AutoModel supports FA2/paged attention #2133
Closed
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
fxmarty
commented
Jun 27, 2024
server/text_generation_server/models/custom_modeling/flash_llama_modeling.py
Show resolved
Hide resolved
5 tasks
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Outdated suggestions cannot be applied.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.
As per title.
The models benefiting from it in Transformers are:
i.e. all models with
supports_cache_class = Trueand_supports_flash_attn_2 = True, following huggingface/transformers#31446 and some more changes in Transformers needed to support a single dim for total sequence length[total_seqlen, hidden_size].models/__init__.pyis kind of bloated and I guess is going to be refactored with the upcoming TRT-LLM support / multi backend.