NielsRogge/transformers

Pull Requests Commits

Address comments

NielsRogge committed 2 years ago

61fb3160

Fix integration test

NielsRogge committed 2 years ago

566e0e7b

Fix push_to_hub for folders

NielsRogge committed 2 years ago

0f2f0067

Fix doc snippets and integration test

NielsRogge committed 2 years ago

e1eeab6c

Remove autocast and print statements

NielsRogge committed 2 years ago

5efc2a70

NielsRogge committed 2 years ago

0e5b23ee

NielsRogge committed 2 years ago

9a1eb220

Add integration test

NielsRogge committed 2 years ago

5385bf59

Add processor tests

NielsRogge committed 2 years ago

e4b2fd04

Update conversion script

NielsRogge committed 2 years ago

80065da2

NielsRogge committed 2 years ago

58f5ee96

Add InstructBlipProcessor

NielsRogge committed 2 years ago

8b0a12b0

Add model to toctree

NielsRogge committed 2 years ago

714486da

NielsRogge committed 2 years ago

6c75c52f

NielsRogge committed 2 years ago

8914ce4f

Make LLama output match

NielsRogge committed 2 years ago

7e8aa688

Cast the weights

NielsRogge committed 2 years ago

70578a48

NielsRogge committed 2 years ago

72c168b9

More improvements

NielsRogge committed 2 years ago

5d762c43

Fix README and auto mappings

NielsRogge committed 2 years ago

f4f83d51

Convert flan-t5-xl checkpoint

NielsRogge committed 2 years ago

8843867d

Add first draft

NielsRogge committed 2 years ago

5c8714f3

[image-to-text pipeline] Add conditional text support + GIT (#23362)

NielsRogge committed 2 years ago

Verified 2f424d79

Update workflow files (#23658)

ydshieh committed 2 years ago

Verified e69feab8

Update all no_trainer with skip_first_batches (#23664)

muellerzr committed 2 years ago

Verified b191d7db

Fix SAM tests and use smaller checkpoints (#23656)

Rocketknight1 committed 2 years ago

Verified 26a06814

changing the requirements to a cpu torch version that works (#23483)

sshahrokhi committed 2 years ago

Verified 6f72e71f

Fix wav2vec2 is_batched check to include 2-D numpy arrays (#23223)

LWprogramming committed 2 years ago

Verified 5de2a6d5

Bugfix: LLaMA layer norm incorrectly changes input type and consumers lots of memory (#23535)

TimDettmers committed 2 years ago

Verified 4ddd9de9

Muellerzr fix deepspeed (#23657)

muellerzr committed 2 years ago

Verified fe34486f

Older