huggingface/transformers

Pull Requests Commits

Add Fast PromptDepthAnything Processor (#40602)

SamuelBarryCS committed 1 year ago

Verified ff26fe83

Use torch.expm1 and torch.log1p for better numerical results (#40860)

cyyever committed 1 year ago

Verified 6254bb4a

Clarify passing is_causal in sdpa_attention_paged_forward (#40838)

cyyever committed 1 year ago

Verified e674e9da

🔴 Move variable output controls to `_prepare_generation_config ` (#40715)

manueldeprada committed 1 year ago

Verified 0957999f

Fix modular consistency (#40883)

Cyrilvallez committed 1 year ago

Verified 5e9ec59d

[`VaultGemma`] Update expectations in integration tests (#40855)

vasqu committed 1 year ago

Verified 3442b2f3

Adding Support for Qwen3-VL Series (#40795)

JJJYmmm committed 1 year ago

Verified c0dbe095

[Qwen3 Next] Use numerically stable `rsqrt` (#40848)

艾力可 committed 1 year ago

Verified fc5f9105

Update model tags and integration references in bug report (#40881)

ArthurZucker committed 1 year ago

Verified 96d3795c

fix: XIELU act parameters not being casted to correct dtype (#40812)

NanoCode012 committed 1 year ago

Verified f5e16418

fix florence kwargs (#40826)

SunMarc committed 1 year ago

Verified ada64ce4

[docstrings / type hints] Update outdated annotations for `past_key_values` (#40803)

gante committed 1 year ago

Verified 93f810e6

[Bug fix #40813] Fix base_model_tp_plan of Starcoder2 model. (#40814)

greg-kwasniewski1 committed 1 year ago

Verified c65fea0b

Redirect MI355 CI results to dummy dataset (#40862)

ahadnagy committed 1 year ago

Verified 9c804f7e

Fix TrainingArguments.parallelism_config NameError with accelerate<1.10.1 (#40818)

albertvillanova committed 1 year ago

Verified 02ea2b34

Use checkpoint in auto_class_docstring (#40844)

cyyever committed 1 year ago

Verified d42e96a2

[generate] Always use decoder config to init cache (#40772)

gante committed 1 year ago

Verified 6eb32558

[tests] move generative tests away from `test_modeling_common.py` (#40854)

gante committed 1 year ago

Verified e682f90f

[test] Fix test_eager_matches_sdpa incorrectly skipped (#40852)

eustlb committed 1 year ago

Verified 8d845913

add: differential privacy research model (#40851)

RyanMullins committed 1 year ago

Verified 291772b6

[Sam2Video] Fix video inference with batched boxes and add test (#40797)

yonigozlan committed 1 year ago

Verified 8502b41b

[SAM2] Fix inconsistent results with original implementation with input boxes (#40800)

yonigozlan committed 1 year ago

Verified f384bb8a

[tests] re-enable aria fast tests (#40846)

gante committed 1 year ago

Verified 4cb41ad2

Fixes for continuous batching (#40828)

remi-or committed 1 year ago

Verified ef053939

Fix the misalignment between the l2norm in GDN of Qwen3-Next and the implementation in the FLA library. (#40842)

bozheng-hit committed 1 year ago

Verified 98a80781

Replace image classification loss functions to `self.loss_function` (#40764)

qubvel committed 1 year ago

Verified 77aa35ee

Update no split modules in T5Gemma model (#40810)

npuichigo committed 1 year ago

Verified 797859c9

Adds Causal Conv 1D kernel for mamba models (#40765)

MekkCyber committed 1 year ago

Verified 6e69b608

Add VideoProcessors to auto-backend requirements (#40843)

Cyrilvallez committed 1 year ago

Verified 827b65c4

Improve torch_dtype checks (#40808)

cyyever committed 1 year ago

Verified 5e2e77fb

Newer Older