Nodes for Stable Diffusion

Question. Is Vuo in a good position with its front end to implement Stable Diffusion? Similar to Comfy UI. If it is. That may be a good FR.

So, at the most basic level, I think this feature request would be for a node that inputs a text phrase and outputs an image created by generative AI. Is that what you had in mind, @krezrock?

Currently Apple’s implementation of Stable Diffusion takes about 20-46 seconds to generate a 1024x1024 image on recent Mac hardware. Would you still find this a potentially useful Vuo feature at that speed, or would it be better to wait until the technology is to the point where you could use it more effectively for real-time generation of visuals?

Do you know if there are any alternatives to Stable Diffusion that are trained on images that the artists granted permission to use?

That is correct, at a basic level.

Great point in regard to Macs processing these models. Maybe the best approach would be nodes to access API’s. I believe there are ways to use the API’s to access external GPU’s instead of using locally.

You bring up a very big issue with Stable Diffusion (and image diffusion in general). The only alternative that is supposed to be ethically trained is Adobe’s Firefly. They have not release API access.