Sparse Matrices 3: Failing at JAX
Read OriginalThis technical blog post details the author's ongoing struggle to implement sparse matrix operations, specifically Cholesky factorizations, within the JAX framework. It examines JAX's transformation limitations and array immutability, analyzing their impact on three core functions needed for sparse matrix support in PyMC to accelerate inference for linear mixed and spatial models.
Comments
No comments yet
Be the first to share your thoughts!
Browser Extension
Get instant access to AllDevBlogs from your browser