transformers
the fix that did not get in
#37370
Merged

the fix that did not get in #37370

Cyrilvallez merged 33 commits into main from flex-fix-regression
ArthurZucker
molbap debugging improvements
2227d491
molbap add debugging details
5d0db09e
molbap add more debugging details
a98c69aa
molbap debug more
35f5bba7
ArthurZucker the fix that did not get in
c7d1e731
github-actions github-actions marked this pull request as draft 255 days ago
github-actions
ArthurZucker ArthurZucker marked this pull request as ready for review 255 days ago
github-actions github-actions requested a review from MekkCyber MekkCyber 255 days ago
github-actions github-actions requested a review from SunMarc SunMarc 255 days ago
HuggingFaceDocBuilderDev
MekkCyber
MekkCyber approved these changes on 2025-04-08
Cyrilvallez First fix flex
c6f442ae
Cyrilvallez fix query offset
98196598
Cyrilvallez fix flex first
5c8c3cd1
Cyrilvallez fix device mask creation for speed
a2761e96
Cyrilvallez small mask creation sdpa
461402af
Cyrilvallez Update flex_attention.py
b9df8463
ArthurZucker ArthurZucker added for patch
ArthurZucker remove chunked prefill from HybridChunkedCache
8a95efaf
ArthurZucker Merge branch 'flex-fix-regression' of github.com:huggingface/transfor…
1d82ac49
ArthurZucker never seen such a fucked up merged
30111923
molbap clean up layers + output
7e1d2209
molbap add summary json file
19095482
Cyrilvallez Efficient general cache
eee77b31
ArthurZucker
ArthurZucker commented on 2025-04-09
Cyrilvallez Update cache_utils.py
f4927976
molbap cleanup
a31018ae
ArthurZucker Merge remote-tracking branch 'origin/model_debugger_upgrades' into fl…
56c607a3
ArthurZucker fix?
030cc7c1
Cyrilvallez fix!
749c1540
Cyrilvallez oups typo
a7b60db2
ArthurZucker Merge branch 'flex-fix-regression' of github.com:huggingface/transfor…
f9b9881b
ArthurZucker not everywhere
02718f98
ArthurZucker more fixes
583cd695
ArthurZucker revert unrelated changes
48ae51a3
Cyrilvallez Fix but ugly for now -> should use pad instead
6b8b5b38
Cyrilvallez oups
9bc8481e
Cyrilvallez re-initialize the cache
31c75c83
Cyrilvallez Use pad to simplify
b1439714
Cyrilvallez style
82c4553c
ArthurZucker
ArthurZucker commented on 2025-04-09
Cyrilvallez correct slicing
0cbdca98
Cyrilvallez Cyrilvallez merged e032d12e into main 254 days ago
Cyrilvallez Cyrilvallez deleted the flex-fix-regression branch 254 days ago

Login to write a write a comment.

Login via GitHub

Assignees
No one assigned
Labels
Milestone