f90f25/src/transformers/models/gemma2/modeling_gemma2.py#L371
Loss curves for attn softcapping vs no attn softcapping:
MFU curves for attn softcapping vs no attn softcapping:
