Search papers, labs, and topics across Lattice.
This paper introduces RefineCut, a novel open-weight planner designed for executable video-editing planning, which integrates a deterministic verifier to ensure adherence to constraints during the editing process. By employing a multi-teacher branching strategy and a closed verifier loop, RefineCut achieves significant improvements in video-editing performance, reaching a Video-Editing Score of 0.924 on the RefineCut-Bench dataset. The results demonstrate that the planner not only outperforms traditional workflow systems but also matches or exceeds the performance of larger frontier models like Llama-3.1-8B and GLM-4-9B without requiring teacher calls during inference.
RefineCut's innovative verifier-grounded approach elevates video-editing performance to new heights, achieving a score of 0.924 while eliminating the need for teacher calls at inference.
Practical video editing is not only pixel generation: an editor must turn a brief, a clip pool, music metadata, and hard constraints into an executable timeline. We study this decision layer as \emph{executable video-editing planning} and introduce RefineCut, which, unlike workflow systems that wrap a prompted frontier model, trains a compact open-weight planner for it. The planner edits a typed timeline through structured patches covering clip selection, trimming, ordering, transitions, and duration and music alignment; a deterministic verifier applies each patch and checks it against an explicit constraint ledger. Because editing has no single ground-truth repair, we do not imitate teachers directly: RefineCut replays every multi-teacher branch through the verifier and keeps verifier-best repairs as supervision. A second stage, RefineCut-Evo, lets the student score its own repairs with the verifier and a task rubric and trains on high-margin preference pairs, so the final $8$B planner runs in a closed verifier loop with no teacher calls at inference. On RefineCut-Bench ($3{,}578$ tasks, $7{,}971$ captioned clips, $499$ music tracks, explicit ledgers), verifier-replayed distillation lifts the planner from $0.620$ to $0.858$ on the protocol-specific Video-Editing Score and RefineCut-Evo reaches $0.924$; the gain transfers to Llama-3.1-8B and GLM-4-9B, and in the same closed loop the $8$B planner matches or exceeds its frontier teachers. Code and RefineCut-Bench are publicly released; see the Data Availability statement.
