VLM即刻：通过神经元分块实现高效的I/O压缩
VLM in a flash: I/O-Efficient Sparsification of Vision-Language Model via Neuron Chunking
AI总结：
本文提出Neuron Chunking方法，通过神经元分块实现高效的I/O压缩，提升边缘部署VLMs的性能。
摘要：
在边缘部署大型视觉-语言模型（VLMs）时，越来越多地依赖基于闪存的权重卸载，其中激活稀疏化用于减少I/O开销。然而，传统稀疏化方法仍以模型为中心，仅根据激活幅度选择神经元，忽略了访问模式对闪存性能的影响。我们提出了Neuron Chunking，一种高效的I/O稀疏化策略，该策略在内存中以块（即连续神经元的组）为单位进行操作，并将神经元的重要性与存储访问成本相结合。该方法通过轻量级的访问连续性抽象建模I/O延迟，并选择具有高实用性的块，定义为神经元重要性归一化后的估计延迟。通过将稀疏化决策与底层存储行为对齐，Neuron Chunking在Jetson Orin Nano和Jetson AGX Orin上分别将I/O效率提高了4.65倍和5.76倍。
英文摘要：
Edge deployment of large Vision-Language Models (VLMs) increasingly relies on flash-based weight offloading, where activation sparsification is used to reduce I/O overhead. However, conventional sparsification remains model-centric, selecting neurons solely by activation magnitude and neglecting how access patterns influence flash performance. We present Neuron Chunking, an I/O-efficient sparsification strategy that operates on chunks (i.e., groups of contiguous neurons in memory) and couples neuron importance with storage access cost. The method models I/O latency through a lightweight abstraction of access contiguity and selects chunks with high utility, defined as neuron importance normalized by estimated latency. By aligning sparsification decisions with the underlying storage behavior, Neuron Chunking improves I/O efficiency by up to 4.65x and 5.76x on Jetson Orin Nano and Jetson AGX Orin, respectively.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
