Model generates variable-length video segments from user inputs
The model generates variable-length segments conditioned on current inputs and cached history, with structured video captions aligned to action-aware causal attention.
A vision-language component converts user requests and multimodal interaction histories into structured conditions for generating the next segment.
The researchers made available model weights, code, benchmarks, and a lightweight adaptation workflow to enable reproducible research on interactive world generation.
Sources behind this report
- Hugging Face
Research · abstract · CC0 1.0
Source published
Read by the newsroom
Report history
No earlier revision dates are retained here. This does not establish that the report has never changed; read any dated updates and corrections.
This brief was made by AI and checked against its cited public sources. How the newsroom works