🚀 Introducing Hy-Nomas: An Open Source Free-Form Interactive Video World Model

Tencent introduced Hy-Nomas, an open source world model that generates interactive video from free-form text prompts and camera trajectories.

• The model is built for long-horizon generation with world memory, so scenes stay consistent as a session runs on.
• Tencent says lightweight LoRA adaptation lets the model be tuned for specialized domains such as robotics.
• It is optimized for inference on consumer GPUs, and Tencent published a website, GitHub repository, page and demo for people to test it and send feedback.