Skip to content

Pull requests: huggingface/transformers

New pull request New
950 Open 21,663 Closed
950 Open 21,663 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Reviews
Filter by reviews
No reviews Review required Approved review Changes requested
Assignee
Filter by who’s assigned
Assigned to nobody Loading
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

Trainer: Pass num_items_in_batch to compute_loss in prediction_step
#41183 opened Sep 26, 2025 by pramodith Loading…
3 of 5 tasks
1
1
Update GLM-4.1V MMRope implementation
#41182 opened Sep 26, 2025 by zRzRzRzRzRzRzR Loading…
1
download and use HF Hub Cache
#41181 opened Sep 26, 2025 by ydshieh Loading…
1
Fix Latex typesetting in documentation
#41177 opened Sep 26, 2025 by cyyever Loading…
1 of 5 tasks
4
Optimize rope_deltas propagation logic in Qwen2.5-VL
#41176 opened Sep 26, 2025 by Xqle Loading…
5 tasks
2
Bump hfh prerelease version
#41175 opened Sep 26, 2025 by Wauplin Loading…
2
🚨 [v5] Delete feature extractors used for vision
#41174 opened Sep 26, 2025 by zucchini-nlp Loading…
Rope for Qwen2--5-vl
#41173 opened Sep 26, 2025 by zucchini-nlp Loading…
1
2
Fix typsetting and content of llm_tutorial_optimization.md
#41172 opened Sep 26, 2025 by cyyever Loading…
1 of 5 tasks
4
🚨 [v5] Remove relative position embeddings (for bert like models)
#41170 opened Sep 25, 2025 by vasqu Loading…
10
Fix TorchDynamo crash in StaticCache by validating offloading and offload_only_non_sliding arguments
#41169 opened Sep 25, 2025 by Flakes342 Loading…
4 of 5 tasks
1
[v5] Remove model_parallel deprecated feature
#41166 opened Sep 25, 2025 by SunMarc Loading…
3
[Trainer] deprecate num_train_tokens
#41165 opened Sep 25, 2025 by SunMarc Loading…
1
[DistilBert] Refactor Attention
#41163 opened Sep 25, 2025 by vasqu Loading…
2
Add Sinhala (සිංහල) translation of README
#41162 opened Sep 25, 2025 by Ranjuna120 Loading…
5 tasks
2
Remove old sagemaker api support
#41161 opened Sep 25, 2025 by SunMarc Loading…
1
[docs] update tips syntax
#41160 opened Sep 25, 2025 by mishig25 Loading…
2
Support setting total_train_batch_size.
#41159 opened Sep 25, 2025 by zhengchenyu Loading…
5 tasks
1
fix qwen text config
#41158 opened Sep 25, 2025 by zucchini-nlp Loading…
1
3
Fix white space in documentation
#41157 opened Sep 25, 2025 by cyyever Loading…
1 of 5 tasks
4
Fix inaccurate train_tokens_per_second when resuming from checkpoint
#41156 opened Sep 25, 2025 by lilin-1 Loading…
2 of 5 tasks
1
5
add rotary kernel support to Qwen3 model
#41147 opened Sep 25, 2025 by kaixuanliu Loading…
10
Support automatic conversion from zero checkpoint to universal checkpoint.
#41144 opened Sep 25, 2025 by zhengchenyu • Draft
5 tasks
1
Adapt to the SDPA interface to enable the NPU to call FlashAttentionScore
#41143 opened Sep 25, 2025 by frozenleaves Loading…
18
Fix typing of train_args
#41142 opened Sep 25, 2025 by cyyever • Draft
1
ProTip! Adding no:label will show everything without a label.