open_clip
[WIP] Allow for attention caching during CoCa generation
#502

Open

[WIP] Allow for attention caching during CoCa generation #502

sramshetty wants to merge 11 commits into mlfoundations:main from sramshetty:inference_caching

initial caching for generation

ecb93d95

remove timing

8ad92401

fix beamsearch caching

71d01f52

WIP Setup base for text encoder caching

09e3fec8

Merge branch 'mlfoundations:main' into inference_caching

df85d0c4

Fix transformer caching default

2c961e6d

avoid passing cache when not necessary

69a936ae

simplify caching argument

8dc14ed8

sramshetty marked this pull request as draft 3 years ago

fix transformer cache typing for true branch

5ed4cf21

reorder typing to address list invariance

c0686dee

remove unnecessary placeholder lists

e561c8f6

Reviewers

No reviews

Assignees

No one assigned

Labels

None yet

Milestone

No milestone