Applied Machine Learning 2026 - Useful Python ML packages

Writing your own ML algorithm from scratch is hard work, and not what this course is about. Rather, one should be using appropriate packages.
Below is a list of those considered useful/important for the course, but you will surely also be using others (your own favorites).
Explicit version suggestions can be found in the below requirement files (also on course GitHub).

General Packages:
  • numpy
  • matplotlib
  • scipy
  • h5py
  • pandas
  • seaborn
  • ipywidgets
  • notebook

    ML Packages:
  • scikit-learn
  • xgboost
  • lightgbm
  • tensorflow
  • torch
  • torch_geometric
  • torchaudio
  • torchvision
  • bayesian-optimization
  • optuna
  • shap
  • scikeras

    These packages (with explicit versions) can also be found in this requirements file.


    Last updated: 13th of April 2026 by Troels Petersen.