The second assignment covers iterative optimization and parametric (deep) models for image classification.
This part is about implementing and experimenting with different flavors of gradient descent.
Download the data from here. Your task is to implement gradient_descent_2d.py. See the code comments for instructions. The fn/ folder contains sampled 2D functions for use with that script. You can add more functions if you want, here is a list of interesting candidates.
The goal of this part is for you to understand gradient descent better by playing around with it. Try different parameters, starting points, and functions. This nicely highlights the function and limitations of gradient descent, which we've already covered in the lecture.