huggingface/transformers

Pull Requests Commits

Merge branch 'main' into fixing_gptq_tests

MekkCyber committed 233 days ago

Verified cb7df519

Add tearDown method to Quark to solve OOM issues (#38234)

MekkCyber committed 233 days ago

Verified 9a962dd9

fix multi-image case for llava-onevision (#38084)

cyr0930 committed 233 days ago

Verified 101b3fa4

[`compile`] re-enable for Qwen-VL models (#38127)

zucchini-nlp committed 233 days ago

Verified a21f11fc

[Falcon H1] Fix Typo in Integration Test (#38256)

dhiaEddineRhaiem committed 233 days ago

Verified 4542086d

[MODEL] Add Falcon H1 (#38249)

younesbelkada committed 233 days ago

Verified 6829936e

tp plan should not be NONE (#38255)

ArthurZucker committed 233 days ago

Verified e288ee00

Revert parallelism temporarily (#38240)

LysandreJik committed 234 days ago

Verified 711d78d1

CI reporting improvements (#38230)

ydshieh committed 234 days ago

Verified feec294d

Protect ParallelInterface

LysandreJik committed 234 days ago

cb513e35

LysandreJik committed 234 days ago

f4ef41c4

[gemma3] fix bidirectional attention mask (#38080)

zucchini-nlp committed 234 days ago

Verified f834d368

[mllama] fix loading and inference (#38223)

zucchini-nlp committed 234 days ago

Verified 2edb0e4b

add to expected output

MekkCyber committed 234 days ago

ebf0e701

Add padding-free to bamba (#35861)

garrett361 committed 234 days ago

Verified 390f1534

Fixing Bitnet after use_rms_norm introduction (#38229)

MekkCyber committed 234 days ago

Verified 2a794713

Enable Quantize KV Cache for Mistral Model (#35042)

Bojun-Feng committed 234 days ago

Verified 96618960

parallelism goes brrr (#37877)

NouamaneTazi committed 234 days ago

Verified 1c2f36b4

Fix Llama4 (#38222)

Cyrilvallez committed 234 days ago

Verified b591d925

Mamba2 remove unecessary test parameterization (#38227)

ivarflakstad committed 234 days ago

Verified 3f0b7d0f

Minor llama4 fixes (#38123)

molbap committed 234 days ago

Verified 9cde2f5d

fix dead flax links modeling_flax_pytorch_utils.py (#38212)

dizer-ti committed 234 days ago

Verified 856f034f

Make `train_dataset` attribute in `_get_train_sampler` optional (#38226)

SunMarc committed 234 days ago

Verified bb3c6426

In Llama4 fix wrongly inverted causal attention mask when using SDPA implementation (#38094)

sogartar committed 234 days ago

Verified 2ad152f8

Disable torchscript tests for AriaForConditionalGenerationModelTest (#38225)

ivarflakstad committed 234 days ago

Verified de70c842

Add support to Marimo Notebooks and Enverge.ai (#38210)

brenoca committed 234 days ago

Verified 8ea61c45

New cache tests and refactored Hybrid Cache (#37972)

manueldeprada committed 234 days ago

Verified d34e21e7

Add `Llama4TextModel` to `AutoModel` mapping (#38162)

Matthew Hoffman committed 234 days ago

Verified 183fb363

Remove trust_remote_code=True tests from bnb quantization tests (MPT now integrated) (#38206)

Titus-von-Koeller committed 234 days ago

Verified f022bf93

[fix] sliding window attention mask (#38045)

zucchini-nlp committed 234 days ago

Verified 0a52bd24

Older