[SFTTrainer] Flash attention support for SFTTrainer
#656
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
younesbelkada
changed the title
SFTTrainer] Flash attention support for SFTTrainerSFTTrainer] Flash attention support for SFTTrainer
|
The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. |
|
Are trl users using padding? |
|
adding |
lvwerra
reviewed
Aug 18, 2023
| ) | ||
| from transformers.trainer_callback import TrainerCallback | ||
| from transformers.trainer_utils import EvalPrediction | ||
| from transformers.utils import ContextManagers |
There was a problem hiding this comment.
hahahah nice!
younesbelkada
changed the title
SFTTrainer] Flash attention support for SFTTrainerSFTTrainer] Flash attention support for SFTTrainer
|
This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread. |
|
Closing for now as huggingface/transformers#25598 might be merged |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 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Outdated suggestions cannot be applied.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.
Related to huggingface/transformers#25265
Users can easily benefit from flash attention, this PR adds a new argument in
SFTTrainerto take care of that and properly document and raise errors when relevant.This leads to some interesting speedups and memory saving that I will detail after doing some experiments, hence putting this PR as draft for now
Only available if you use pytorch nightlies and use
packing=Trueas SDPA + flash attention does not support paddingcc @lvwerra @fxmarty @vwxyzjn