NumPy
NumPy is “the fundamental package for scientific computing with Python” (numpy.org). It provides an N-dimensional array object with vectorisation, indexing and broadcasting, plus mathematical functions, random number generators, linear algebra and Fourier transforms. The core is optimised C. The site says it forms the basis of libraries such as scikit-learn and SciPy, and that array libraries including Dask, CuPy, JAX, Xarray, Sparse, PyTorch and TensorFlow build on or are compatible with its API.
Repository snapshot (GitHub API, 2026-11-19)
| Item | Detail |
|---|---|
| Repo | numpy/numpy |
| Licence | The site says “liberal BSD license”; GitHub does not classify it (the LICENSE.txt is headed “Copyright (c) 2005-2026, NumPy Developers”) |
| Created | 2010-09-13 |
| Activity | last push 2026-11-19; latest GitHub release v2.5.5 (2026-11-15); the site’s headline release is NumPy 2.5.0 (2026-06-21) |
| Size | 33,423 stars, 13,113 forks |
| Governance | Described as developed publicly by a community; the README badge says “Powered by NumFOCUS” |
Where it fits
NumPy is the array foundation under pandas, scipy, Scikit-learn and (by API compatibility) pytorch.
Sources
- numpy.org, the NumPy README and GitHub API (2026-11-19)