NielsRogge/transformers

Pull Requests Commits

Merge remote-tracking branch 'upstream/main' into feature/fix_bitsandbytes

NielsRogge committed 1 year ago

c9ab0f2b

Trainer - add cache clearing and the option for batched eval metrics computation (#28769)

FoamoftheSea committed 1 year ago

Verified df475bf8

Trainer._load_from_checkpoint - support loading multiple Peft adapters (#30505)

claralp committed 1 year ago

Verified e0769530

Fix llava next tie_word_embeddings config (#30640)

SunMarc committed 1 year ago

Verified aa64f086

Quantization / HQQ: Fix HQQ tests on our runner (#30668)

younesbelkada committed 1 year ago

Verified 9c772ac8

Hotfix-change-ci (#30669)

ArthurZucker committed 1 year ago

Verified a45c5148

Check if the current compiled version of pytorch supports MPS (#30664)

jiaqianjing committed 1 year ago

Verified 09edd77f

[`CI update`] Try to use dockers and no cache (#29202)

ArthurZucker committed 1 year ago

Verified 307f632b

Avoid duplication in PR slow CI model list (#30634)

ydshieh committed 1 year ago

Verified 91d155ea

Prevent `TextGenerationPipeline._sanitize_parameters` from overriding previously provided parameters (#30362)

yting27 committed 1 year ago

Verified deb7605a

HQQ: PEFT support for HQQ (#30632)

younesbelkada committed 1 year ago

Verified d0c72c15

Fix W&B run name (#30462)

qubvel committed 1 year ago

Verified 66f675eb

add mlp bias for llama models (#30031)

mayank31398 committed 1 year ago

Verified 425e1a04

Fix CI after #30410 (#30612)

zucchini-nlp committed 1 year ago

Verified a0e77a1f

Add HQQ quantization support (#29637)

mobicham committed 1 year ago

Verified 59952994

Output `None` as attention when layer is skipped (#30597)

jonghwanhyeon committed 1 year ago

Verified 4c940934

Fix FX tracing issues for Llama (#30619)

michaelbenayoun committed 1 year ago

Verified 39359e5b

Generate: fix `SinkCache` on Llama models (#30581)

gante committed 1 year ago

Verified 9719202d

Docs: add missing `StoppingCriteria` autodocs (#30617)

gante committed 1 year ago

Verified 66abe139

Docs: fix `generate`-related rendering issues (#30600)

gante committed 1 year ago

Verified aa55ff44

phi3 chat_template does not support system role (#30606)

amitportnoy committed 1 year ago

Verified 801894e0

Use `contiguous()` in clip checkpoint conversion script (#30613)

ydshieh committed 1 year ago

Verified f57f0149

fix:missing `output_router_logits` in SwitchTransformers (#30573)

lausannel committed 1 year ago

Verified a65da83d

Fix copies for DBRX - neuron fix (#30610)

amyeroberts committed 1 year ago

Verified 4ad5adaf

🚨 Update image_processing_vitmatte.py (#30566)

rb-synth committed 1 year ago

Verified f9530258

Fix memory leak with CTC training script on Chinese languages (#30358)

lucky-bai committed 1 year ago

Verified 12c5544d

Fix for Neuron (#30259)

michaelbenayoun committed 1 year ago

Verified fbabd674

Fix: failing CI after #30568 (#30599)

zucchini-nlp committed 1 year ago

Verified 5cf3e6bf

Bump torch from 1.9.0+cpu to 1.13.1 in /examples/flax/vision (#21168)

dependabot[bot] committed 1 year ago

Verified c681b58b

Bump pillow from 10.0.1 to 10.2.0 in /examples/research_projects/decision_transformer (#28655)

dependabot[bot] committed 1 year ago

Verified 3a36597a

Older