AI Paper Atlas

Research map / Video / Video-language understanding

FlashVID: Efficient Video Large Language Models via Training-free Tree-based Spatiotemporal Token Merging

Ziyang Fan, Keyu Chen, Ruilong Xing, Yulin Li, Li Jiang, Zhuotao Tian

ICLR 2026 · Video › Video-language understanding · approach: Video LLMs with fine-grained understanding · 16 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use ADTS (attention + diversity) selection instead of attention-only (ATS) (ablation)FlashVID · 4 video benchmarks, R=10%relative accuracy: 96.9 -> 99.1Improved
use ADTS instead of diversity-only (DTS) (ablation)FlashVID · 4 video benchmarks, R=10%relative accuracy: 97.1 -> 99.1Improved
add [CLS] attention calibration to diversity selection (ablation)ADTS · 4 video benchmarks, R=10%relative accuracy: 97.1 -> 98.5Improved
add event relevance calibration to diversity selection (ablation)ADTS · 4 video benchmarks, R=10%relative accuracy: 97.1 -> 97.6Improved
combine ADTS and TSTM (alpha=0.7) rather than TSTM alone (ablation)FlashVID · 4 video benchmarks, R=10%relative accuracy: 97.4 -> 99.1Improved
combine ADTS and TSTM (alpha=0.7) rather than ADTS alone (ablation)FlashVID · 4 video benchmarks, R=10%relative accuracy: 98.5 -> 99.1Improved
constrain spanning tree depth and breadth in TSTM (ablation)TSTM · video benchmarksaccuracyNo clear effect
use FlashVID instead of FastVID at 10% retentionLLaVA-OneVision · video benchmarks avgrelative accuracy: 97.8 -> 99.1Improved
use FlashVID instead of VisionZip at low retentionLLaVA-OneVision · video benchmarks avgrelative accuracy: 92.5 -> 99.1 (R=10%)Improved
prune tokens with FlashVID at 25% retention vs full tokensLLaVA-OneVision · video benchmarks avgrelative accuracy: 100.0 -> 100.3Improved
extend input from 16 to 160 frames with FlashVID under same token budgetQwen2.5-VL · 5 video benchmarks avgrelative accuracy: 100.0 -> 108.6Improved
use FlashVID instead of FastVID on LLaVA-Video at 20% retention, VideoMME overallLLaVA-Video · VideoMME Overallaccuracy: 62.6 (FastVID) vs 62.2 (FlashVID)Hurt

Later papers that compared with this method

2 later accepted papers report FlashVID as a baseline in their results tables; 2 of them report a better number on most of the results they share with it.

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy