4. Our Values

This documentation serves as a guide for developers and code reviewers to ensure our actions align with our development values, which are of paramount importance as an open-source project.

These values include:

1. Be authentic

We should ensure this repository stays authentic regarding all features and all claims we produce. We will always uphold the principle that our project is developed based on the motivation to help the community better tackle large-scale deep learning training and tightly integrated it with our innovations.

2. Be user-centered

As an open-source project, we should value our users as they are the important ones to drive this project forward. We should consider how our features can help the users better train their models and how they can use our project easily without mastering the system expertise.

3. Be reliable

We should closely follow the test-driven development to ensure our project is reliable in different use cases. An array of testing, including unit testing, integrating testing, system testing, and compatibility testing, should be conducted during the development process.

4. Be helpful

This value covers a wide range of development practices and the gist is that when you take an action, you should always consider how to make it easier for the recipient of your action. For example, when you create a pull request, you should make it easy for code reviewers to review your code without scratching their heads. Another case would be when a user creates an issue, the developer should offer responses in time so that the user is not left unanswered for a long time.

5. Be careful

As a developer or code reviewer, we should always make sure that the code is original. Otherwise, it is essential to acknowledge and credit any relevant work/project to align with the integrity guideline in the open-source community, academia, and industry.