GS-Voxel: Fitting-Free Structured Latents for Large-Scale 3DGS Generation

acvlab · hf · 2026-08-19

New project on Hugging Face: GS-Voxel tackles large-scale 3D Gaussian Splatting generation by converting unstructured 3D Gaussian reconstructions into sparse structured latents, removing the need for per-scene fitting. On top of this representation, flow models enable scalable aerial scene generation.

Original post →

More from Multimodal

Multimodal channel →