On: Real-time image editing using latent consistency models

[mungoman2]

Come on, this is the wrong attitude. What they asked for could be prototyped in a day to see if it makes sense. Have the left image in the demo be fed from a rendering from a simple 3d scene of a camera and simple colored primitives.

[dvrp → mungoman2]

I was being sarcastic! I’m super excited too!

[stavros → dvrp]

Now make it 5D.

[dvrp → stavros]

Introducing Tesseract Diffusion.