VAST releases EdenLive: 5B real-time video model rendering at 24+ FPS on a single GPU
AdinaYakup · x · 2026-10-18
VAST-AI Research has released EdenLive, a new video model for real-time generative rendering from any visual reference. Key specs: 24+ FPS rendering with minute-scale generation, 480P/720P output, and a 5B variant that runs on a single GPU. Built on WAN 2.1/2.2, it takes a proxy/reference video stream, a target first frame, and a text prompt, with control injected via a post-trained DINOv3 vision encoder conditioning the diffusion transformer. Weights (about 210GB across 14B/5B few-step and bidirectional DiT variants) are on Hugging Face; upstream code maintains interaction state while the model handles visual rendering.
Related event: VAST Open-Sources EdenLive Real-Time Generative Renderer(3 posts)→
More from Embodied
- China installs 9x more industrial robots than US, and DC is worried — binarybits · 2026-10-18
- Giving the rabbit r1 a second life with llama.cpp and on-device s2s apps — mervenoyann · 2026-10-18
- Boston Dynamics' brutal lesson: Handle out, Stretch in as DHL signs for 1,000+ robots — TinfoilTricorn · 2026-10-18
- Stanford AA 274A Principles of Robot Autonomy lectures by Marco Pavone are freely available — caglar_ee · 2026-10-18
- Meta glasses trick: AI running route plus Google Maps Ask AI equals instant route map — IanAndrewsDC · 2026-10-18
- Dongyi Robotics shows 100x efficiency gains with 4-layer embodied AI stack at Canton Fair — 智东西 · 2026-10-18