huggingface/transformers

Pull Requests Commits

check deepspeed multi-gpu

ydshieh committed 2 years ago

5303b69c

[Mistral, Mixtral] Improve docs (#29084)

NielsRogge committed 2 years ago

Verified dabe8556

[Gemma] Fix eager attention (#29187)

sanchit-gandhi committed 2 years ago

Verified 2a9b1f80

Add training version check for AQLM quantizer. (#29142)

BlackSamorez committed 2 years ago

Verified fc37f389

FIX [`Gemma`] Fix bad rebase with transformers main (#29170)

younesbelkada committed 2 years ago

Verified ae49b218

[ `gemma`] Adds support for Gemma 💎 (#29167)

ArthurZucker committed 2 years ago

Verified 594c1277

[`Maskformer`] safely get backbone config (#29166)

asartran committed 2 years ago

Verified 58245ba6

support SDPA Attention in stablelm (#29106)

eaidova committed 2 years ago

Verified 1d0ea7ab

`torch.compile` compatibility with `generate` + static cache (#29114)

fxmarty committed 2 years ago

Verified cc4a664b

🚨 Llama: update rope scaling to match static cache changes (#29143)

gante committed 2 years ago

Verified 3994fa5b

ArthurZucker committed 2 years ago

1a77f07f

[`pipeline`] Add pool option to image feature extraction pipeline (#28985)

asartran committed 2 years ago

Verified e770f031

Fix drop path being ignored in DINOv2 (#29147)

fepegar committed 2 years ago

Verified c47576ca

Added image_captioning version in es and included in toctree file (#29104)

gisturiz committed 2 years ago

Verified 3c00b885

Generate: missing generation config eos token setting in encoder-decoder tests (#29146)

gante committed 2 years ago

Verified 857fd8ea

Raise unused kwargs image processor (#29063)

molbap committed 2 years ago

Verified 1c81132e

[Phi] Add support for sdpa (#29108)

hackyon committed 2 years ago

Verified b8b16475

Save (circleci) cache at the end of a job (#29141)

ydshieh committed 2 years ago

Verified 7688d8df

Add support for fine-tuning CLIP-like models using contrastive-image-text example (#29070)

tjs-intel committed 2 years ago

Verified ee3af60b

Revert low cpu mem tie weights (#29135)

asartran committed 2 years ago

Verified 0996a100

[`Core tokenization`] `add_dummy_prefix_space` option to help with latest issues (#28010)

ArthurZucker committed 2 years ago

Verified 15cfe389

FIX [`PEFT` / `Trainer` ] Handle better peft + quantized compiled models (#29055)

younesbelkada committed 2 years ago

Verified efdd4366

[`cuda kernels`] only compile them when initializing (#29133)

ArthurZucker committed 2 years ago

Verified 5e95dcab

Generate: unset GenerationConfig parameters do not raise warning (#29119)

gante committed 2 years ago

Verified a7755d24

Llama: fix batched generation (#29109)

gante committed 2 years ago

Verified 7d312ad2

FIX [`bnb` / `tests`] Propagate the changes from #29092 to 4-bit tests (#29122)

younesbelkada committed 2 years ago

Verified ff76e7c2

Abstract image processor arg checks. (#28843)

molbap committed 2 years ago

Verified 1c9134f0

FEAT [`Trainer` / `bnb`]: Add RMSProp from `bitsandbytes` to HF `Trainer` (#29082)

younesbelkada committed 2 years ago

Verified f7ef7cec

Move misplaced line (#29117)

kno10 committed 2 years ago

Verified a7ff2f23

[`gradient_checkpointing`] default to use it for torch 2.3 (#28538)

ArthurZucker committed 2 years ago

Verified 9094abe8

Older