Overview
TensorFlow Decision Forests (TF-DF) is a library for training, running, and interpreting decision forest models within the TensorFlow ecosystem. It supports various machine learning tasks including classification, regression, ranking, and uplift modeling. TF-DF offers implementations of Random Forests and Gradient Boosted Trees, allowing users to leverage these algorithms for tabular data analysis. The library provides tools for model interpretation, helping users understand the behavior and predictions of their models. YDF (Yggdrasil Decision Forests) extends TF-DF with new features, a simplified API, and faster training times. TF-DF integrates seamlessly with TensorFlow, enabling users to deploy decision forest models in TensorFlow Serving and other TensorFlow-based environments. It natively handles numeric, categorical, and missing features, reducing the need for extensive preprocessing.
Common tasks