jaxlie

Works natively with JAX transformations like vmap, jit, and grad, enabling automatic differentiation and efficient GPU acceleration for optimization tasks, as highlighted in the vmap example.

Provides intuitive classes for SO2, SE2, SO3, and SE3 with methods like exp() and log(), making Lie group operations more accessible compared to low-level C++ libraries like Sophus.

Includes manifold optimization helpers and Taylor approximations near singularities, specifically useful for robotics and vision applications, as demonstrated in the se3_optimization.py example.

Supports broadcasting for leading axes and pytree compatibility, allowing batch operations and seamless integration with JAX's ecosystem for high-performance computing.