huggingface/transformers

Pull Requests Commits

[cleanup] delete deprecated kwargs in qwen2_audio 🧹 (#38404)

gante committed 1 year ago

Verified f85fd904

[CSM] update model id (#38211)

eustlb committed 1 year ago

Verified b9f8f863

Add report_repo_id to mi300 workflow (#38401)

ivarflakstad committed 1 year ago

Verified 07dd6b24

[CSM] infer codec model with no_grad + audio eos label (#38215)

eustlb committed 1 year ago

Verified 3142bd85

Fix Qwen2.5-VL Video Processor (#38366)

yeliudev committed 1 year ago

Verified 10ae443e

[chat] use the checkpoint's `generation_config.json` as base parameterization (#38330)

gante committed 1 year ago

Verified 80902ae9

Fix convert to original state dict for VLMs (#38385)

hiyouga committed 1 year ago

Verified 008e0d87

[chat] improvements for thinking models and reduce default verbosity (#38322)

gante committed 1 year ago

Verified c7694831

guard size mismatch check to only quantized models (#38397)

SunMarc committed 1 year ago

Verified 55f23333

[aya vision] fix processor for vLLM (#38371)

zucchini-nlp committed 1 year ago

Verified 1a5be2f5

[video utils] group and reorder by number of frames (#38374)

zucchini-nlp committed 1 year ago

Verified 19fdb75c

[paligemma] fix processor with suffix (#38365)

zucchini-nlp committed 1 year ago

Verified b0735dc0

[transformers x vLLM] standardize processors (#37915)

zucchini-nlp committed 1 year ago

Verified 9e1017b4

Fix image token mask in Gemma3 (#38295)

Cyrilvallez committed 1 year ago

Verified b5ececb9

Add AMD MI300 CI caller leveraging self-hosted runner scale set workflow in hf-workflows (#38132)

jitesh-gupta committed 1 year ago

Verified c4e71e8f

Stop autoconverting custom code checkpoints (#37751)

Rocketknight1 committed 1 year ago

Verified 706b0092

update gemma tests (#38384)

ydshieh committed 1 year ago

Verified 07848a84

[cli] cli usable without torch (#38386)

gante committed 1 year ago

Verified cd0f3ce7

:rotating_light: :rotating_light: Fix custom code saving (#37716)

Rocketknight1 committed 1 year ago

Verified ba6d7222

Stop TF weight rename reDOS (#38325)

Rocketknight1 committed 1 year ago

Verified 701caef7

fix typo: `tokenizer` -> `tokenize` (#38357)

foldl committed 1 year ago

Verified 0a4e8e28

fix typos (#38336)

DeVikingMark committed 1 year ago

Verified 63964b7c

Better check in `initialize_weights` (#38382)

Cyrilvallez committed 1 year ago

Verified 8b03c8ea

Use one `utils/notification_service.py` (#38379)

ydshieh committed 1 year ago

Verified eb74cf97

for now disable compile (#38383)

ArthurZucker committed 1 year ago

Verified 98328fd9

Improved cache docs (#38060)

manueldeprada committed 1 year ago

Verified 78079abe

[Falcon H1] Fix slow path forward pass (#38320)

dhiaEddineRhaiem committed 1 year ago

Verified 7a9b071b

Protect `get_default_device` for torch<2.3 (#38376)

Cyrilvallez committed 1 year ago

Verified b5b76b55

Fix incorrect batching audio index calculation for Phi-4-Multimodal (#38103)

Isotr0py committed 1 year ago

Verified bff32678

Fix all import errors based on older torch versions (#38370)

Cyrilvallez committed 1 year ago

Verified 9f0402bc

Newer Older