Version 2.0 incorporates textual and visual knowledge into the diffusion model, improving the quality of images.

80lv's platform is a resource for game developers and 3D artists, offering insights into game development techniques, industry trends, and visual effects (VFX). Through articles, interviews, and case studies, 80lv offers insights into 3D modeling, texturing, animation, and rendering. Developers and artists can learn about game engines, shader programming, and advanced rendering techniques to create immersive and visually stunning games and animations.

80 LEVEL

Chinese text-to-image diffusion model ERNIE-ViLG got upgraded to version 2.0. It can scale up the model to 24 billion parameters, which is 10 times more than in Stable Diffusion.

ERNIE-ViLG 2.0: the Biggest Text-to-Image Model with 24B Parameters