WorldFM Open Source: Real-Time Multi-View Diffusion from Target Poses

tom_doerr · x · 2026-08-14

A new project named WorldFM has been open-sourced on GitHub. It is a real-time multi-view diffusion model that generates images at new viewpoints given a single reference image and target camera poses.

The repository includes complete setup scripts and integrates submodules like HunyuanWorld-1.0, MoGe, Real-ESRGAN, and ZIM, allowing developers to quickly build the environment and run the inference pipeline.

Original post →

More from Multimodal

Multimodal channel →