NielsRogge/transformers

Pull Requests Commits

NielsRogge committed 2 years ago

fbeb7310

Fix SDPA dispatch & make SDPA CI compatible with torch<2.1.1 (#27940)

fxmarty committed 2 years ago

Verified 9f18cc6d

[LLaVa] Some improvements (#27895)

NielsRogge committed 2 years ago

Verified 7ea21f1f

Fix `SeamlessM4Tv2ModelIntegrationTest` (#27911)

ylacombe committed 2 years ago

Verified 5e620a92

Skip `UnivNetModelTest::test_multi_gpu_data_parallel_forward` (#27912)

ydshieh committed 2 years ago

Verified e96c1de1

[BEiT] Fix test (#27934)

NielsRogge committed 2 years ago

Verified 8d8970ef

[DETA] fix backbone freeze/unfreeze function (#27843)

SangbumChoi committed 2 years ago

Verified 235be085

Fix typo (#27918)

f4hy committed 2 years ago

Verified df5c5c62

[integration] Update Ray Tune integration for Ray 2.7 (#26499)

justinvyu committed 2 years ago

Verified 5fa66df3

[CLAP] Replace hard-coded batch size to enable dynamic ONNX export (#27790)

xenova committed 2 years ago

Verified ffd426ee

F.scaled_dot_product_attention support (#26572)

fxmarty committed 2 years ago

Verified 80377eb0

Generate: SinkCache can handle iterative prompts (#27907)

gante committed 2 years ago

Verified ce0bbd51

fix typo in image_processing_blip.py Wwhether -> Whether (#27899)

zhc7 committed 2 years ago

Verified 94c76538

[Doc] Spanish translation of pad_truncation.md (#27890)

aaronjimv committed 2 years ago

Verified d6c3a3f1

Allow `resume_from_checkpoint` to handle `auto_find_batch_size` (#27568)

muellerzr committed 2 years ago

Verified 6757ed28

fix llava (#27909)

ArthurZucker committed 2 years ago

Verified aa7ab98e

Llama conversion script: adjustments for Llama Guard (#27910)

pcuenca committed 2 years ago

Verified e0b617d1

Fix 2 tests in `FillMaskPipelineTests` (#27889)

ydshieh committed 2 years ago

Verified e3669375

Fix `notification_service.py` (#27903)

ydshieh committed 2 years ago

Verified 79e76559

mark `test_initialization` as flaky in 2 model tests (#27906)

ydshieh committed 2 years ago

Verified 3b720ad9

Fix CLAP converting script (#27153)

ylacombe committed 2 years ago

Verified 7f07c356

Fix remaining issues in beam score calculation (#27808)

VsonicV committed 2 years ago

Verified b31905d1

Fix beam score calculation issue for Tensorflow version (#27814)

VsonicV committed 2 years ago

Verified 3ac9945e

fix: non-atomic checkpoint save (#27820)

thundergolfer committed 2 years ago

Verified 4c5ed1d0

Added passing parameters to "reduce_lr_on_plateau" scheduler (#27860)

CharbelAD committed 2 years ago

Verified fe8d1302

Fix: Raise informative exception when `prefix_allowed_tokens_fn` return empty set of tokens (#27797)

Saibo-creator committed 2 years ago

Verified 56be5e80

[⚠️ removed a default argument] Make `AttentionMaskConverter` compatible with `torch.compile(..., fullgraph=True)` (#27868)

fxmarty committed 2 years ago

Verified 307a7d0b

Generate: New `Cache` abstraction and Attention Sinks support (#26681)

tomaarsen committed 2 years ago

Verified 633215ba

Translate `model_doc` files from `clip` to `cpm` to JP (#27774)

rajveer43 committed 2 years ago

Verified 0ea42ef0

Updates the distributed CPU training documentation to add instructions for running on a Kubernetes cluster (#27780)

dmsuehir committed 2 years ago

Verified 79b79ae2

Older