Existing approaches for improving the efficiency of Large Vision-Language Models (LVLMs) are largely based on the concept of visual token reduction. This approach, however, creates an information bottleneck that impairs performance, especially on challenging tasks that require fine-grained understanding and reasoning. In this work, we challenge this paradigm by introducing VISion On Request (VISOR), a method that reduces inference cost without discarding visual information. Instead of compressing the image, VISOR improves efficiency by sparsifying the interaction between image and text tokens. Specifically, the language model attends to the full set of high-resolution visual tokens through a small, strategically placed set of attention layers: general visual context is provided by efficient cross-attention between text-image, while a few well-placed and dynamically selected self-attention layers refine the visual representations themselves, enabling complex, high-resolution reasoning when needed. Based on this principle, we first train a single universal network on a range of computational budgets by varying the number of self-attention layers, and then introduce a lightweight policy mechanism that dynamically allocates visual computation based on per-sample complexity. Extensive experiments show that VISOR drastically reduces computational cost while matching or exceeding state-of-the-art results across a diverse suite of benchmarks, and excels in challenging tasks that require detailed visual understanding.

翻译：现有提升大型视觉语言模型效率的方法主要基于视觉令牌压缩理念。然而，这种策略会造成信息瓶颈，在需要精细理解与推理的复杂任务中损害模型性能。本研究提出按需视觉交互（VISOR）方法，通过在不丢弃视觉信息的前提下降低推理成本，挑战了现有范式。VISOR并非压缩图像，而是通过稀疏化图像与文本令牌之间的交互提升效率。具体而言，语言模型通过少量精心布置的自注意力层关注完整高分辨率视觉令牌集：通用视觉语境通过高效的图文跨注意力机制提供，而少量精心选择且动态调用的自注意力层则用于优化视觉表征本身，从而在需要时实现复杂高分辨率推理。基于该原理，我们首先通过改变自注意力层数量，在多个计算预算下训练单一通用网络，随后引入轻量级策略机制，根据样本复杂度动态分配视觉计算资源。大量实验表明，VISOR在显著降低计算成本的同时，在多样化基准测试中达到或超越最先进水平，尤其在需要精细视觉理解的复杂任务中表现卓越。