This section explains how developers can add functionality to TensorFlow's capabilities. Begin by reading the following architectural overview:
The following guides explain how to extend particular aspects of TensorFlow:
- Adding a New Op, which explains how to create your own operations.
- Adding a Custom Filesystem Plugin, which explains how to add support for your own shared or distributed filesystem.
- Custom Data Readers, which details how to add support for your own file and record formats.
- Creating Estimators in tf.contrib.learn, which explains how to write your own custom Estimator. For example, you could build your own Estimator to implement some variation on standard linear regression.
Python is currently the only language supported by TensorFlow's API stability promises. However, TensorFlow also provides functionality in C++, Java, and Go, plus community support for Haskell and Rust. If you'd like to create or develop TensorFlow features in a language other than these languages, read the following guide:
To create tools compatible with TensorFlow's model format, read the following guide: