Search papers, labs, and topics across Lattice.
2
0
4
0
Editing long videos with multiple instructions can be done without hallucinations or loss of temporal continuity, thanks to a novel agentic framework that combines LLMs and VLMs.
CoinVE-200K enables video editing models to understand and execute complex, multi-faceted editing instructions with unprecedented accuracy and quality.
