huggingface/transformers

Pull Requests Commits

ydshieh committed 295 days ago

0fa2d206

Support loading Quark quantized models in Transformers (#36372)

fxmarty-amd committed 295 days ago

Verified 1a374799

Use pyupgrade --py39-plus to improve code (#36843)

cyyever committed 295 days ago

Verified ce091b1b

Fix hqq skipped modules and dynamic quant (#36821)

mobicham committed 295 days ago

Verified 3e8f0fbf

Fix ONNX export for sequence classification head (#36332)

echarlaix committed 295 days ago

Verified 055afdb6

Shieldgemma2 (#36678)

RyanMullins committed 295 days ago

Verified 487dab1b

Fix: remove the redundant snippet of _whole_word_mask (#36759)

HuangBugWei committed 295 days ago

Verified a63e92e2

Gemma 3: Adding explicit GenerationConfig and refactoring conversion … (#36833)

RyanMullins committed 295 days ago

Verified 8124a234

Fix import for torch 2.0, 2.1 - guard typehint for "device_mesh" (#36768)

qubvel committed 295 days ago

Verified cf8091c0

Update min safetensors bis (#36823)

SunMarc committed 295 days ago

Verified 388e6659

[generate] clarify docstrings: when to inherit `GenerationMixin` (#36605)

gante committed 295 days ago

Verified b47d9b2f

[modular] Sort modular skips (#36304)

gante committed 295 days ago

Verified 8e97b440

Pass state dict (#35234)

phos-phophy committed 295 days ago

Verified 63380b77

[qwen2 audio] remove redundant code and update docs (#36282)

gante committed 295 days ago

Verified 957b05b4

Update deprecated Jax calls (#35919)

rasmi committed 295 days ago

Verified f0d5b2ff

Fix fp16 ONNX export for RT-DETR and RT-DETRv2 (#36460)

qubvel committed 295 days ago

Verified 1ddb6493

Pass num_items_in_batch directly to loss computation (#36753)

eljandoubi committed 295 days ago

Verified e7337ee7

Saving `Trainer.collator.tokenizer` in when `Trainer.processing_class` is `None` (#36552)

yutong_liu committed 295 days ago

Verified 8b479e39

fix tiktoken convert to pass AddedToken to Tokenizer (#36566)

itazap committed 295 days ago

Verified 3f03c379

[ForCausalLMLoss] allow users to pass shifted labels (#36607)

stas00 committed 295 days ago

Verified 8f64b177

Disable inductor config setter by default (#36608)

HDCharles committed 295 days ago

Verified 94555437

Fix swanlab global step (#36728)

Zeyi-Lin committed 295 days ago

Verified 8733297b

Move the warning to the documentation for DataCollatorWithFlattening (#36707)

qgallouedec committed 295 days ago

Verified b815fae3

Just import torch AdamW instead (#36177)

Rocketknight1 committed 295 days ago

Verified 9be4728a

Update configuration_qwen2.py (#36735)

michaelfeil committed 295 days ago

Verified 51bd0ceb

quick fix fast_image_processor register error (#36716)

JJJYmmm committed 295 days ago

Verified 107fedc1

Add Space to Bitsandbytes doc (#36834)

MekkCyber committed 296 days ago

Verified 258dd9cc

Support tracable dynamicKVcache (#36311)

tugsbayasgalan committed 296 days ago

Verified f39f4960

One more fix for reviewer assignment (#36829)

Rocketknight1 committed 296 days ago

Verified 63c31165

[gemma 3] multimodal checkpoints + AutoModelForCausalLM (#36741)

gante committed 296 days ago

Verified 7c233980

Older