概览

深层神经网络中使用的人工智能 (AI) 技术已为 IT 领域提供了新的价值。 虽然很多人希望用 AI 技术实现嵌入式应用，但 AI 处理需要大量计算，由于性能不足或功耗需求较大，导致难以使用搭载 CPU 或 GPU 的传统解决方案来适应嵌入式设备。 此外，AI 技术在不断演进，而新的算法也在不断发展。

在 AI 快速发展的过程中，瑞萨开发出了高性能、低能耗且能够随着演进做出响应的 AI 加速器 (DRP-AI) 和软件（DRP-AI 转换器）。 DRP-AI 和 DRP-AI 转换器相结合可实现高功效的 AI 推理，而这是当前的 AI 技术无法支持的。

随着 DRP-AI 转换器的不断更新，AI 模型还可得到扩展。

DRP-AI Features

DRP-AI 的功能

DRP-AI 由 AI-MAC（乘积累加处理器）和 DRP（可重构处理器）组成。 AI 处理可通过在卷积层和全连接层为运算分配 AI-MAC 来高速执行，而且 DRP 也适用于预处理和池化层等其他复杂的处理。

DRP-AI consists of AI-MAC (multiply-accumulate processor) and DRP (reconfigurable processor)

有关 DRP-AI 的更多详细的技术信息，请参考以下白皮书。
白皮书：Embedded AI-Accelerator DRP-AI (PDF | English, 日本語)

While most AI accelerators specialize only in AI inference and rely on the CPU for pre- and post-processing, DRP-AI integrates pre- and post-processing and AI inference into a single DRP-AI hardware to achieve superior AI processing performance.

AI Processing Performance

软件列表

工具：DRP-AI Translator / DRP-AI TVM*1

DRP-AI Translator and DRP-AI TVM are tools that are available to convert trained AI models into a format that can run on DRP-AI. This section describes the features of these two tools.
The DRP-AI Translator is a tool that is tuned to maximize DRP-AI performance. DRP-AI achieves high-speed performance, low power consumption and reduced CPU load by enabling DRP-AI to perform all the operations of an AI model.
DRP-AI TVM applies the DRP-AI accelerator to the proven ML compiler framework Apache TVM*2. This enables support for multiple AI frameworks (ONNX, PyTorch, TensorFlow, etc.). In addition, it enables operation in conjunction with the CPU, allowing more AI models to be run.

These two tools can be selected according to the customer's product application.

Selectable tools

The table below lists the AI model formats and products (MPUs) supported by each tool.

supported by each tool

The table below shows the download site for each tool, a brief description of the tool and information on deliverables.

  DRP-AI Translator DRP-AI TVM
Tool Download DRP-AI Translator (Renesas Web) DRP-AI TVM (GitHub)
Tool summary description DDRP-AI Translator explanation page DRP-AI TVM explanation page
Implementation guide Included in the DRP-AI Support Package README on GitHub to guide you through the implementation process
Sample Code Performance evaluation samples available on GitHub
AI evaluation software
DRP-AI drivers Included in the DRP-AI Support Package
Linux Available in Linux Package Available in Linux Package

软件：DRP-AI 支持包

DRP-AI 支持包提供驱动器和需要操作 DRP-AI 的指导。 立即下载 DRP-AI，体验从开放式软件到设备实施的无缝 AI 发展。

DRP-AI Support Package

*1. DRP-AI TVM is powered by EdgeCortix MERATM Compiler Framework
*2. For more information on Apache TVM, please refer to https://tvm.apache.org
RZ/V2M DRP-AI Support Package [V7.40]
This product provides the software and documentation for DRP-AI embedded within RZ/V2M.
 Software Package 瑞萨电子
RZ/V2L DRP-AI Support Package [V7.41]
This product provides the software and documentation for DRP-AI embedded within RZ/V2L.
 Software Package 瑞萨电子
RZ/V2MA DRP-AI Support Package[V7.40]
This product provides the software and documentation for DRP-AI embedded within RZ/V2MA.
Software Package 瑞萨电子
DRP-AI Translator [V1.84]
This is an AI model conversion tool (DRP-AI Translator) for DRP-AI equipped products. Please check the Release Notes and User's Manual first before using this product.
 Software Package 瑞萨电子
DRP-AI TVM (GitHub)
We provide an AI model conversion tool (DRP-AI TVM) for DRP-AI-equipped products. When using this product, please check the contents of the linked README.md first.
 Software Package 瑞萨电子
5 items
Boost Vision AI Power Efficiency Using DRP-AI Accelerator

DRP-AI Accelerator embedded in RZ/V series MPUs provides high-speed AI processing while keeping high power efficiency at the endpoints.

Video List
Boost Vision AI Power Efficiency Using DRP-AI Accelerator
RZ/V2M Pose Estimation Operating Heat Generation

新闻和博客

博客 2024年4月4日
博客 2022年4月18日
博客 2022年2月15日
博客 2021年11月15日