TensorFlow for R

TensorFlow™ is an open source software library for numerical computation using data flow graphs. Nodes in the graph represent mathematical operations, while the graph edges represent the multidimensional data arrays (tensors) communicated between them. The flexible architecture allows you to deploy computation to one or more CPUs or GPUs in a desktop, server, or mobile device with a single API.

The TensorFlow API is composed of a set of Python modules that enable constructing and executing TensorFlow graphs. The tensorflow package provides access to the complete TensorFlow API from within R.

Installation

To get started, install the tensorflow R package from GitHub as follows:

devtools::install_github("rstudio/tensorflow")

Then, call py_require_tensorflow() at the start of each R session, before using TensorFlow:

library(tensorflow)
py_require_tensorflow()

You can confirm that TensorFlow is available with:

hello <- tf$constant("Hello")
print(hello)

This is all you need for the default setup. Reticulate will take care of the details: it will choose or create a suitable Python environment, install TensorFlow and its Python dependencies, and make that environment available to the tensorflow R package. You do not need to install Python packages or configure a Python environment manually.

In most cases, you no longer need to call install_tensorflow(). If you want to create a persistent virtual environment explicitly, use:

install_tensorflow()

See the article on installation to learn about more advanced options, including installing a version of TensorFlow that takes advantage of Nvidia GPUs if you have the correct CUDA libraries installed.

Documentation

See the package website for additional details on using the TensorFlow API from R: https://tensorflow.rstudio.com

See the TensorFlow API reference for details on all of the modules, classes, and functions within the API: https://www.tensorflow.org/api_docs/python/tf/all_symbols

The tensorflow package provides code completion and inline help for the TensorFlow API when running within the RStudio IDE. In order to take advantage of these features you should also install the Current Release of RStudio.