add sentencepiece as a soft dependency #9065
Conversation
|
The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update. |
sayakpaul
left a comment
sayakpaul
left a comment
There was a problem hiding this comment.
Thanks. Just a single comment.
| is_note_seq_available, | ||
| is_onnx_available, | ||
| is_scipy_available, | ||
| is_sentencepiece_available, |
There was a problem hiding this comment.
Actually sentencepiece is needed by all the pipelines that need T5. Should we add them here as well?
There was a problem hiding this comment.
technically, they can use these pipelines without t5, no? i.e. use pre-computed text embeddings
| KolorsPipeline, | ||
| UNet2DConditionModel, | ||
| ) | ||
| from diffusers.pipelines.kolors import ChatGLMModel, ChatGLMTokenizer |
There was a problem hiding this comment.
I think we already install sentencepiece for tests. So should be good:
Line 126 in b1f43d7
There was a problem hiding this comment.
yeah, that was the main reason I didn't catch this
* add sentencepiece as soft dependency for kolors * up --------- Co-authored-by: Sayak Paul <[email protected]>
fix #9034