TinyDéjàVu让传感器神经网络推理内存降90%，速度不变。
TinyDéjàVu: Smaller RAM and Faster Inference with Neural Networks on MCUs for Sensor Data Streams
- 通过优化数据流，重用中间结果降低内存占用
- 在嵌入式MCU上实现比之前方法少90%的内存使用
- 适合资源受限的物联网设备实时推理场景
嵌入式智能应用广泛采用小型神经网络部署在无线传感器和执行器上，需持续对感知的时间序列数据进行推理。为满足电池供电下的寿命与能耗要求，此类硬件通常采用极低内存的微控制器（如128 kB RAM）。在此背景下，优化神经网络各层间的数据流动至关重要。本文提出新框架TinyDéjàVu及配套算法，显著降低典型微控制器硬件上针对传感器时序数据的神经网络推理所需内存预算。我们开源了实现，并在常见微控制器（Arm Cortex-M）上进行了可复现的基准测试。结果表明，对于重叠滑动窗口输入，TinyDéjàVu相比先前工作（StreamiNNC）在计算延迟相同的情况下，可节省高达90%的内存使用。
原文摘要 · Abstract (English)
Examples of embedded intelligence include a wide variety of tiny neural networks used on-board wireless sensors and actuators, which are expected to continuously perform inference on time-series of the data they sense. In order to fit lifetime and energy consumption requirements when operating on battery, such hardware is exclusively based on microcontroller with as little memory as possible, e.g., 128 kB of RAM. In this context, optimizing data flows during inference across neural network layers becomes crucial. In this paper, we introduce a new framework, TinyDéjàVu, and novel algorithms we designed to drastically reduce the RAM budget required by inference using various neural network models for sensor data time-series on typical microcontroller hardware. We publish the implementation of TinyDéjàVu as open source, and we perform reproducible benchmarks on common microcontroller hardware (Arm Cortex-M). We show that TinyDéjàVu can save up to 90\% of RAM usage with equal compute latency compared to prior work (StreamiNNC) on overlapping sliding window inputs.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。