Scholarlink

This chapter provides a tutorial for beginners on training their own CNN model at the software level and designing a custom CNN hardware accelerator. It introduces mainstream methods for CNN training, including Python-based frameworks such as TensorFlow. Additionally, the chapter covers VISIT AI, a tool for implementing CNN models on the Xilinx platform, and guides on building custom CNN ASIC hardware accelerators using SystemVerilog.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A Tutorial on CNN Design from Model Training to Hardware Implementation

  • 章节

  • 发表日期：

    2026年

查看全文
  • Peiyao Sun,
  • Haosen Yu,
  • Tom J. Kazmierski,
  • Basel Halak

摘要

This chapter provides a tutorial for beginners on training their own CNN model at the software level and designing a custom CNN hardware accelerator. It introduces mainstream methods for CNN training, including Python-based frameworks such as TensorFlow. Additionally, the chapter covers VISIT AI, a tool for implementing CNN models on the Xilinx platform, and guides on building custom CNN ASIC hardware accelerators using SystemVerilog.