huggingface/transformers

Pull Requests Commits

[WIP] Emu3: add model (#33770)

zucchini-nlp committed 1 year ago

Verified 52e1f87c

Fix flex_attention in training mode (#35605)

Cyrilvallez committed 1 year ago

Verified ccc0381d

Remove `benchmark.py` after #34275

ArthurZucker committed 1 year ago

a9bd1e62

Chat template: return vectorized output in processors (#34275)

zucchini-nlp committed 1 year ago

Verified e0646f3d

Add Moonshine (#34784)

eustlb committed 1 year ago

Verified 5f087d13

Skip `torchscript` tests if a cache object is in model's outputs (#35596)

ydshieh committed 1 year ago

Verified 6f127d3f

ModernBert: reuse GemmaRotaryEmbedding via modular + Integration tests (#35459)

tomaarsen committed 1 year ago

Verified 6b73ee89

Add flex_attn to diffllama (#35601)

muellerzr committed 1 year ago

Verified 8de7b1ba

ModernBERT bug fixes (#35404)

warner-benjamin committed 1 year ago

Verified 1e3ddcb2

add `_supports_flex_attn = True` for models that do support it (#35598)

ArthurZucker committed 1 year ago

Verified e97d7a5b

[doc] deepspeed universal checkpoint (#35015)

huyiwen committed 1 year ago

Verified c9c682d1

Refactor/fix Cohere2 (#35594)

Cyrilvallez committed 1 year ago

Verified 3a4ae6ea

[`tokenizers`] Ensure that add_prefix_space is propagated to backend_tokenizer.pre_tokenizer (#35593)

tomaarsen committed 1 year ago

Verified 32e0db8a

Fix modular edge case + modular sorting order (#35562)

Cyrilvallez committed 1 year ago

Verified 46276f9a

PR for Issue #22694: Fixed Training Evaluation table display for VSCode (#35557)

dedsec995 committed 1 year ago

Verified d3fe9fa3

Small fix rope kwargs (#35589)

molbap committed 1 year ago

Verified 395b114b

Fix flaky `SwitchTransformersModelTest::test_training_gradient` (#35587)

ydshieh committed 1 year ago

Verified 82dd6c14

`tokenizer` train from iterator without pre_tokenizers (#35396)

ArthurZucker committed 1 year ago

Verified eb4579cf

feat: add TP plan for granite (#35573)

kmehant committed 1 year ago

Verified 320512df

[Idefics3] Move image features to same device as input embeds (#35100)

shyshin committed 1 year ago

Verified 633da1b1

Add inputs_embeds param to ModernBertModel (#35373)

jxmorris12 committed 1 year ago

Verified 832c6191

Fix flaky `test_batching_equivalence` (#35564)

ydshieh committed 1 year ago

Verified 1b2f942a

Setup loss_type in config at model init time (#34616)

ChanderG committed 1 year ago

Verified 4adc415b

Re-add missing __all__ for Cohere and Phi3 (#35578)

Cyrilvallez committed 1 year ago

Verified c8ab6ce6

Minor fix in video text 2 text docs (#35546)

merveenoyan committed 1 year ago

Verified 487c31a2

More model refactoring! (#35359)

Cyrilvallez committed 1 year ago

Verified 965a2fb3

Don't show warning for `inv_freq` buffers (#35255)

zucchini-nlp committed 1 year ago

Verified 137965ca

Fix multi-gpu loss (#35395)

ArthurZucker committed 1 year ago

Verified 8cad65a6

update code owners (#35576)

ArthurZucker committed 1 year ago

Verified 2e2f8015

[i18n-ar] Translated file: `docs/source/ar/tasks/multiple_choice.md` into Arabic (#35199)

AhmedAlmaghz committed 1 year ago

Verified a6256ec0

Newer Older