Pith. sign in

REVIEW 4 cited by

DynTok: Dynamic Compression of Visual Tokens for Efficient and Effective Video Understanding

Not yet reviewed by Pith; the record is open.

Re-run · record.json Download PDF Read on arXiv ↗

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2506.03990 v1 pith:FVZZXBKP submitted 2025-06-04 cs.CL cs.CV

Hongzhi Zhang , Jingyuan Zhang , Xingguang Ji , Qi Wang , Fuzheng Zhang This is my paper
classification cs.CLcs.CV
keywords videovisualtokenscompressiondyntokeffectivenumberbackbone
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

Typical video modeling methods, such as LLava, represent videos as sequences of visual tokens, which are then processed by the LLM backbone for effective video understanding. However, this approach leads to a massive number of visual tokens, especially for long videos. A practical solution is to first extract relevant visual information from the large visual context before feeding it into the LLM backbone, thereby reducing computational overhead. In this work, we introduce DynTok, a novel \textbf{Dyn}amic video \textbf{Tok}en compression strategy. DynTok adaptively splits visual tokens into groups and merges them within each group, achieving high compression in regions with low information density while preserving essential content. Our method reduces the number of tokens to 44.4% of the original size while maintaining comparable performance. It further benefits from increasing the number of video frames and achieves 65.3% on Video-MME and 72.5% on MLVU. By applying this simple yet effective compression method, we expose the redundancy in video token representations and offer insights for designing more efficient video modeling techniques.

Discussion (0). Sign in to comment.

Forward citations

Cited by 4 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. FlexiSLM: A Dynamic and Controllable Frame Rate Spoken Language Model

    cs.SD 2026-06 unverdicted novelty 7.0 of 10

    FlexiSLM is the first spoken language model supporting dynamic and controllable frame rates on speech input and output, outperforming fixed-rate 7B models at high quality and enabling faster inference at lower rates l...

  2. CRAFT: Compression via Recursive Adaptive Fusion of Video Tokens for Vision-Language Models

    cs.CV 2026-08 conditional novelty 6.0 of 10

    CRAFT recursively merges video tokens with training-free similarity selection plus learnable gated fusion, retaining ~97% of average accuracy at 8x compression across six benchmarks.

  3. Stateful Token Reduction for Long-Video Hybrid VLMs

    cs.CV 2026-02 conditional novelty 6.0 of 10

    For hybrid Mamba–Transformer video models, keeping 25% of visual tokens with a query-based progressive schedule gives 3.8–4.2x prefilling speedups with near-baseline accuracy; the paper attributes this to stateful com...

  4. ViCoStream: Streaming VideoLLMs Can Run Beyond 100 FPS with Stage-Wise Coordinated Inference

    cs.CV 2026-06 unverdicted novelty 5.0 of 10

    ViCoStream is a new coordinated pipeline framework for streaming VideoLLMs that achieves 134 FPS video throughput and less than 50 ms TTFT on A100 while keeping accuracy near full-history baselines.

Pith tools