World Labs releases real-time generative world model (RTFM): a single H100 GPU can generate 3D worlds in real time
-
最新报道：According to a report on Bijie.com on October 16th, WorldLabs, a spatial intelligence company founded by Stanford University Professor Fei-Fei Li, officially released a research preview of its next-generation real-time generative model, RTFM (Real-Time Frame Model). This model can generate videos and 3D scenes in real time during interaction, enabling persistent virtual worlds. RTFM is designed around three core principles: efficiency, scalability, and persistence. The model can run at interactive frame rates on a single NVIDIA H100 GPU and can generate scenes with complex effects such as reflections, shadows, and gloss from a single image without explicit 3D modeling. Its "spatial memory" mechanism maintains world persistence through frame pose modeling and "context juggling," supporting long-term user interaction and exploration.
下一篇
Ripple announces $1 billion acquisition of GTreasury, a financial management system company
- According to Bijie.com, on October 16th, Ripple announced the acquisition of GTreasury, a ...
24小时热点
热点专题
418366