NielsRogge/transformers

Pull Requests Commits

Niels Rogge committed 3 years ago

62a48c30

Add support for hybrid checkpoints

Niels Rogge committed 3 years ago

f6ae0d74

Convert BiT checkpoints

Niels Rogge committed 3 years ago

7c593bb9

More improvements

Niels Rogge committed 3 years ago

d77073bb

Update conversion script

Niels Rogge committed 3 years ago

59691f5b

Make forward pass work

Niels Rogge committed 3 years ago

a0f3d982

Niels Rogge committed 3 years ago

79dc342f

Add Donut image processor (#20425)

amyeroberts committed 3 years ago

Verified ae1cffaf

Extract warnings from CI artifacts (#20474)

ydshieh committed 3 years ago

Verified 28247e78

[Maskformer] Add MaskFormerSwin backbone (#20344)

NielsRogge committed 3 years ago

Verified 6dc884ab

add timeout option for deepspeed engine (#20443)

henghuiz committed 3 years ago

Verified 955780d3

chore: add link to the video cls notebook. (#20386)

sayakpaul committed 3 years ago

Verified d59d5a61

Include image processor in add-new-model-like (#20439)

amyeroberts committed 3 years ago

Verified 321ef388

[AutoBackbone] Improve API (#20407)

NielsRogge committed 3 years ago

Verified 0bae286d

fix both failing RoCBert tests (#20469)

ArthurZucker committed 3 years ago

Verified 39a72125

Safetensors offload (#20321)

sgugger committed 3 years ago

Verified 30163921

[FLAX] Add dtype to embedding for bert/bart/opt/t5 (#20340)

merrymercy committed 3 years ago

Verified ac2f6674

Replace assertion with ValueError exceptions in run_image_captioning_flax.py (#20365)

katiele47 committed 3 years ago

Verified 667ccea7

[Doctest] Add configuration_fsmt.py (#19936)

sha016 committed 3 years ago

Verified 0a619325

Replace assertions with value errors on distilbert model (#20463)

JuheonChu committed 3 years ago

Verified 98122794

[CLIPTokenizer] Improve warning (#20458)

patrickvonplaten committed 3 years ago

Verified 134a8e21

with pytorch cpu only version. without --no_cuda, using --bf16 will trigger error like "Your setup doesn't support bf16/gpu. You need torch>=1.10, using Ampere GPU with cuda>=11.0" (#20445)

sywangyi committed 3 years ago

Verified de53e4bf

update cpu related doc (#20444)

sywangyi committed 3 years ago

Verified ca3b652b

make tensors in function build_relative_position created on proper device instead of always on cpu (#20434)

qq775294390 committed 3 years ago

Verified 8f7078e8

More TF int dtype fixes (#20384)

Rocketknight1 committed 3 years ago

Verified de4159a3

Fix ESM checkpoints for tests (#20436)

Rocketknight1 committed 3 years ago

Verified 72b19ca6

Fix doctests for audio models (#20468)

ydshieh committed 3 years ago

Verified f244a978

Fix links for `contrastive_loss` (#20455)

ydshieh committed 3 years ago

Verified df938fc1

Fix device issues in `CLIPSegModelIntegrationTest` (#20467)

ydshieh committed 3 years ago

Verified 2cdac665

Fix typo in FSMT Tokenizer (#20456)

kamalkraj committed 3 years ago

Verified 61d3928b

Older