Wednesday, April 24, 2024

A system for producing 3D level clouds from advanced prompts

Must read

Whereas latest work on text-conditional 3D object technology has proven promising outcomes, the state-of-the-art strategies usually require a number of GPU-hours to supply a single pattern. That is in stark distinction to state-of-the-art generative picture fashions, which produce samples in quite a lot of seconds or minutes. On this paper, we discover an alternate technique for 3D object technology which produces 3D fashions in solely 1-2 minutes on a single GPU. Our technique first generates a single artificial view utilizing a text-to-image diffusion mannequin, after which produces a 3D level cloud utilizing a second diffusion mannequin which circumstances on the generated picture. Whereas our technique nonetheless falls in need of the state-of-the-art when it comes to pattern high quality, it’s one to 2 orders of magnitude quicker to pattern from, providing a sensible trade-off for some use circumstances. We launch our pre-trained level cloud diffusion fashions, in addition to analysis code and fashions, at this https URL.

Supply hyperlink

More articles


Please enter your comment!
Please enter your name here

Latest article