-
Notifications
You must be signed in to change notification settings - Fork 32k
Pull requests: huggingface/transformers
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
adding BC for custom toks accessing slow tok attrs deprecated in v5
#43898
opened Feb 10, 2026 by
itazap
Loading…
Decorate cache updates with no_grad, just in case
#43897
opened Feb 10, 2026 by
Rocketknight1
Loading…
Support for BharatGen's Param2MoE model architecture
#43888
opened Feb 10, 2026 by
bhargav-patel-29
Loading…
2 of 5 tasks
Fix UMT5EncoderModel embedding weights not being tied after loading
#43880
opened Feb 10, 2026 by
jiqing-feng
Loading…
Remove remaining vestiges of the TranslationPipeline
#43869
opened Feb 9, 2026 by
Rocketknight1
Loading…
fix: correct typo 'quantizatin_operations' to 'quantization_operations'
#43861
opened Feb 9, 2026 by
thecaptain789
Loading…
Allow to bypass remote code if we want to try and convert it
#43857
opened Feb 9, 2026 by
ArthurZucker
•
Draft
Fix translation task validation when translation pipeline is unavailable
#43849
opened Feb 9, 2026 by
OiPunk
Loading…
Fix _from_config silently skipping weight initialization under DeepSpeed ZeRO-3
#43847
opened Feb 8, 2026 by
tohtana
Loading…
2 of 5 tasks
fix(cli): Fix TypeAdapter NameError when pydantic is not installed
#43842
opened Feb 8, 2026 by
Mr-Neutr0n
Loading…
fix(moe): Handle dtype mismatch in torch._grouped_mm with autocast
#43839
opened Feb 8, 2026 by
Mr-Neutr0n
Loading…
fix: wrapped TypeAdpater in string literals (for now)
#43836
opened Feb 8, 2026 by
pragnyanramtha
Loading…
fix: ensure dtype consistency in grouped_mm under autocast
#43833
opened Feb 8, 2026 by
nulone
Loading…
1 of 3 tasks
Previous Next
ProTip!
Updated in the last three days: updated:>2026-02-07.