You don't need to be a developer or a technical writer to make a significant impact on the TensorFlow documentation—just a GitHub account. This guide shows how to make contributions to tensorflow.org.
See the TensorFlow docs contributor guide for guidance. For questions, the docs@tensorflow.org mailing list is available.
Questions about TensorFlow usage are better addressed on Stack Overflow or the discuss@tensorflow.org mailing list.
To contribute to the TensorFlow code repositories, see the TensorFlow code contributor guide and the TensorFlow contribution guidelines.
We love patches! To publish your changes, you must sign either the individual or corporate Contributor License Agreement (CLA):
- If you are an individual writing original documentation or source code and you're sure you own the intellectual property, sign an individual CLA.
- If you work for a company that wants to allow you to contribute your work, sign a corporate CLA.
We can accept your pull requests after you sign the CLA. We can only receive original documentation and source code from you and other people that have signed the CLA.
The TensorFlow documentation is written in Markdown and Jupyter/Colab notebooks.
The root of tensorflow.org/ is found in the
site/en directory.
Not all technical content on tensorflow.org is located in site/en. Some
projects have their repositories under
github.com/tensorflow and they contain
project-specific documentation. These projects are navigable from the
tensorflow/docs site/en directory and include a redirect link to where the
docs can be updated.
The API reference is generated from the source code located in the core tensorflow/tensorflow repository and other projects.
Additionally, some non-technical content, images, and design elements are not
located in the tensorflow/docs repository.