NielsRogge/transformers

Pull Requests Commits

efficiency of MoE to not use unused experts

SangbumChoi committed 1 year ago

b21bb06b

allow flag to enable custom model

SangbumChoi committed 1 year ago

0dd96133

refactor single loop

SangbumChoi committed 1 year ago

a9bb08f0

Revert "fix pos_embed"

SangbumChoi committed 1 year ago

e2fbb269

NielsRogge committed 1 year ago

ba7373f9

Merge branch 'add_vitpose_autobackbone' of https://github.com/NielsRogge/transformers into vitpose-niels

SangbumChoi committed 1 year ago

c12eb355

SangbumChoi committed 1 year ago

2c56a480

[run-slow] vitpose, vitpose_backbone

SangbumChoi committed 1 year ago

e86751a5

rename split module

SangbumChoi committed 1 year ago

28623b1c

Merge branch 'main' into add_vitpose_autobackbone

SangbumChoi committed 1 year ago

Verified 3d4d5595

Fix CI slack reporting issue (#34833)

ydshieh committed 1 year ago

Verified 40821a24

Fix CI by tweaking torchao tests (#34832)

SunMarc committed 1 year ago

Verified 3cb8676a

Fix hyperparameter search when optuna+deepseed (#34642)

corentin-ryr committed 1 year ago

Verified bf42c3bd

Torchao weights only + prequantized compability (#34355)

SunMarc committed 1 year ago

Verified 67890de3

[run-slow] vitpose, vitpose_backbone

SangbumChoi committed 1 year ago

5610d5dd

ruff check --fix

SangbumChoi committed 1 year ago

241211d1

[run-slow] vitpose, vitpose_backbone

SangbumChoi committed 1 year ago

ac004016

SangbumChoi committed 1 year ago

831f70da

Merge branch 'add_vitpose_autobackbone' of https://github.com/NielsRogge/transformers into vitpose-niels

SangbumChoi committed 1 year ago

ac3712fe

enable all conversion

SangbumChoi committed 1 year ago

78fe1b9c

Update src/transformers/models/vitpose/convert_vitpose_to_hf.py

SangbumChoi committed 1 year ago

Verified e588f4ff

[run-slow] vitpose, vitpose_backbone

SangbumChoi committed 1 year ago

d8e6e2ec

SangbumChoi committed 1 year ago

a9a36454

SangbumChoi committed 1 year ago

d4aa3ee6

Fix: take into account meta device (#34134)

tibor-reiss committed 1 year ago

Verified f297af55

fix(DPT,Depth-Anything) `torch.export` (#34103)

philkuz committed 1 year ago

Verified 8cadf76e

Fix the memory usage issue of logits in generate() (#34813)

kjohew committed 1 year ago

Verified 9d16441e

Fix low memory beam search (#34746)

zucchini-nlp committed 1 year ago

Verified 9470d653

LLaVA OV: fix unpadding precision (#34779)

zucchini-nlp committed 1 year ago

Verified 145fbd46

Translate attention.md into Chinese (#34716)

wwwbai committed 1 year ago

Verified 30335093

Older