huggingface/transformers

Pull Requests Commits

Copy the template resolution logic from the base apply_chat_template to Voxtral (#45117)

Rocketknight1 committed 56 days ago

Verified 5c7190f0

add kwargs to all methods in the CallbackHandler class (#45353)

wilnn committed 56 days ago

Verified 520005e3

Close file handler (#45187)

ydshieh committed 56 days ago

Verified c9f1b95a

fix: restore mypy type checking for PreTrainedConfig subclasses (#45071) (#45240)

shhKnight30 committed 56 days ago

Verified ce83f16d

`cohere_asr`: fix device issue for `test_model_parallel_beam_search` (#45214)

kaixuanliu committed 56 days ago

Verified e07692b3

Fix AttributeError in Gemma3ForConditionalGeneration and Gemma3ForSequenceClassification when config.return_dict=False (#45277)

kamalrajkannan78 committed 56 days ago

Verified 507a37d1

fix bug for videomt model device mismatch (#45204)

kaixuanliu committed 56 days ago

Verified 4fd862fc

fix gemma4 gradient accumulation loss and last token incorrect labels (#45354)

winglian committed 56 days ago

Verified 47d77657

Logger has `[transformers]` prefix in non-verbose mode (#45316)

zucchini-nlp committed 56 days ago

Verified d13b44b6

Fix AttributeError in AssistantToTargetTranslator.unmap_input_ids with cross-vocab models (#45320)

Regata3010 committed 56 days ago

Verified c43f15c7

Fix Qwen2.5-VL temporal RoPE scaling applied to still images (#45330)

Kash6 committed 56 days ago

Verified a9f5b3a8

musicflamingo: add test support for Intel XPU device (#45212)

kaixuanliu committed 56 days ago

Verified d1cca998

nomic_bert: make the test suitable for general device. (#45209)

kaixuanliu committed 56 days ago

Verified 020ace76

Skip invalid flash-attn tests for `pi0` model (#45011)

kaixuanliu committed 56 days ago

Verified f6ff4ed8

Add cuda compatibility check for using `grouped_mm` (#45001)

Sai-Suraj-27 committed 56 days ago

Verified 68abf9ba

Load adapter with TP (#45155)

michaelbenayoun committed 57 days ago

Verified c585eeaa

[docs] tp training (#44613)

stevhliu committed 57 days ago

Verified 277677d3

[docs] training performance (#44342)

stevhliu committed 57 days ago

Verified 5c79cf02

[docs] optimizers, hyperparam search, training features (#44290)

stevhliu committed 57 days ago

Verified f965b10b

Remove unused parameters and improve add_tensor_parallel_hooks_t… (#44768)

michaelbenayoun committed 57 days ago

Verified 9d840ea5

Use torchvision `decode_image` to load images in the torchvision backend (#45195)

yonigozlan committed 57 days ago

Verified d6a89042

[gemma4] Fix device map auto (#45347)

Cyrilvallez committed 57 days ago

Verified 6546bec4

Refactor CLIP-like models (#44431)

zucchini-nlp committed 57 days ago

Verified 353a9dc4

refactor: display test duration (#45344)

tarekziade committed 57 days ago

Verified 1e931b8f

http retries on audio file downloads (#45126)

tarekziade committed 57 days ago

Verified 25c80bc4

Fix `Wav2Vec2Config.vocab_size` type to allow `None` (#45108)

jiqing-feng committed 57 days ago

Verified 83c3672b

fix(testing): Fix Kyutai Speech-To-Text and LongCatFlash test failures on main CI (#44695)

harshaljanjani committed 57 days ago

Verified 655707b9

[Qwen3_5]Remove unnecessary masked_fill_ in torch_chunk_gated_delta_rule attention computation: "attn = (q_i @ k_i.transpose(-1, -2) * decay_mask[:, :, i]).masked_fill_(mask, 0)" (#45215)

ENg-122 committed 57 days ago

Verified 0aff0dbf

Add THD support in ESM (#44145)

balvisio committed 57 days ago

Verified 3170e369

[gemma4] Remove all shared weights, and silently skip them during loading (#45336)

Cyrilvallez committed 57 days ago

Verified 9f8ddaab

Newer Older