warwick-sciml

Support modules for the University of Warwick module on Predictive Modelling, Uncertainty Quantification and Scientific Machine Learning (PX914 / ES98E).

pip install warwick-sciml              # core: BLR, sensitivity analysis, fluid/FE solvers
pip install "warwick-sciml[gp]"        # + Gaussian process utilities (JAX, tinygp)
pip install "warwick-sciml[atomistic]" # + atomistic examples (ASE, atomistica, matscipy, JAX)

In a marimo notebook, add it to the PEP 723 header instead, e.g.

# /// script
# dependencies = ["warwick-sciml[atomistic]>=2026.1", "marimo>=0.25"]
# ///

then from sciml.blr import design_matrix, from sciml import tersoff_jax, etc.

Module	Contents
sciml.blr	Bayesian linear regression: basis functions, design matrices, posteriors
sciml.sensitivity	finite-difference sensitivities and Taylor expansions
sciml.fluidsolvers	Poiseuille flow solvers
sciml.femsolvers	2D plane-stress finite element solver
sciml.gputils	Gaussian process fitting and plotting (extra gp)
sciml.atomistic	Tersoff calculators, vacancy formation energies (extra atomistic)
sciml.tersoff_jax	differentiable Tersoff potential in JAX (extra atomistic)
sciml.progress	tqdm wrappers for lecture notebooks

Versions follow the academic year (2026.x for 2026/27).

Development

uv run --extra test pytest