Decision forests, which train multiple decision trees and combine their predictions, can be used for tasks including classification, regression, and survival analysis. Recently, the TensorFlow Decision Forest (TF-DF) library, a collection of Keras models, was added to the TensorFlow framework. TF-DF includes state-of-the-art algorithms for training, serving, and interpreting Decision Forest models. It wraps the Yggdrasil Decision Forest C++ libraries.