Learning Wizard

Purpose

This document provides guidance for the development of deep neural network (DNN) apps from existing models through the use of C language APIs provided by the Ascend Computing Language (AscendCL), for such purposes as target recognition and image classification.

With this document, you will:

• Learn about the AscendCL architecture, basic terms, and typical API call sequences.
• Learn about the procedure and implementation of app development using AscendCL APIs.
• Know how to develop additional apps based on the provided samples.

To understand this document better, you are supposed to have ability in C/C++ programming and a basic understanding of machine learning or deep learning.

Document Usage Suggestions

If you use this document for the first time or are not clear about the following questions, you are advised to learn the overall application development process of Getting Started first, then learn the overview, and then learn the API call sequence and sample code in sections Developing a Basic Inference Application, Image/Video Data Processing, and Single-Operator Execution.

• Where is AscendCL in the CANN architecture?
• What are the functions of the devices, streams, and contexts in AscendCL?
• What is the basic workflow for developing apps using AscendCL APIs?

If you want to dive deeper, refer to the following app development wizard.