基于FPGA的深度强化学习硬件加速技术分析
Analysis of Hardware Acceleration Technology for Deep Reinforcement Learning Based on FPGA
刘峥嵘1
作者信息
- 1. 大连理工大学 微电子学院,辽宁 116001
- 折叠
摘要
阐述语音识别、图像识别处理、优化策略和硬件模板中的FPGA加速器技术特点,以及在FPGA上加速深度学习的可行性,分析FPGA加速器的应用,探讨FPGA加速器加速深度强化学习的硬件模板、性能和功耗水平,提出针对FPGA加速器的训练方法.
Abstract
This paper expounds the characteristics of FPGA accelerator technology in speech recognition,image recognition processing,optimization strategies,and hardware templates,as well as the feasibility of accelerating deep learning on FPGA.It analyzes the application of FPGA accelerator,explores the hardware template,performance,and power consumption level of FPGA accelerator for accelerating deep reinforcement learning,and proposes a training method for FPGA accelerator.
关键词
电路设计/FPGA/语音识别/图像识别处理/优化策略/硬件模板Key words
circuit design/FPGA/speech recognition/image recognition processing/optimization strategies/hardware templates引用本文复制引用
出版年
2024
NETL
NSTL
万方数据