huggingface/transformers

Pull Requests Commits

ydshieh committed 2 years ago

705edb59

[docs] Backbone (#28739)

stevhliu committed 2 years ago

Verified abbffc45

Add models from deit (#28302)

rajveer43 committed 2 years ago

Verified 23ea6743

[docs] fix some bugs about parameter description (#28806)

zspo committed 2 years ago

Verified d98591a1

enable graident checkpointing in DetaObjectDetection and add tests in Swin/Donut_Swin (#28615)

SangbumChoi committed 2 years ago

Verified e19c12e0

Add tip on setting tokenizer attributes (#28764)

Rocketknight1 committed 2 years ago

Verified 7bc6d763

Fix symbolic_trace with kv cache (#28724)

fxmarty committed 2 years ago

Verified 709dc432

Make `is_torch_bf16_available_on_device` more strict (#28796)

ydshieh committed 2 years ago

Verified eb8e7a00

Adding [T5/MT5/UMT5]ForTokenClassification (#28443)

hackyon committed 2 years ago

Verified 0d26abdd

[docs] Correct the statement in the docstirng of compute_transition_scores in generation/utils.py (#28786)

Ki-Seki committed 2 years ago

Verified 7b2bd1fb

Split daily CI using 2 level matrix (#28773)

ydshieh committed 2 years ago

Verified 47358661

Add artifact name in job step to maintain job / artifact correspondence (#28682)

ydshieh committed 2 years ago

Verified 95346e9d

DeepSpeed: hardcode `torch.arange` dtype on `float` usage to avoid incorrect initialization (#28760)

gante committed 2 years ago

Verified beb2a096

Flax mistral (#26943)

kiansierra committed 2 years ago

Verified f7076cd3

Wrap Keras methods to support BatchEncoding (#28734)

Rocketknight1 committed 2 years ago

Verified 7a496100

canonical repos moves (#28795)

julien-c committed 2 years ago

Verified 721e2d94

Resolve DeepSpeed cannot resume training with PeftModel (#28746)

lh0x00 committed 2 years ago

Verified bebeeee0

[Whisper] Refactor forced_decoder_ids & prompt ids (#28687)

patrickvonplaten committed 2 years ago

Verified 65a926e8

[`HFQuantizer`] Remove `check_packages_compatibility` logic (#28789)

younesbelkada committed 2 years ago

Verified f9f1f2ac

don't initialize the output embeddings if we're going to tie them to input embeddings (#28192)

tom-p-reichel committed 2 years ago

Verified ae0c27ad

Prevent MLflow exception from disrupting training (#28779)

codiceSpaghetti committed 2 years ago

Verified a937425e

[`bnb`] Fix bnb slow tests (#28788)

younesbelkada committed 2 years ago

Verified d703eaae

Pin Torch to <2.2.0 (#28785)

Rocketknight1 committed 2 years ago

Verified 74c9cfea

Add tf_keras imports to prepare for Keras 3 (#28588)

Rocketknight1 committed 2 years ago

Verified 415e9a09

Task-specific pipeline init args (#28439)

amyeroberts committed 2 years ago

Verified 1d489b3e

[`Backbone`] Use `load_backbone` instead of `AutoBackbone.from_config` (#28661)

amyeroberts committed 2 years ago

Verified 2fa1c808

Further pin pytest version (in a temporary way) (#28780)

ydshieh committed 2 years ago

Verified c24c5245

Fix transformers.utils.fx compatibility with torch<2.0 (#28774)

fxmarty committed 2 years ago

Verified 6f7d5db5

Use Conv1d for TDNN (#25728)

gau-nernst committed 2 years ago

Verified 5c8d941d

[`HfQuantizer`] Move it to "Developper guides" (#28768)

younesbelkada committed 2 years ago

Verified 866253f8

Older