add Flash Attention Support for Helsinki-NLP/opus models #36169
Labels
Feature request
Request for a new feature
Good Second Issue
Issues that are more difficult to do than "Good First" issues - give it a try if you want!
Feature request
I would like to propose adding support for flash attention to the Helsinki-NLP/opus models in the Hugging Face library. Judging by what I am seeing here, it seems not implemented.
Motivation
performance gains: adding flash attention would allow for significant speedups
Your contribution
I am happy to help test the implementation if needed. Thank you for considering this enhancement, and I look forward to the discussion!
The text was updated successfully, but these errors were encountered: