Sparrow: Text-Anchored Window Attention with Visual-Semantic Glimpsing for Speculative Decoding in Video LLMs

TL;DR

This paper introduces Sparrow, a novel framework that enhances speculative decoding in Video Large Language Models by leveraging visual-semantic internalization, resulting in significant speedups and improved performance on long video sequences.

Contribution

Sparrow employs text-anchored window attention and semantic-rich intermediate states to mitigate attention dilution and visual noise, enabling efficient long video inference in Vid-LLMs.

Findings

Achieves 2.82x speedup on long video sequences

Resolves performance degradation in long sequences

Effectively handles 25k visual tokens in real-time tasks

Abstract

Although speculative decoding is widely used to accelerate Vision-Language Models (VLMs) inference, it faces severe performance collapse when applied to Video Large Language Models (Vid-LLMs). The draft model typically falls into the trap of attention dilution and negative visual gain due to key-value cache explosion and context window mismatches. We observe a visual semantic internalization phenomenon in Vid-LLMs, indicating that critical visual semantics are implicitly encoded into text hidden states during deep-layer interactions, which renders raw visual inputs structurally redundant during deep inference. To address this, we propose the Sparrow framework, which first utilizes visually-aware text-anchored window attention via hidden state reuse to fully offload visual computation to the target model, and leverages intermediate-layer visual state bridging to train the draft model with semantic-rich intermediate states, thereby filtering out low-level visual noise. Additionally, a multi-token prediction strategy is introduced to bridge the training-inference distribution shift. Experiments show that Sparrow achieves an average speedup of 2.82x even with 25k visual tokens, effectively resolving the performance degradation in long sequences and offering a practical solution for real-time long video tasks.