huggingface/transformers

Pull Requests Commits

update to check instance

itazap committed 30 days ago

ca3742b0

Merge branch 'main' into tokenizer-validation

itazap committed 30 days ago

Verified f1f73931

validate tokenizer components

itazap committed 30 days ago

56550972

Fix integration test in Owlv2 image processing tests (#42783)

yonigozlan committed 30 days ago

Verified dfe6e4c0

[kernels] Final kernel removal 🥳 (#41664)

MekkCyber committed 30 days ago

Verified de055d6d

Vision docs 📝 (#42096)

merveenoyan committed 30 days ago

Verified eaa3d4dd

[Model] Add PaddleOCR-VL Model Support (#42178)

zhang-prog committed 30 days ago

Verified 8c84144b

[CI] fix wav2vec test (#42810)

MekkCyber committed 30 days ago

Verified 78b29929

fix awq (#42776)

SunMarc committed 30 days ago

Verified f8e8ddb0

Better continuous batching tests (#42699)

remi-or committed 30 days ago

Verified f8e5ae6a

[Quantization] FBgemm FP8 for XPU (#42773)

MekkCyber committed 30 days ago

Verified 86644be4

Revert "🚨🚨🚨🚨🚨🚨🚨🚨🚨 default to `"auto"` dtype (#34919)"

LysandreJik committed 30 days ago

a8f32a0e

[`Padding-Free Attention`] Fix packed FA attention with pos ids only (#42801)

vasqu committed 30 days ago

Verified d1eda63f

🚨🚨🚨🚨🚨🚨🚨🚨🚨 default to `"auto"` dtype (#34919)

ArthurZucker committed 30 days ago

Verified f5aa90d0

feat(granitemoe*): Remove logits upcast when computing loss (#42753)

gabe-l-hart committed 31 days ago

Verified 0af2381f

allow registration of custom checkpoint conversion mappings (#42634)

winglian committed 31 days ago

Verified 45d8168e

Fully deprecate AutoGPTQ and AutoAWQ for GPT-QModel (#41567)

Qubitium committed 31 days ago

Verified 8ebfd84f

[core] fix fp-quant (#42613)

MekkCyber committed 31 days ago

Verified c3acdd57

[Quantization] Fix Static FP8 Quantization (#42775)

MekkCyber committed 31 days ago

Verified 15735a43

Override Transformers defaults by GGUF defaults (#42770)

a4lg committed 31 days ago

Verified 51a66739

fix: support tensor labels in DataCollatorWithFlattening (#42620)

hqkqn32 committed 31 days ago

Verified f54647c8

Command-a-vision fix (#42642)

dongluw committed 31 days ago

Verified 1b8ccf1c

Stricter checks for mistral patch (#42743)

ArthurZucker committed 31 days ago

Verified b9951b4e

[kernels] Fix kernel CI (#42764)

MekkCyber committed 31 days ago

Verified 508a9764

Ensure e_score_correction_bias dtype of DeepSeek-V3/R1 is FP32 (#42580)

xin3he committed 31 days ago

Verified 2e29a9a6

[Quantization] Fixing some tests (#42763)

MekkCyber committed 31 days ago

Verified 8f978e5b

Only call `torch.autocast` if it will have an effect (#42747)

hmellor committed 32 days ago

Verified 6d0adb5b

Only default `rope_parameters` to empty `dict` if there is something to put in it (#42651)

hmellor committed 32 days ago

Verified 3230fb50

fix links in `CONTRIBUTING.md` (#42745)

casinca committed 32 days ago

Verified 471d7ce9

Add an alternative scenario to EoMT `post_process_semantic_segmentation` in case `path_offsets` is None (#42716)

simonreise committed 32 days ago

Verified 5b4d72c5

Older