jaxlie documentation

   

jaxlie is a Lie theory library for rigid body transformations and optimization in JAX.

Current functionality:

• SO(2), SE(2), SO(3), and SE(3) Lie groups implemented as high-level dataclasses.

• exp(), log(), adjoint(), multiply(), inverse(), and identity() implementations for each Lie group.

• Pytree registration for all dataclasses.

• Broadcasting for leading axes.

• Helpers + analytical Jacobians for tangent-space optimization (jaxlie.manifold).

Source code on Github.

API Reference

• jaxlie
  • jaxlie.hints
  • jaxlie.manifold
  • jaxlie.utils

Example usage

• Basics
• SE(3) Optimization
• jax.vmap Usage