[Submitted on 2 Nov 2026 (v1), last revised 19 Nov 2026 (this version, v4)]
Abstract:Complex video editing requires understanding complex instructions while jointly fulfilling multiple editing requests. Meeting these demands through dense model scaling substantially increases computational cost. We introduce VideoEditor-MoE, the first open-source mixture-of-experts framework for video editing. Beyond sparse MLP-MoE layers, MoE-Attention expands attention capacity through routed key–value experts while preserving global interactions among video tokens. Initialized from a 5B video editing model, VideoEditor-MoE scales to 14B and 56B parameters, with approximately 8B parameters activated per token in both variants. Comparisons with existing methods and two complementary baselines, Explicit-MoE and MLP-MoE, assess its benefits over explicit expert guidance and MLP-only expert architectures. Further analyses demonstrate that all experts remain active and reveal that learned routing responds more strongly to sample content and diffusion noise levels than to human-defined editing categories. Spatial analysis further indicates sensitivity to object-level semantic features. We also identify decoupled weight decay as a strategy for stabilizing expert scaling.
Submission history
From: Xiao Cao [view email]
[v1] Mon, 2 Nov 2026 12:42:51 UTC (6,654 KB)
[v2] Mon, 9 Nov 2026 13:11:07 UTC (6,658 KB)
[v3] Wed, 11 Nov 2026 11:48:20 UTC (6,655 KB)
[v4] Thu, 19 Nov 2026 10:12:22 UTC (6,787 KB)
