In this notebook, we'll build from scratch a gradient boosted trees regression model that includes a learning rate hyperparameter, and then use it to fit a noisy nonlinear function.