transformers
Continuous batching thread safety
#44924
Merged

Continuous batching thread safety #44924

Qubitium
Qubitium fix torch.cuda.graph should operate in thread_local mode
24edc8bb
Qubitium fix tie_weights skipping logic is not thread-safe
1f641acb
Qubitium doc
34050ec7
Qubitium cleanup
f9234af4
Qubitium Qubitium changed the title Continuos batching paged attention threads Continuous batching paged attention thread safety 48 days ago
Qubitium Merge branch 'main' into continuos-batching-paged-attention-threads
6aba9849
Qubitium
ArthurZucker ArthurZucker requested a review from remi-or remi-or 47 days ago
ArthurZucker
ArthurZucker commented on 2026-03-23
Qubitium Merge branch 'main' into continuos-batching-paged-attention-threads
c9c81f64
Qubitium revert tie_weight() concurrency bug fix. push to another pr
030340c3
Qubitium Qubitium changed the title Continuous batching paged attention thread safety Continuous batching thread safety 47 days ago
Qubitium Merge branch 'main' into continuos-batching-paged-attention-threads
b4e846c3
Qubitium
Qubitium Merge branch 'main' into continuos-batching-paged-attention-threads
bdeb5035
Qubitium Merge branch 'main' into continuos-batching-paged-attention-threads
9bf17825
remi-or
Qubitium cleanup unit test to only check for `thread_local` error_mode
0bf8b6af
Qubitium
Qubitium Merge branch 'main' into continuos-batching-paged-attention-threads
48d8a30f
Qubitium
Qubitium add true model test
5f27b406
Qubitium Merge branch 'continuos-batching-paged-attention-threads' of https://…
19255c3c
Qubitium
Qubitium commented on 2026-03-23
Qubitium Merge branch 'main' into continuos-batching-paged-attention-threads
4ab6fffa
remi-or
Qubitium remove error_mode set unit test
b9cf5c03
Qubitium
Qubitium Merge branch 'continuos-batching-paged-attention-threads' of https://…
7ac4b32c
Qubitium Merge branch 'main' into continuos-batching-paged-attention-threads
e348311b
remi-or
Qubitium remove unit test
a1057a97
Qubitium Merge branch 'continuos-batching-paged-attention-threads' of https://…
f5accd95
Qubitium
Qubitium
github-actions
remi-or
HuggingFaceDocBuilderDev
ArthurZucker
ArthurZucker approved these changes on 2026-03-23
ArthurZucker ArthurZucker merged dda54684 into main 47 days ago
Qubitium Qubitium deleted the continuos-batching-paged-attention-threads branch 47 days ago
Qubitium Qubitium restored the head branch 46 days ago

Login to write a write a comment.

Login via GitHub

Reviewers
Assignees
No one assigned
Labels
Milestone