NumPy For Beginners

This LibGuide gives SJSU students and faculty an introduction to NumPy—what arrays are, why they’re fast, and how to do everyday tasks like creating arrays, indexing, broadcasting, and basic I/O.

Things To Know

Interoperability (the 10-second version): NumPy is built to work with other array libraries (including GPU tools).

It can:

(1) read or share other arrays without copying (via standard interfaces or np.from_dlpack(...)),
(2) hand off NumPy calls to those libraries so work runs on their hardware (__array_function__, __array_ufunc__, or creators with like=), and
(3) return the right type back so metadata and behavior are preserved.

In practice: convert explicitly when you switch tools, watch for copies vs. views, and know whether you’re on CPU or GPU.

For more information, please look here.