add flash attn 2 #99
Open
add flash attn 2 #99
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
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Outdated suggestions cannot be applied.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.
Marginal improvements over xformers, note that xformers itself isn't a tremendous improvement over vanilla attention.
We have relatively small sequence lengths of 256 for self attention and 256x77 for cross attention. Our head dimensions of 16 heads of 64 inner dim as well are on the smaller side.
Only linking f8 runs, measurements are same for f16 as we have the same attention block configs/sequence lengths
https://wandb.ai/williamberman/muse/runs/9heg02rd?workspace=user-williamberman
https://wandb.ai/williamberman/muse/runs/2tc3ueir?workspace=user-williamberman
https://wandb.ai/williamberman/muse/runs/f4e8r72x/workspace?workspace=user-williamberman