News for hackers. Stuff that matters.

A comment on Reverse engineer Stable Diffusion images. Parent comment

Re:Reverse engineer Stable Diffusion images by rf15 on Wednesday February 08, 2023 @04:38AM (#34706055)
This sounds a lot like the results you get from CLIP interrogation. Maybe they just use that and made another online service for it?
[ Parent | on HN ]
Re:Reverse engineer Stable Diffusion images by sahil_chaudhary on Wednesday February 08, 2023 @07:05AM (#34706972)
It is actually based on a different approach, it uses an image-captioning model finetuned on image-prompt pairs
[ Parent | on HN ]
Re:Reverse engineer Stable Diffusion images by xkapastel on Wednesday February 08, 2023 @08:39AM (#34707820)
CLIP Interrogator uses BLIP, an image captioning model, as well as trying a bunch of prompts with CLIP.

Is the code for this one available?

[ Parent | on HN ]
Re:Reverse engineer Stable Diffusion images by sahil_chaudhary on Wednesday February 08, 2023 @10:58AM (#34709869)
Ah yes, this model treats this purely as image captioning. The model isn't open source yet.
[ Parent | on HN ]