feat-unify-encode-function-0622
#19
by
nan
- opened
make the encode_images and encode_texts consistent with https://huggingface.co/jinaai/jina-clip-implementation/blob/main/modeling_clip.py#L384
nan
changed pull request status to
open
Shouldn't we return a single tensor, instead of a List(tensor), if a single text/image is passed?
Shouldn't we return a single tensor, instead of a List(tensor), if a single text/image is passed?
yep, makes sense. I've fixed it.
nan
changed pull request status to
merged