huggingface/transformers

Pull Requests Commits

moved labels to the same device as logits for BLOOM, GPT Neo, GPT NeoX, RoBERTa and VIT models (#22663)

iamarunbrahma committed 2 years ago

Verified 656e869a

Revert migration of setup to pyproject.toml (#22658)

sgugger committed 2 years ago

Verified 6db23af5

Generate: add API warning to streamers (#22659)

gante committed 2 years ago

Verified 3f96e0b4

[OPT] Fix default attention mask size (#22649)

ArthurZucker committed 2 years ago

Verified f3341926

[tokenization] do not push special file (#22657)

ArthurZucker committed 2 years ago

Verified b1b3dc3e

Small nit, (#22653)

ArthurZucker committed 2 years ago

Verified 117a0f6a

🌐 [i18n-KO] Translated `pipeline_tutorial.mdx` to Korean (#22508)

wonhyeongseo committed 2 years ago

Verified fc1ba6fd

Fix `MegaModel` CI (#22652)

ydshieh committed 2 years ago

Verified 14d5b2b6

Fix typo (#22650)

Ronalmoo committed 2 years ago

Verified f2cc8ffd

Move labels to the same device as logits for LlamaForSequenceClassification and Blip2 (#22596)

xssChauhan committed 2 years ago

Verified 1de8ce9e

🌐[i18n-KO] Translate `autoclass_tutorial` to Korean and Fix the typo of `quicktour` (#22533)

gabrielwithappy committed 2 years ago

Verified d59034ff

fix FSDP version related issues (#22489)

pacman100 committed 2 years ago

Verified ee8e80a0

Update tiny model summary file for recent models (#22637)

ydshieh committed 2 years ago

Verified c7ec71ba

[`Blip`] Fix slow tests and doctests with correct values (#22632)

younesbelkada committed 2 years ago

Verified ed672864

LlamaTokenizerFast Fix (.., from_slow=True). (#22630)

Narsil committed 2 years ago

Verified 6a02e980

[`bnb`] 8bit models should not be converted to `DDP` (#22628)

younesbelkada committed 2 years ago

Verified 09a9888f

A script to add/update `pipeline_model_mapping` systematically (#22180)

ydshieh committed 2 years ago

Verified d0b83fe2

update_pip_test_mapping (#22606)

ydshieh committed 2 years ago

Verified fa01127a

docs: Fix broken link to generation strategies (#22623)

Connor Henderson committed 2 years ago

Verified 321b0908

Make tiny model creation + pipeline testing more robust (#22500)

ydshieh committed 2 years ago

Verified 2c22bc79

Backbone add mixin tests (#22542)

amyeroberts committed 2 years ago

Verified 12d51db2

Seq2SeqTrainer: use unwrapped model to retrieve the generation config (#22584)

gante committed 2 years ago

Verified 48706c71

Revert error back into warning for byte fallback conversion. (#22607)

Narsil committed 2 years ago

Verified 0aa1153f

Adding Llama FastTokenizer support. (#22264)

Narsil committed 2 years ago

Verified 1670be4b

feat(model parallelism): moving the labels to the same device as the logits for gpt2 and bart (#22591)

kausmeows committed 2 years ago

Verified 15641892

Use native TF checkpoints for the BLIP TF tests (#22593)

Rocketknight1 committed 2 years ago

Verified e577bd0f

Add DePlot + MatCha on `transformers` (#22528)

younesbelkada committed 2 years ago

Verified 176ceff9

Adding support for BPE merge creation from scores instead of ids. (#22582)

Narsil committed 2 years ago

Verified 126eafe3

Fix a typo in one of the BLIP pretrained checkpoint names (#22588)

Rocketknight1 committed 2 years ago

Verified 12f1a3bb

Sync preprocesses before loading the processor at run_speech_recognition_ctc.py (#21926)

mpenagar committed 2 years ago

Verified d5239bab

Newer Older