Does the colossalAI support to train conditional model based on image condition rather than text?

[pokameng]

Hello! Does the colossalAI support to train conditional model based on image condition rather than text? @Fazziekey The configs in colossalAI are based on text to image, but i want to train a new model based on image.

[Fazziekey]

It‘s ok to train from image condition, but you need to config the right model and input

[pokameng]

It‘s ok to train from image condition, but you need to config the right model and input

Ok can you share the example config with me?

[pokameng]

It‘s ok to train from image condition, but you need to config the right model and input

@Fazziekey

[Fazziekey]

sorry, I don't have an example, but welcome to contribute a example for us