[REQUEST] Comfy workflow for prompt generation from image input
#17
by
jspsoli
- opened
Since this effort is fully open source, it would be nice to have a worflow that includes all of the used tools (and their config) to generate prompts based on a given image - the same way it was done for this model's training data.
This way we can easily replicate existing images using the best prompts for this model and if nothing else it would allow us to better understand what prompts work best and all kinds of tricks we can use for better results.