Trains and plots models with different decision forest learning algorithms (Cart, Random Forest, Oblique Random Forests, Gradient Boosted Trees, Gradient Boosted Trees with Global optimization, Dart, Extremly Randomized Trees) on a 2d ellipse pattern and with various amount of training data (from 10 to 10k examples). Models are trained with TensorFlow Decision Forests. The code is available in this Colab.