numerical-linear-algebra

Why we included this project

For anyone who keeps reaching for NumPy without really knowing what it is doing under the hood, this free course from fast.ai is a strong place to fix that. It is built around Jupyter notebooks that focus on doing matrix computations fast enough and accurately enough for real machine learning work, with hands-on examples like PageRank, compressing images with SVD, and efficient gradient descent. Most lessons stay in Python with NumPy and Scikit-Learn, while Numba and PyTorch appear in a few chapters when speed or GPU topics come up. Since every notebook renders in the browser and there is a matching YouTube lecture playlist, you can work through it at your own pace or jump back to specific topics when a later project trips you up. It originally ran as a graduate course at the University of San Francisco, and the notebooks still read like teaching material rather than reference documentation.