Hi, looks like a very promising library, this bit in the docs got me interested:
Unlike the SciPy implementation of Newton's method, the Optimistix version also works for vector-valued (or PyTree-valued) y.
Does this mean that the function passed to the root finder has to be vectorized in the traditional sense? Do the root finders here support functions which rely on vmap? I couldn't find anything in the docs about this. Thanks!
Hi, looks like a very promising library, this bit in the docs got me interested:
Does this mean that the function passed to the root finder has to be vectorized in the traditional sense? Do the root finders here support functions which rely on
vmap
? I couldn't find anything in the docs about this. Thanks!