FPGA加速下，0.36秒完成CNN微调，适合物联网实时数据适应。
InstantFT: An FPGA-Based Runtime Subsecond Fine-tuning of CNN Models
- 基于FPGA优化前向与反向计算，实现参数高效微调。
- 比LoRA方法快17.4倍，仅需0.36秒完成微调，精度相当。
- 能效提升16.3倍，适合资源受限设备的在线适应。
训练深度神经网络（DNN）所需的计算和内存远超推理，使资源受限的物联网平台难以实现实时模型调整。本文提出InstantFT，一种基于FPGA的超快速卷积神经网络（CNN）微调方法，通过优化参数高效微调（PEFT）中的前向与反向计算。在存在概念漂移的数据集上的实验表明，InstantFT将预训练CNN的微调速度提升至现有基于低秩适配（LoRA）方法的17.4倍，同时保持相当的准确率。基于FPGA的InstantFT将微调时间缩短至仅0.36秒，能效提升16.3倍，支持在非平稳数据分布下对CNN进行即时适应。
原文摘要 · Abstract (English)
Training deep neural networks (DNNs) requires significantly more computation and memory than inference, making runtime adaptation of DNNs challenging on resource-limited IoT platforms. We propose InstantFT, an FPGA-based method for ultra-fast CNN fine-tuning on IoT devices, by optimizing the forward and backward computations in parameter-efficient fine-tuning (PEFT). Experiments on datasets with concept drift demonstrate that InstantFT fine-tunes a pre-trained CNN 17.4x faster than existing Low-Rank Adaptation (LoRA)-based approaches, while achieving comparable accuracy. Our FPGA-based InstantFT reduces the fine-tuning time to just 0.36s and improves energy-efficiency by 16.3x, enabling on-the-fly adaptation of CNNs to non-stationary data distributions.
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。