Release Notes

tinygp 0.2.0 (2022-03-03)

Features

• Added new interface for conditioning GP models. condition method now returns a tinygp.GaussianProcess object describing the conditional distribution. (#32)

• Added new experimental scalable solver using quasiseparable matrices. See kernels.quasisep package, solvers.quasisep package, and tinygp.solvers.quasisep.solver.QuasisepSolver for more information. (#47)

Improved Documentation

• Updated benchmarks to include quasiseparable solver. (#49)

• Major overhaul of API documentation. Added many docstrings and expanded text thoughout the API docs pages. (#52)

• Added 3 new tutorials: (1) An Introduction to tinygp, giving a general introduction to tinygp, (2) Fitting a Mean Function, showing how tinygp can be used with a non-trivial mean function, and (3) Scalable GPs with Quasiseparable Kernels, introducing the scalable solver for quasiseparable kernels. (#54)

• Added support for towncrier generated release notes. (#55)

Deprecations and Removals

• Breaking change: Removed existing condition method and deprected predict method. (#32)