Closed K-O-N-B closed 3 months ago
yes the assistant has a system prompt that guides the description to be more like SD prompts you need to have a model with good reasoning normally you just need to ask something like "make a SD prompt for the image" and use a vision model for it to work