huggingface/transformers

Pull Requests Commits

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

CKeibel committed 2 years ago

Verified aac7099c

Safe import of LRScheduler (#29919)

amyeroberts committed 2 years ago

Verified 855b95ce

Add beam search visualizer to the doc (#29876)

aymeric-roucher committed 2 years ago

Verified c9d2e855

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

gante committed 2 years ago

Verified 248d5d23

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

faaany committed 2 years ago

Verified 7c19fafe

Adding Flash Attention 2 Support for GPT2 (#29226)

EduardoPach committed 2 years ago

Verified 22d159dd

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

ArthurZucker committed 2 years ago

Verified 3a7e6836

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

ArthurZucker committed 2 years ago

Verified 543889f3

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

ArthurZucker committed 2 years ago

Verified b256516a

Fix typo in T5Block error message (#29881)

Mingosnake committed 2 years ago

Verified d9dc993f

MixtralSparseMoeBlock: add gate jitter (#29865)

lorenzoverardo committed 2 years ago

Verified a25037be

add Cambricon MLUs support (#29627)

huismiling committed 2 years ago

Verified 75769744

Move `eos_token_id` to stopping criteria (#29459)

zucchini-nlp committed 2 years ago

Verified 0efcf323

fix fuyu device_map compatibility (#29880)

SunMarc committed 2 years ago

Verified 31c575bc

Reimplement "Automatic safetensors conversion when lacking these files" (#29846)

LysandreJik committed 2 years ago

Verified 4d8427f7

Fix 29807, sinusoidal positional encodings overwritten by post_init() (#29813)

hovnatan committed 2 years ago

Verified a81cf9ee

Mamba `slow_forward` gradient fix (#29563)

vasqu committed 2 years ago

Verified cefb819f

Add Qwen2MoE (#29377)

bozheng-hit committed 2 years ago

Verified 1c39974a

Support `num_attention_heads` != `num_key_value_heads` in Flax Llama Implementation (#29557)

bminixhofer committed 2 years ago

Verified 8e08acad

Set custom_container in build docs workflows (#29855)

Wauplin committed 2 years ago

Verified f01e1609

Disable AMD memory benchmarks (#29871)

IlyasMoutawwakil committed 2 years ago

Verified 07d79520

Add `cosine_with_min_lr` scheduler in Trainer (#29341)

liuyanyi committed 2 years ago

Verified ef609958

Allow `bos_token_id is None` during the generation with `inputs_embeds` (#29772)

LZHgrla committed 2 years ago

Verified 998b5bb5

[docs] Indent ordered list in add_new_model.md (#29796)

windsonsea committed 2 years ago

Verified b9ceb03d

Fix header in IFE task guide (#29859)

merveenoyan committed 2 years ago

Verified de81a677

Replace 'decord' with 'av' in VideoClassificationPipeline (#29747)

Tyx-main committed 2 years ago

Verified b32bf85b

Add warnings if training args differ from checkpoint trainer state (#29255)

jonflynng committed 2 years ago

Verified b5a6d6ee

remove quotes in code example (#29812)

johko committed 2 years ago

Verified 7eb3ba82

[`revert commit`] revert 00a09ed448082da3d6d35fb23a37b7d04f7b4dcd

ArthurZucker committed 2 years ago

e3e16ddc

ArthurZucker committed 2 years ago

00a09ed4

Newer Older