huggingface/transformers

Pull Requests Commits

Merge commit '10bc8b4e' into dynamic_length_on_b6eb708b

ydshieh committed 2 years ago

58c831bd

Merge commit 'cbf98115' into dynamic_length_on_b6eb708b

ydshieh committed 2 years ago

2fc9e12e

0ae789e0 + dynamic length in static cache

ydshieh committed 2 years ago

cbf98115

fix EosTokenCriteria

ydshieh committed 2 years ago

10bc8b4e

ArthurZucker committed 2 years ago

b6eb708b

Add fixed resize and pad strategy for object detection (#30742)

qubvel committed 2 years ago

Verified bf646fbf

update release script (#30880)

ArthurZucker committed 2 years ago

Verified e9a8041d

Support arbitrary processor (#30875)

ArthurZucker committed 2 years ago

Verified 0a9300f4

[whisper] fix multilingual fine-tuning (#30865)

sanchit-gandhi committed 2 years ago

Verified 57edd84b

Fix dependencies for image classification example (#30842)

jla524 committed 2 years ago

Verified 977ce58a

Enable device map (#30870)

darshana1406 committed 2 years ago

Verified 3802e786

Remove deprecated logic and warnings (#30743)

amyeroberts committed 2 years ago

Verified 57c965a8

TEST: Add llama logits tests (#30835)

younesbelkada committed 2 years ago

Verified 3d7d3a87

Fix VideoLlava imports (#30867)

amyeroberts committed 2 years ago

Verified 15c74a28

TST / Quantization: Reverting to torch==2.2.1 (#30866)

younesbelkada committed 2 years ago

Verified 4e17e7dc

Docs: update example with assisted generation + sample (#30853)

gante committed 2 years ago

Verified f4014e75

Video-LLaVa: Fix docs (#30855)

zucchini-nlp committed 2 years ago

Verified 95b3c381

Make `Gemma` work with `torch.compile` (#30775)

ydshieh committed 2 years ago

Verified 1b3dba94

Disable the FA backend for SDPA on AMD GPUs (#30850)

mht-sharma committed 2 years ago

Verified 0753134f

Cache: add new flag to distinguish models that `Cache` but not static cache (#30800)

gante committed 2 years ago

Verified 9d889f87

[Idefics2] Improve docs, add resources (#30717)

NielsRogge committed 2 years ago

Verified 17cc71e1

add sdpa to ViT [follow up of #29325] (#30555)

hyenal committed 2 years ago

Verified 1c21f48a

[LLaVa-NeXT] Small fixes (#30841)

NielsRogge committed 2 years ago

Verified 9fd606db

Fix llama model sdpa attention forward function masking bug when output_attentions=True (#30652)

Aladoro committed 2 years ago

Verified 4b3eb19f

Use `torch 2.3` for CI (#30837)

ydshieh committed 2 years ago

Verified 2d83324e

FEAT / Bitsandbytes: Add `dequantize` API for bitsandbytes quantized models (#30806)

younesbelkada committed 2 years ago

Verified 3f435823

Deprecate models script - correctly set the model name for the doc file (#30785)

amyeroberts committed 2 years ago

Verified 58faa7b8

Better llava next. (#29850)

nxphi47 committed 2 years ago

Verified 5ca085b8

Update ds_config_zero3.json (#30829)

pacman100 committed 2 years ago

Verified bdfefbad

Missing `Optional` in typing. (#30821)

xkszltl committed 2 years ago

Verified 92544cb8

Older