Generated images donot match the prompt
#10
by zky1 - opened
Great works! I try this model in diffusers but I found this model doesnot generate correct color as I mentioned in the prompt. Such as I wrote "white stones, green grapes" in the prompt but the model generates grey stones and A bunch of carrots. However, the lines as matched as input. I test the canny and softedge model, they perfomed better. Why doe this happen and any solution? Thanks a lot.
You might need to check the base model that you used to generate. I believe this has nothing to do with the controlnet.
StrugglerXYH changed discussion status to closed