huggingface/transformers

Pull Requests Commits

Update VITS modeling to enable ONNX export (#28141)

echarlaix committed 2 years ago

Verified 7226f3d2

fix FA2 when using quantization for remaining models (#28341)

susnato committed 2 years ago

Verified cadf93a6

[DETA] Improvement and Sync from DETA especially for training (#27990)

SangbumChoi committed 2 years ago

Verified 899d8351

Fix pos_mask application and update tests accordingly (#27892)

ferjorosa committed 2 years ago

Verified 57e9c832

Don't check the device when device_map=auto (#28351)

yuanwu2017 committed 2 years ago

Verified 03b98099

README: install transformers from conda-forge channel (#28313)

kevherro committed 2 years ago

Verified 5d36025c

Fix error in M4T feature extractor (#28340)

ylacombe committed 2 years ago

Verified 35e9d2b2

enable training mask2former and maskformer for transformers trainer (#28277)

SangbumChoi committed 2 years ago

Verified 4a66c0d9

[docs] Sort es/toctree.yml | Translate performance.md (#28262)

aaronjimv committed 2 years ago

Verified 6b8ec258

Translate contributing.md into Chinese (#28243)

Mayfsz committed 2 years ago

Verified 3ea88336

Remove token_type_ids from model_input_names (like #24788) (#28325)

Apsod committed 2 years ago

Verified 45b1dfa3

Add FastSpeech2Conformer (#23439)

Connor Henderson committed 2 years ago

Verified d83ff5ee

fix documentation for zero_shot_object_detection (#28267)

not-lain committed 2 years ago

Verified 6eba901d

Bump tj-actions/changed-files from 22.2 to 41 in /.github/workflows (#28311)

dependabot[bot] committed 2 years ago

Verified c2d283a6

Remove fast tokenization warning in Data Collators (#28213)

dbuos committed 2 years ago

Verified aa4a0f8e

[Whisper] Fix errors with MPS backend introduced by new code on word-level timestamps computation (#28288)

ercaronte committed 2 years ago

Verified 5be46dfc

fix bug:divide by zero in _maybe_log_save_evaluate() (#28251)

frankenliu committed 2 years ago

Verified 87ae2a46

Fix trainer saving safetensors: metadata is None (#28219)

hiyouga committed 2 years ago

Verified 502a10a6

Update docs around mixing hf scheduler with deepspeed optimizer (#28223)

dwyatte committed 2 years ago

Verified cad9f5c6

small typo (#28229)

stas00 committed 2 years ago

Verified 3cefac1d

fix FA2 when using quantization (#28203)

pacman100 committed 2 years ago

Verified 3b7675b2

[`Awq`] Enable the possibility to skip quantization for some target modules (#27950)

younesbelkada committed 2 years ago

Verified fa21ead7

[`Llava`] Fix llava index errors (#28032)

younesbelkada committed 2 years ago

Verified 29e7a1e1

update the logger message with accordant weights_file_name (#28181)

izyForever committed 2 years ago

Verified 68fa1e85

Fixing visualization code for object detection to support both types of bounding box. (#27842)

Anindyadeep committed 2 years ago

Verified 74d9d0ce

[Whisper] Fix word-level timestamps with bs>1 or num_beams>1 (#28114)

ylacombe committed 2 years ago

Verified 5da3db3f

Drop `feature_extractor_type` when loading an image processor file (#28195)

ydshieh committed 2 years ago

Verified c4df7c16

Fix the check of models supporting FA/SDPA not run (#28202)

ydshieh committed 2 years ago

Verified bb3bd447

Bug: `training_args.py` fix missing import with accelerate with version `accelerate==0.20.1` (#28171)

michaelfeil committed 2 years ago

Verified e37ab52d

Add Swinv2 backbone (#27742)

NielsRogge committed 2 years ago

Verified c9fb250a

Newer Older