cs.CVNov 1, 2026

FinishStrong: Learning to End Well after Imperfect Start in Video Temporal Grounding

Authors: Yuda Zou, Boxiang Zhou, Xin Zhou, Yibo Chen, Dejia Song, Xu Tang, Yao Hu, Yongchao Xu

Organizations: School of Computer Science, Wuhan University · Xiaohongshu Inc.

Abstract

A poor start lowers the ceiling, but does not negate the value of finishing well. In autoregressive video temporal grounding, models systematically exhibit an endpoint performance asymmetry, predicting ends significantly worse than starts. A committed flawed start acts as an immutable physical constraint, irreversibly capping the attainable overlap as a sunk loss. While models should minimize the remaining conditional regret, standard post-training paradigms fail to teach this adaptation. Specifically, teacher-forced SFT suffers from strict prefix exposure bias, and holistic trajectory evaluation in standard GRPO inadvertently penalizes conditionally optimal ends if they follow a poor start. To address this, we introduce FinishStrong, a novel two-stage post-training framework explicitly designed to teach optimal continuations. First, Online-start SFT exposes the end prediction to the model's own generated contexts. Second, Conditional-tree GRPO restructures the sampling topology, centering end rewards exclusively within the same-start branch to mathematically isolate the conditional end quality from the inherited start penalty. FinishStrong-9B achieves a state-of-the-art 61.7% overall IoU on TimeLens-Bench, remarkably outperforming flagship closed-source models like Gemini-3.8-Flash and Qwen3.8-Max. Crucially, when forced to continue from imperfect starts, it reduces the subsequent end prediction error by 36.5%, proving its exceptional ability to end well after flawed starts. Code will be open.

Figures & tables

Appendix figures & tables15 assets

Supplementary material from the paper’s appendix.

Appendix

Explore similar work

CardsList
  1. A Paradigm Shift: Fully End-to-End Training for Temporal Sentence Grounding in Videos

    Apr 3, 2026Allen He, Qi Liu, Kun Liu +2Temporal Video GroundingParameter-Efficient Fine-Tuning

  2. TimeLens2: Generalist Video Temporal Grounding with Multimodal LLMs

    Jul 19, 2026Yuhan Zhu, Changlian Ma, Xiangyu Zeng +12Temporal Video GroundingReinforcement Learning

  3. Temporal-Aware Reasoning Optimization for Video Temporal Grounding

    Jun 8, 2026Minghang Zheng, Zihao Yin, Yi Yang +2Temporal Video GroundingTemporal Reasoning