NielsRogge/transformers

Pull Requests Commits

Let `EarlyStoppingCallback` not require `load_best_model_at_end` (#35101)

muellerzr committed 346 days ago

Verified b02828e4

Added error when sequence length is bigger than max_position_embeddings (#32156)

Taha1506 committed 346 days ago

Verified 0aaf124f

Use inherit tempdir makers for tests + fix failing DS tests (#35600)

muellerzr committed 346 days ago

Verified 1211e616

Fix flaky `test_custom_4d_attention_mask` (#35606)

ydshieh committed 346 days ago

Verified bbc00046

ArthurZucker committed 346 days ago

f63829c8

[WIP] Emu3: add model (#33770)

zucchini-nlp committed 346 days ago

Verified 52e1f87c

Fix flex_attention in training mode (#35605)

Cyrilvallez committed 346 days ago

Verified ccc0381d

Remove `benchmark.py` after #34275

ArthurZucker committed 346 days ago

a9bd1e62

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

zucchini-nlp committed 346 days ago

Verified e0646f3d

Add Moonshine (#34784)

eustlb committed 346 days ago

Verified 5f087d13

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

ydshieh committed 346 days ago

Verified 6f127d3f

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

tomaarsen committed 346 days ago

Verified 6b73ee89

Add flex_attn to diffllama (#35601)

muellerzr committed 346 days ago

Verified 8de7b1ba

ModernBERT bug fixes (#35404)

warner-benjamin committed 346 days ago

Verified 1e3ddcb2

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

ArthurZucker committed 346 days ago

Verified e97d7a5b

[doc] deepspeed universal checkpoint (#35015)

huyiwen committed 346 days ago

Verified c9c682d1

Refactor/fix Cohere2 (#35594)

Cyrilvallez committed 346 days ago

Verified 3a4ae6ea

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

tomaarsen committed 346 days ago

Verified 32e0db8a

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

Cyrilvallez committed 346 days ago

Verified 46276f9a

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

dedsec995 committed 347 days ago

Verified d3fe9fa3

Small fix rope kwargs (#35589)

molbap committed 347 days ago

Verified 395b114b

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

ydshieh committed 347 days ago

Verified 82dd6c14

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

ArthurZucker committed 347 days ago

Verified eb4579cf

feat: add TP plan for granite (#35573)

kmehant committed 347 days ago

Verified 320512df

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

shyshin committed 347 days ago

Verified 633da1b1

Add inputs_embeds param to ModernBertModel (#35373)

jxmorris12 committed 347 days ago

Verified 832c6191

Fix flaky `test_batching_equivalence` (#35564)

ydshieh committed 347 days ago

Verified 1b2f942a

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

ChanderG committed 347 days ago

Verified 4adc415b

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

Cyrilvallez committed 347 days ago

Verified c8ab6ce6

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

merveenoyan committed 347 days ago

Verified 487c31a2

Newer Older