← 返回 JSSC 论文列表
📄 下载 JSSC 原文 PDF
JSSC 2022第10期Digital Circuits28nm

Trainer An Energy-Efficient Edge-Device Training Processor Supporting Dynamic Wei

提出一种支持动态权重剪枝的能效边缘设备训练处理器Trainer
28nm CMOS, 20.96mm²面积
边缘计算稀疏训练能效优化动态剪枝批归一化
推测机制消除隐式冗余操作
动态稀疏自适应数据流解决重用不平衡
计算依赖解耦的批归一化单元减少重复数据访问
Abstract
Transfer learning, which transfers knowledge from source datasets to target datasets, is practical for adaptive deep neural network (DNN) applications. When considering user pri- vacy and communication bandwidth issues, edge devices’ training is essential for transfer learning. Nevertheless, training requires repeating feedforward (FF), backpropagation (BP), and weight gradient (WG) millions of times, introducing prohibitive compu- tation for edge devices. A promising method to reduce training c