Updated on 2022-03-13 GMT+08:00

Intended Audience

This document is intended for developers who program operators using Ascend 310. It aims to help you:

  • Understand the end-to-end (E2E) development process of a custom TE operator, including developing a custom operator and integrating it into the network to run on E2E applications.
  • Learn the operator development and operator plug-in development principles by referring to the reduction operator sample in this document, and be able to develop other custom operators.
To better understand this document, you should have:
  • Capability of developing Python/C++/C language programs
  • Good understanding of mathematical expressions
  • Good understanding of machine learning and deep learning
  • Good understanding of the Ascend 310 architecture
  • Good understanding of TVM and open source framework Caffe