From 317e094ae5aa81321694bfd6b2ff1fdab2a4ac33 Mon Sep 17 00:00:00 2001 From: Haihan Jiang Date: Sun, 14 Jun 2026 02:15:20 -0700 Subject: [PATCH] docs: add scipy least squares API docs --- docs/api.rst | 2 ++ docs/nonlinear_least_squares.rst | 14 ++++++++++++++ 2 files changed, 16 insertions(+) diff --git a/docs/api.rst b/docs/api.rst index ee42dd66..b57944bc 100644 --- a/docs/api.rst +++ b/docs/api.rst @@ -98,6 +98,8 @@ Nonlinear least squares jaxopt.GaussNewton jaxopt.LevenbergMarquardt + jaxopt.ScipyLeastSquares + jaxopt.ScipyBoundedLeastSquares Root finding ------------ diff --git a/docs/nonlinear_least_squares.rst b/docs/nonlinear_least_squares.rst index e416c205..6cef3034 100644 --- a/docs/nonlinear_least_squares.rst +++ b/docs/nonlinear_least_squares.rst @@ -98,3 +98,17 @@ parameters: where :math:`\mathbf{J}` is the Jacobian of the residual function w.r.t. parameters and :math:`\mu` is the damping parameter. + +SciPy wrappers +-------------- + +.. autosummary:: + :toctree: _autosummary + + jaxopt.ScipyLeastSquares + jaxopt.ScipyBoundedLeastSquares + +For users who need SciPy's ``least_squares`` implementation with PyTree and +implicit differentiation support, :class:`jaxopt.ScipyLeastSquares` wraps +unconstrained nonlinear least squares problems and +:class:`jaxopt.ScipyBoundedLeastSquares` handles problems with box constraints.