🤗 Transformers: the model-definition framework for state-of-the-art machine learning models in text, vision, audio, and multimodal models, for both inference and training.
COMMITS
January 30, 2026
Y
Better new failures reporting for PR comment CI (#43629)
Yih-Dar committed
S
[docs] serving (#42853)
Steven Liu committed
K
add XPU expected output for MixedInt8GPT2Test (#43615)
kaixuanliu committed
M
Don't modify mappings in tests (#43634)
Matt committed
C
Allow Attention and Experts to be used as standalone modules (#43622)
Cyril Vallez committed
R
Don't modify `tied_weight_keys` in-place (#43619)
Raushan Turganbay committed
A
[`Rope`] Revert #43410 and make inheritance implicit again (#43620)
Anton Vlasjuk committed
C
[vllm compat] Separate renaming from conversion ops (#43621)
Cyril Vallez committed
F
refactor + robusts tests for Tensor Parallel (#42809)
Ferdinand Mom committed
K
K
add xpu expectation for lw_detr model (#43339)
kaixuanliu committed
K
minimax_m2: fix failed test case for XPU (#43324)
kaixuanliu committed
Y
Improve new failures reporting (#43628)
Yih-Dar committed
T
Fix extras on all supported Python versions (#43490)
Tarek Ziade committed
Y
🚨 Remove deprecated AnnotionFormat (#42983)
Yuanyuan Chen committed
H
fix(models): Fix suno/bark-small CPU offload device mismatch causing CI failures (#43607)
Harshal Janjani committed
R
[CB] [Serve] Fix broken serve tests (#43594)
Rémi Ouazan committed
F
Docs: fix typo in weight converter guide (#43610)
Fahad Alghanim committed
Y
[MoE] Use int input for histc on CUDA to support deterministic algorithms (#43583)
YangKai0616 committed
R
Fixes configuration default values (#43592)
Raushan Turganbay committed
R
Fix `make_batched_video` with 5D arrays (#43486)
Raushan Turganbay committed
January 29, 2026
M
Operation Green CI II (#43537)
Matt committed
J
enable cpu paged cache (#42869)
jiqing-feng committed
R
Qwen3 omni - fix get video features (#43588)
Raushan Turganbay committed
J
[GLM-Image] Add batch > 1 support and fix configuration defaults (#43342)
Jared Wen committed
Y
[Model] Refactor modernbert with the attention interface (#43030)
YangKai0616 committed
C
Regex post processing in loading (#43585)
Cyril Vallez committed
I
simplify extra tokens logic in base (#43230)
Ita Zaporozhets committed
Y
Add XPU support to the tests for solar_open (#43579)
YangKai0616 committed
W
remove FbgemmFp8LinearTest (#43545)
Wang, Yi committed