Skip to content

Pull requests: huggingface/transformers

New pull request New
1,040 Open 22,648 Closed
1,040 Open 22,648 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

[Quantization] fix dequant when block size is none & static quantization
#42545 opened Dec 2, 2025 by MekkCyber Loading…
1
[SAM3] Compute masks once instead of per-layer, fix fa2 crash
#42543 opened Dec 2, 2025 by yonigozlan Loading…
2
handle get_input_embeddings() on models like gemma3 gracefully
#42542 opened Dec 2, 2025 by winglian Loading…
5 tasks
[docs] optimizations quickstart
#42538 opened Dec 1, 2025 by stevhliu Loading…
1
[Ministral 3] Small fix config
#42537 opened Dec 1, 2025 by patrickvonplaten Loading…
1
extend FA2 and other cases to XPU,
#42536 opened Dec 1, 2025 by yao-matrix Loading…
1
Fix documentation for Trainer.place_model_on_device parameter
#42534 opened Dec 1, 2025 by Yuyutsu01 Loading…
1
[torchao] safetensors
#42529 opened Dec 1, 2025 by liangel-02 Loading…
6
Added doctests for SwiftFormer model
#42527 opened Dec 1, 2025 by ankitmalakar7 Loading…
4 of 5 tasks
1
1
Fix interactions between require_read_token and staticmethod
#42522 opened Dec 1, 2025 by Rocketknight1 Loading…
1
Fix FSDP2 defaulting to version 1 in TrainingArguments; add dynamic plugin param passthrough
#42521 opened Dec 1, 2025 by amanzoni1 Loading…
2 tasks
Add backward compatibility for methods which have been moved to RotaryEmbeddingConfigMixin
#42517 opened Dec 1, 2025 by hmellor Loading…
2
Add torch compile to CB run-benchmark
#42516 opened Dec 1, 2025 by remi-or Loading…
1
Fix tokenization fallback
#42514 opened Dec 1, 2025 by ArthurZucker Loading…
3
Move original_max_position_embeddings to rope params
#42513 opened Dec 1, 2025 by zucchini-nlp Loading…
1
Fix Qwen-VL family with prompt tuning
#42508 opened Dec 1, 2025 by zucchini-nlp Loading…
10
Add ModernVBERT models
#42504 opened Dec 1, 2025 by paultltc Loading…
9 of 12 tasks
1
1
Add skip_post_init flag
#42497 opened Nov 30, 2025 by Aravind-11 • Draft
3 of 4 tasks
1
3
feat: allow CP with trainer
#42496 opened Nov 29, 2025 by kmehant Loading…
1 of 5 tasks
2
Fix PEFT gradient checkpointing with frozen input embeddings
#42494 opened Nov 29, 2025 by Aaraviitkgp Loading…
1
fix: remove trailing os sep in local pretrained model path
#42493 opened Nov 29, 2025 by junjunjd • Draft
5 tasks
Fix failing test in Glm4vMoeIntegrationTest
#42488 opened Nov 28, 2025 by Sai-Suraj-27 Loading…
1 of 5 tasks
2
Fix kernel mapping bug when multiple devices exist (Issue #42451)
#42486 opened Nov 28, 2025 by ada-ggf25 Loading…
1
3
more tts pipeline exampel
#42484 opened Nov 28, 2025 by Deep-unlearning Loading…
9
Fixed convert_batch_to_list_format staticmethod function call
#42476 opened Nov 28, 2025 by Sai-Suraj-27 Loading…
1 of 5 tasks
5
ProTip! no:milestone will show everything without a milestone.