Machine Learning Refined: Python Jupyter notebook collection

See our blog here for interactive versions of the notebooks in this repo. These posts describe a range of topics in machine learning / deep learning including a wide variety of topics in supervised learning, mathematical optimization and automatic differentiation / the back propagation algorithm, and reinforcement learning.

In order to effectively run the Jupyter notebooks contained in this repo on your own machine we strongly recommend using the Anaconda Python 3 distribution which can be downloaded here since the default install contains most of the library dependencies used here as well as as Jupyter notebook, with the exception of autograd, which can be installed using pip by typing

pip install autograd

at the terminal.

To re-run the animations contained withiin these jupyter notebooks you can initialize your jupyter session with the following adjusted command in place of the standard 'jupyter notebook' initialization command - which increases the rate you can plot images to a Jupyter notebook cell

jupyter notebook --NotebookApp.iopub_data_rate_limit=10000000000

This repository contains various supplementary Jupyter notebooks, Python and MATLAB files, presentations associated with the textbook Machine Learning Refined (Cambridge University Press). Visit http://www.mlrefined.com for free chapter downloads and tutorials, and our Amazon site here for details regarding a hard copy of the text.