NielsRogge/transformers

Pull Requests Commits

Address comments

NielsRogge committed 1 year ago

22d3d70c

NielsRogge committed 1 year ago

743a9674

NielsRogge committed 1 year ago

40bd7315

Merge remote-tracking branch 'upstream/main' into feature/use_processor

NielsRogge committed 1 year ago

b64070cb

Rework tests to compare trainer checkpoint args (#29883)

muellerzr committed 1 year ago

Verified 3b8e2932

[`BC`] Fix BC for AWQ quant (#29965)

TechxGenus committed 1 year ago

Verified 6e584070

Update model card and link of blog post. (#29928)

bozheng-hit committed 1 year ago

Verified 46d63681

Reset alarm signal when the function is ended (#29706)

coldnight committed 1 year ago

Verified f6701bc6

fix: get mlflow version from mlflow-skinny (#29918)

Alexander Jipa committed 1 year ago

Verified e644b600

Add warning message for `run_qa.py` (#29867)

jla524 committed 1 year ago

Verified 156d30da

Fix rope theta for OpenLlama (#29893)

jla524 committed 1 year ago

Verified 6fd93fe9

Super tiny fix 12 typos about "with with" (#29926)

fzyzcjy committed 1 year ago

Verified 5ad7f170

Mark `test_eager_matches_sdpa_generate` flaky for some models (#29479)

ydshieh committed 1 year ago

Verified 43d17c18

Update installs in image classification doc (#29947)

MariaHei committed 1 year ago

Verified ba56ed08

[`LlamaSlowConverter`] Slow to Fast better support (#29797)

ArthurZucker committed 1 year ago

Verified 536ea2ac

Fix doc issue #29758 in DebertaV2Config class (#29842)

vinayakkgarg committed 1 year ago

Verified e2036468

[`BC`] Fix BC for other libraries (#29934)

ArthurZucker committed 1 year ago

Verified 2bbbf1be

Allow GradientAccumulationPlugin to be configured from AcceleratorConfig (#29589)

fabianlim committed 1 year ago

Verified 4df5b9b4

[ `TokenizationLlama`] fix the way we convert tokens to strings to keep leading spaces 🚨 breaking fix (#29453)

ArthurZucker committed 1 year ago

Verified a2a7f716

[`Mamba`] from pretrained issue with `self.embeddings` (#29851)

ArthurZucker committed 1 year ago

Verified e677479c

RoPE models: add numerical sanity-check test for RoPE scaling (#29808)

gante committed 1 year ago

Verified 441de62f

add functions to inspect model and optimizer status to trainer.py (#29838)

CKeibel committed 1 year ago

Verified aac7099c

Safe import of LRScheduler (#29919)

amyeroberts committed 1 year ago

Verified 855b95ce

Add beam search visualizer to the doc (#29876)

aymeric-roucher committed 1 year ago

Verified c9d2e855

Tests: replace `torch.testing.assert_allclose` by `torch.testing.assert_close` (#29915)

gante committed 1 year ago

Verified 248d5d23

[doc] fix some typos and add `xpu` to the testing documentation (#29894)

faaany committed 1 year ago

Verified 7c19fafe

Adding Flash Attention 2 Support for GPT2 (#29226)

EduardoPach committed 1 year ago

Verified 22d159dd

[`pipeline`]. Zero shot add doc warning (#29845)

ArthurZucker committed 1 year ago

Verified 3a7e6836

[`GptNeox`] don't gather on pkv when using the trainer (#29892)

ArthurZucker committed 1 year ago

Verified 543889f3

[`make fix-copies`] update and help (#29924)

ArthurZucker committed 1 year ago

Verified b256516a

Older