huggingface/transformers

Pull Requests Commits

draft, wip, but that's the idea

molbap committed 1 year ago

1ae0cbde

[cache refactor] Move all the caching logic to a per-layer approach (#39106)

manueldeprada committed 1 year ago

Verified c338fd43

General weight initialization scheme (#39579)

Cyrilvallez committed 1 year ago

Verified b16688e9

Add AMD GPU expectations for LLaVA tests (#39486)

ahadnagy committed 1 year ago

Verified 015b62bf

Kernels flash attn (#39474)

ArthurZucker committed 1 year ago

Verified efceeaf2

Add AMD expectations to Mistral3 tests (#39481)

ahadnagy committed 1 year ago

Verified b62557e7

[docs] Create page on inference servers with transformers backend (#39550)

zucchini-nlp committed 1 year ago

Verified 18065833

[docs] update attention implementation and cache docs (#39547)

zucchini-nlp committed 1 year ago

Verified cd98c1fe

Add AMD test expectations to DETR model (#39539)

ahadnagy committed 1 year ago

Verified ef99537f

[timm_wrapper] add support for gradient checkpointing (#39287)

Yozer committed 1 year ago

Verified 30567c28

Fixes needed for n-d parallelism and TP (#39562)

winglian committed 1 year ago

Verified a44dcbe5

Bump AMD container for 2.7.1 PyTorch (#39458)

ahadnagy committed 1 year ago

Verified 0cae633c

Add EfficientLoFTR model (#36355)

sbucaille committed 1 year ago

Verified a88ea9cb

[gemma3] fix bidirectional image mask (#39396)

zucchini-nlp committed 1 year ago

Verified 3bc726b3

Update OLMoE model card (#39344)

nlhmnlhmnlhm committed 1 year ago

Verified fbeaf96f

Update modernbertdecoder docs (#39453)

orionw committed 1 year ago

Verified 641aaed7

[`CI`] Fix post merge ernie 4.5 (#39561)

vasqu committed 1 year ago

Verified 049a674e

[Fast image processors] Improve handling of image-like inputs other than images (segmentation_maps) (#39489)

yonigozlan committed 1 year ago

Verified b3ebc761

[`Ernie 4.5`] Add ernie text models (#39228)

vasqu committed 1 year ago

Verified b4115a42

Refactor embedding input/output getter/setter (#39339)

molbap committed 1 year ago

Verified 69b15826

🌐 [i18n-KO] Translated `perf_infer_gpu_multi.md` to Korean (#39441)

nevertmr committed 1 year ago

Verified 2da97f09

[Fast image processor] refactor fast image processor glm4v (#39490)

yonigozlan committed 1 year ago

Verified 82807e56

fix ndim check of device_mesh for TP (#39538)

winglian committed 1 year ago

Verified 4b4f04fc

Refactor `MambaCache` to `modeling_mamba.py` (#38086)

manueldeprada committed 1 year ago

Verified 1aa7256f

Fix Docstring of BarkProcessor (#39546)

st81 committed 1 year ago

Verified a419a402

use the enable_gqa param in torch.nn.functional.scaled_dot_product_at… (#39412)

sywangyi committed 1 year ago

Verified 9323d087

Fix missing initializations for models created in 2023 (#39239)

bvantuan committed 1 year ago

Verified 6b3a1f2f

Raise `TypeError` instead of ValueError for invalid types (#38660)

Sai-Suraj-27 committed 1 year ago

Verified 970d9a75

Fix pylint warnings (#39477)

cyyever committed 1 year ago

Verified 822c5e45

Fix Qwen Omni integration test (#39553)

Cyrilvallez committed 1 year ago

Verified dc017cd7

Older