huggingface/transformers

Pull Requests Commits

Rocketknight1 committed 2 years ago

a8cf3e03

Rocketknight1 committed 2 years ago

39407389

Fix bug for checkpoint saving on multi node training setting (#28078)

dumpmemory committed 2 years ago

Verified 1c286be5

make torch.load a bit safer (#27282)

julien-c committed 2 years ago

Verified dec84b32

Make GPT2 traceable in meta state (#28054)

kwen2501 committed 2 years ago

Verified 74cae670

[LLaVa] Add past_key_values to _skip_keys_device_placement to fix multi-GPU dispatch (#28051)

aismlv committed 2 years ago

Verified e2b6df79

Skip M4T `test_retain_grad_hidden_states_attentions` (#28060)

ylacombe committed 2 years ago

Verified deb72cb6

[`Mixtral`] update conversion script to reflect new changes (#28068)

younesbelkada committed 2 years ago

Verified d269c4b2

doc: Correct spelling mistake (#28064)

caiyili committed 2 years ago

Verified 70a127a3

Remove SpeechT5 deprecated argument (#28062)

ylacombe committed 2 years ago

Verified c817c17d

[Flax LLaMA] Fix attn dropout (#28059)

sanchit-gandhi committed 2 years ago

Verified 6af3ce77

[Flax BERT] Update deprecated 'split' method (#28012)

sanchit-gandhi committed 2 years ago

Verified 7e876dca

[`Modeling` / `Mixtral`] Fix GC + PEFT issues with Mixtral (#28061)

younesbelkada committed 2 years ago

Verified e737446e

[`FA-2`] Fix fa-2 issue when passing `config` to `from_pretrained` (#28043)

younesbelkada committed 2 years ago

Verified 1e209317

Remove warning when Annotion enum is created (#28048)

amyeroberts committed 2 years ago

Verified 1a585c12

Replace build() with build_in_name_scope() for some TF tests (#28046)

Rocketknight1 committed 2 years ago

Verified 3060899b

Proper build() methods for TF (#27794)

Rocketknight1 committed 2 years ago

Verified 050e0b44

[Seamless] Fix links in docs (#27905)

sanchit-gandhi committed 2 years ago

Verified 52c37882

Generate: Mistral/Mixtral FA2 cache fix when going beyond the context window (#28037)

gante committed 2 years ago

Verified 388fd314

Fixed spelling error in T5 tokenizer warning message (s/thouroughly/t… (#28014)

jeddobson committed 2 years ago

Verified 0ede7626

Fix languages covered by M4Tv2 (#28019)

ylacombe committed 2 years ago

Verified bb1d0d0d

SeamlessM4T: `test_retain_grad_hidden_states_attentions` is flaky (#28035)

gante committed 2 years ago

Verified e2b16485

Generate: assisted decoding now uses `generate` for the assistant (#28030)

gante committed 2 years ago

Verified 9e5c28c5

Fix AMD push CI not triggered (#28029)

ydshieh committed 2 years ago

Verified dde6c427

[`core` / `modeling`] Fix training bug with PEFT + GC (#28031)

younesbelkada committed 2 years ago

Verified 73de5108

[`SeamlessM4TTokenizer`] Safe import (#28026)

ArthurZucker committed 2 years ago

Verified 2788f8d8

well well well (#28011)

ArthurZucker committed 2 years ago

Verified 131a528b

add `modules_in_block_to_quantize` arg in GPTQconfig (#27956)

SunMarc committed 2 years ago

Verified 17506d12

Add model_docs from cpmant.md to derformable_detr.md (#27884)

rajveer43 committed 2 years ago

Verified fe44b1f1

LysandreJik committed 2 years ago

3ed3e319

Older