Optimistix v0.0.9

This is a compatibility release: JAX 0.4.34 changed how custom autodifferentiation rules work. As of this update we should now be compatible with this. (#87)

Full Changelog: v0.0.8...v0.0.9

Optimistix v0.0.8

• Compatibility with latest JAX (fixed some warnings, fixed some crashes from the jax.ShapeDtypeStruct.weak_dtype breaking change).
• Many solvers that need to calculate Jacobians will now preferentially use forward mode over reverse mode where possible. This generally improves speed. (#61)
• OptaxMinimiser now supports Optax optimizers that need the current parameter state. (Thanks @NeilGirdhar! #77)
• Bisection now optionally supports expanding the interval to find the root. (Thanks @NeilGirdhar! #78)
• Added warning that complex numbers are still a work in progress -- turns out that some facets of optimizing functions C->R autodifferentiably is perhaps still an open research question (?)
• Doc fixes. (Thanks @johannahaffner! #69)
• Added py.type file to signal static type checking compatibility. (Thanks @NeilGirdhar! #80)

New Contributors

• @johannahaffner made their first contribution in #69
• @NeilGirdhar made their first contribution in #77

Full Changelog: v0.0.7...v0.0.8

Optimistix v0.0.7

Features

• Support for complex numbers! In theory these should now be supported everywhere. In practice we're still considering this a little experimental at the moment, just in case we've missed something. (Huge thanks to @Randl! #53)
• optimistix.{AbstractGaussNewton, GaussNewton, LevenbergMarquardt, IndirectLevenbergMarquardt, Dogleg} should now all support using reverse-mode autodiff to calculate Jacobians. (#51)

Bugfixes

• Fixed a crash when using jax.disable_jit (See patrick-kidger/diffrax#368, #43)
• Fixed terminating on the first step for functions with initial value zero. (Thanks @NightWinkle! #49)

Other

• Documentation fixes. (Thanks @ColCarroll! #32)
• Now compatible with jax_numpy_rank_promotion=raise and jax_numpy_dtype_promotion=strict.

New Contributors

• @ColCarroll made their first contribution in #32
• @NightWinkle made their first contribution in #49
• @Randl made their first contribution in #53

Full Changelog: v0.0.6...v0.0.7

Optimistix v0.0.6

Bugfix release!

• Fixed recompilation happening on every step when iterating step-by-step with a Gauss--Newton solver. (#30)
• sol.state previously included only the array-valued parts of the state (for sol = optx.{minimise, least_squares, root_find, fixed_point}(...)). It now includes everything.
• Fixed optx.internal.implicit_jvp misbehaving for non-Optimistix use cases.
• Fixed optx.{Newton,Chord}(cauchy_termination=False) failing when started close to the solution.

Full Changelog: v0.0.5...v0.0.6

Optimistix v0.0.5

Very simple release!

• Added optimistix.compat.minimize as a replacement for jax.scipy.optimize.minimize. (#14)

Full Changelog: v0.0.4...v0.0.5

Initial release!

Hurrah! How exciting.