AI Dynamics

Global AI News Aggregator

About

New Multimodal Transformer-Based Architecture for 3D Avatar Generation

LHM leverages: A multimodal transformer 3D Gaussian splatting A novel body-head architecture Zero post-processing No fancy capture rigs. No manual tweaks. Just feed it a single image → get high-fidelity 3D avatars in seconds. This is going to change everything!

→ View original post on X — @godofprompt