transformers
[`distributed`] Context parallelism: padding masks, sliding windows and chunked attention with sdpa
#50516
Closed

[`distributed`] Context parallelism: padding masks, sliding windows and chunked attention with sdpa #50516

3outeille wants to merge 2 commits into cp-7-auto-pad from cp-8-masks
3outeille
3outeille [distributed] Rebuild the attention mask of the whole sequence under …
4615fb03
3outeille [distributed] Test padding, sliding windows and chunked attention wit…
fb226b82
HuggingFaceDocBuilderDev
3outeille 3outeille force-pushed the cp-7-auto-pad branch from 4249c6db to 043fd08c 1 day ago
3outeille 3outeille force pushed from 4a6cf258 to fb226b82 1 day ago
3outeille
3outeille 3outeille closed this 15 hours ago
3outeille 3outeille deleted the cp-8-masks branch 15 hours ago

Login to write a write a comment.

Login via GitHub

Reviewers
No reviews
Assignees
No one assigned
Labels
Milestone