SOTA pipeline for fully automatic conditional image generation

Hello,

Would you mind clarifying the paragraph in the README about "a SOTA pipeline for fully automatic conditional image [generation]"?

Note that in the guess mode, you will still be able to input prompts. The only difference is that the model will "try harder" to guess what is in the control map even if you do not provide the prompt. Just try it yourself!

Besides, if you write some scripts (like BLIP) to generate image captions from the "guess mode" images, and then use the generated captions as prompts to diffuse again, you will get a SOTA pipeline for fully automatic conditional image generating.

I am not sure if I understand the suggestion. My understanding is that it would:

1. generate an image in "guess" mode, i.e. without any text prompt
2. generate a caption for the generated image
3. feed the text prompt along an image, but which one: the original input image or the generated image?

Moreover, you mention several images. Is the point to generate several captions to feed along the original input image?

Start with the quoted README paragraph and the linked captioning example, then trace the described guess-mode and captioning steps. Clarify whether the second diffusion pass uses the original or generated image, and explain the reference to several images; the README is done when those questions are answered unambiguously.