Tutorials

Hands-on tutorials.

A growing set of walkthroughs. Pick one based on what you want next: get something running, see how a classical ML problem becomes a QP, or learn to debug from the spike raster.

  1. 01
    Quickstart

    Install snn_opt, solve your first 2-D QP, read off the result. Five lines of Python.

    ~5 min any
  2. 02
    An SVM is a QP

    Turn a textbook support-vector-machine dual into a problem the solver can take, end to end. The bridge between classical ML and snn_opt.

    ~15 min intermediate
  3. 03
    Reading the spike raster

    What the dots mean. What they tell you about your problem. What to look for when something is wrong.

    ~10 min intermediate

What's coming

  • Receding-horizon control — using warm starts to make MPC fast.
  • Sphere-constrained problems — what changes when the constraint set is non-convex (the PCA case).
  • From CVXPY to snn_opt — translating a problem expressed in a modeling language into the explicit `(A, b, C, d)` form.

Suggestions for tutorials are welcome — open an issue on GitHub with what you'd like to see covered.