# Video-OPD: Efficient Post-Training of Multimodal Large Language Models for Temporal Video Grounding via On-Policy Distillation Source: [Video-OPD: Efficient Post-Training of Multimodal Large Language Models for Temporal Video Grounding via On-Policy Distillation]() Collection index: [llms.txt]() Data updated: 2026-06-04T05:00:05.280Z Authors: Jiaze Li, Hao Yin, Haoran Xu, Boshen Xu, Wenhui Tan, Zewen He, Jianzhong Ju, Zhenbo Luo, Jian Luan Published: 2026-02-03T02:05:48.000Z ## Abstract Reinforcement learning has emerged as a principled post-training paradigm for Temporal Video Grounding (TVG) due to its on-policy optimization, yet existing GRPO-based methods remain fundamentally constrained by sparse reward signals and substantial computational overhead. We propose Video-OPD, an efficient post-training framework for TVG inspired by recent advances in on-policy distillation. Video-OPD optimizes trajectories sampled directly from the current policy, thereby preserving alignment between training and inference distributions, while a frontier teacher supplies dense, token-level supervision via a reverse KL divergence objective. This formulation preserves the on-policy property critical for mitigating distributional shift, while converting sparse, episode-level feedback into fine-grained, step-wise learning signals. Building on Video-OPD, we introduce Teacher-Validated Disagreement Focusing (TVDF), a lightweight training curriculum that iteratively prioritizes trajectories that are both teacher-reliable and maximally informative for the student, thereby improving training efficiency. Empirical results demonstrate that Video-OPD consistently outperforms GRPO while achieving substantially faster convergence and lower computational cost, establishing on-policy distillation as an effective alternative to conventional reinforcement learning for TVG. ## Recorded benchmark results A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. - [Temporal Video Grounding on Charades-TimeLens (test)](): Temporal Video Grounding; Charades-TimeLens (test); R@0.3 (IoU=0.3): 73.1. - [Temporal Video Grounding on QVHighlights TimeLens (test)](): Temporal Video Grounding; QVHighlights TimeLens (test); Recall @ IoU=0.3: 73.8. - [Temporal Video Grounding on ActivityNet TimeLens (test)](): Temporal Video Grounding; ActivityNet TimeLens (test); Recall@0.3: 60.5. ## Original sources and implementation [Original paper]() [PDF]()