User’s Guide#

The user guide is a relatively general MegEngine operation manual, the content of which is organized according to the basic process of machine learning and combined with the actual use of the user. We will continue to enrich the content of the guide and improve the quality of the guide based on the feedback from the community.

Instructions for use

  • Users who are new to MegEngine should start from the Beginner’s guide link. We assume that the readers of the user guide MegEngine Deep Learning Getting Started Tutorial and enter the daily use phase of the framework.

  • For experienced users, they can selectively browse the topics they are interested in, and there is no strong dependency between different topics. But we also try our best to organize the content of the user guide in an orderly manner to meet the needs of users who are accustomed to reading in order.

Model Development (Basics)📕

Basics will expand the extension :ref:using common scenarios MegEngine each child deep-learning mentioned in the package.

Model development (advanced article) 📗

The advanced part will introduce more features to help you fully grasp the use of MegEngine and give full play to its strengths.

Reasoning deployment chapter 📘

Introduce how to deploy the trained model to the actual production environment for efficient reasoning.

Tools and Plugins📙

Understand the various tools and plug-ins provided in MegEngine.

You may be reading the user guide with a specific question 🤔, you can try to find the solution below ⬇️ ⬇️ ⬇️

MegEngine Cookbook#

Seek more support#

For some rare problems (that is, when the usage scenarios are relatively unpopular), the solutions may not be organized in the MegEngine User Guide. You can communicate and discuss through the MegEngine official forum or `GitHub Issues <MegEngine/MegEngine#issues>When the frequency of mentioning a certain type of problem has increased significantly, we will start to consider putting it and the corresponding solution in the user guide as a necessary reference.

Although our community ecology also provides online live chat channels, we need to spend some time to settle the discussion process and solutions when necessary. As the saying goes, “Predecessors planted trees, and future generations enjoyed the shade.” In order to allow our exchanges and discussions to provide reference information for later generations, please use historical records that can be retrieved on the Internet to communicate and discuss as much as possible. We believe that in this way, you can consolidate your experience gained from practice in MegEngine and magnify its value.

We also welcome you to share stories related to MegEngine through your social platform account, so that more people can see MegEngine~