ComfyUI icon indicating copy to clipboard operation
ComfyUI copied to clipboard

Published 20 hours ago • verified publisher icon comfyanonymous

[Feature Request] Add Kandinsky support

Open jexom opened this issue 1 year ago • 5 comments

Add support for loading Kandinsky models, which allow for multilanguage prompts and image fusion https://github.com/ai-forever/Kandinsky-2 https://huggingface.co/ai-forever/Kandinsky_2.1

jexom avatar Apr 12 '23 07:04 jexom

Is this only for the russian language? I tried in french with their website and it does not seem to work for french. Maybe I missed a language setting somewhere?

PladsElsker avatar Apr 12 '23 17:04 PladsElsker

Is this only for the russian language? I tried in french with their website and it does not seem to work for french. Maybe I missed a language setting somewhere?

It works well in English

Gorislav avatar Apr 13 '23 19:04 Gorislav

Great suggestion and I fully support it! I wanted to make a similar request. I would really like to see Kandinsky 2.1 in ComfyUI

Gorislav avatar Apr 13 '23 19:04 Gorislav

Is this only for the russian language? I tried in french with their website and it does not seem to work for french. Maybe I missed a language setting somewhere?

They claim that the dataset was multilingual, that said it could mean just 2 languages. The model is still evolving so they may add more languages sooner or later

jexom avatar Apr 14 '23 05:04 jexom

ANy update?

MrlolDev avatar May 28 '23 11:05 MrlolDev