huggingface/transformers

Pull Requests Commits

ydshieh committed 2 years ago

733d214d

ydshieh committed 2 years ago

1a5ea48d

ydshieh committed 2 years ago

c57911dc

Fix weights not properly initialized due to shape mismatch (#28122)

ydshieh committed 2 years ago

Verified 7938c8c8

move code to Trainer.evaluate to enable use of that function with multiple datasets (#27844)

peter-sk committed 2 years ago

Verified 769a9542

[gpt-neox] Add attention_bias config to support model trained without attention biases (#28126)

dalgarak committed 2 years ago

Verified cd9f9d63

Fix FA2 integration (#28142)

pacman100 committed 2 years ago

Verified def581ef

Remove deprecated CPU dockerfiles (#28149)

ashahba committed 2 years ago

Verified b134f685

[docs] Fix mistral link in mixtral.md (#28143)

aaronjimv committed 2 years ago

Verified 38611086

Update modeling_utils.py (#28127)

mzelling committed 2 years ago

Verified 23f8e4db

[`Mixtral`] Fix loss + nits (#28115)

ArthurZucker committed 2 years ago

Verified 4a04b4cc

Generate: speculative decoding (#27979)

gante committed 2 years ago

Verified ac974199

Update split string in doctest to reflect #28087 (#28135)

amyeroberts committed 2 years ago

Verified bd7a3561

When save a model on TPU, make a copy to be moved to CPU (#27993)

qihqi committed 2 years ago

Verified 5aec50ec

[Doc] Fix token link in What 🤗 Transformers can do (#28123)

aaronjimv committed 2 years ago

Verified 4edffda6

Fix a typo in tokenizer documentation (#28118)

mssalvatore committed 2 years ago

Verified c52b515e

[docs] General doc fixes (#28087)

stevhliu committed 2 years ago

Verified a52e180a

Fix indentation error - semantic_segmentation.md (#28117)

rajveer43 committed 2 years ago

Verified 08a6e7a7

More TF fixes (#28081)

Rocketknight1 committed 2 years ago

Verified 71d47f0a

Remove warning if `DISABLE_TELEMETRY` is used (#28113)

Wauplin committed 2 years ago

Verified 0695b242

Disable jitter noise during evaluation in SwitchTransformers (#28077)

DaizeDong committed 2 years ago

Verified 7c5408da

fix ConversationalPipeline docstring (#28091)

not-lain committed 2 years ago

Verified a0522de4

in peft finetune, only the trainable parameters need to be saved (#27825)

sywangyi committed 2 years ago

Verified e6cb8e05

Spelling correction (#28110)

saeneas committed 2 years ago

Verified 7f2a8f92

[`Llava` / `Vip-Llava`] Add SDPA into llava (#28107)

younesbelkada committed 2 years ago

Verified b8378b65

Fix the deprecation warning of _torch_pytree._register_pytree_node (#27803)

cyyever committed 2 years ago

Verified e6dcf8ab

4D `attention_mask` support (#27539)

poedator committed 2 years ago

Verified f85a1e82

fix resuming from ckpt when using FSDP with FULL_STATE_DICT (#27891)

pacman100 committed 2 years ago

Verified 238d2e3c

[docs] MPS (#28016)

stevhliu committed 2 years ago

Verified ebfdb9ca

[docs] Trainer (#27986)

stevhliu committed 2 years ago

Verified 0d63d177

Older