Read the original at HF Daily Papers
Researchers introduce WorldSonus, a framework that generates real-time spatial audio for interactive world models using a streaming causal autoregressive diffusion architecture.
Carried by: HF Daily Papers. First seen: .