asdex

Automatic Sparse Differentiation in JAX.

asdex (pronounced Aztecs) exploits sparsity structure to efficiently compute sparse Jacobians and Hessians. It implements a custom Jaxpr interpreter that uses abstract interpretation to detect sparsity patterns from the computation graph, then uses graph coloring to minimize the number of AD passes needed.

Installation

pip install asdex

Or with uv:

uv add asdex

Example

import numpy as np
from asdex import jacobian

def f(x):
    return (x[1:] - x[:-1]) ** 2

jac_fn = jacobian(f, input_shape=50)
# ColoredPattern(49×50, nnz=98, sparsity=96.0%, JVP, 2 colors)
#   2 JVPs (instead of 49 VJPs or 50 JVPs)
# ⎡⠙⢦⡀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⎤   ⎡⣿⎤
# ⎢⠀⠀⠙⢦⡀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⎥   ⎢⣿⎥
# ⎢⠀⠀⠀⠀⠙⢦⡀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⎥   ⎢⣿⎥
# ⎢⠀⠀⠀⠀⠀⠀⠙⢦⡀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⎥   ⎢⣿⎥
# ⎢⠀⠀⠀⠀⠀⠀⠀⠀⠙⢦⡀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⎥   ⎢⣿⎥
# ⎢⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠙⢦⡀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⎥   ⎢⣿⎥
# ⎢⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠙⢦⡀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⎥ → ⎢⣿⎥
# ⎢⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠙⢦⡀⠀⠀⠀⠀⠀⠀⠀⠀⎥   ⎢⣿⎥
# ⎢⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠙⢦⡀⠀⠀⠀⠀⠀⠀⎥   ⎢⣿⎥
# ⎢⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠙⢦⡀⠀⠀⠀⠀⎥   ⎢⣿⎥
# ⎢⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠙⢦⡀⠀⠀⎥   ⎢⣿⎥
# ⎢⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠙⢦⡀⎥   ⎢⣿⎥
# ⎣⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠉⎦   ⎣⠉⎦

for x in inputs:
    J = jac_fn(x)

Instead of 49 VJPs or 50 JVPs, asdex computes the full sparse Jacobian with just 2 JVPs.

Documentation

• Getting Started — step-by-step tutorial
• How-To Guides — task-oriented recipes
• Explanation — how and why it works
• API Reference — full API documentation

Acknowledgements

This package is built with Claude Code based on previous work by Adrian Hill, Guillaume Dalle, and Alexis Montoison in the Julia programming language:

• An Illustrated Guide to Automatic Sparse Differentiation, A. Hill, G. Dalle, A. Montoison (2025)
• Sparser, Better, Faster, Stronger: Efficient Automatic Differentiation for Sparse Jacobians and Hessians, A. Hill & G. Dalle (2025)
• Revisiting Sparse Matrix Coloring and Bicoloring, A. Montoison, G. Dalle, A. Gebremedhin (2025)
• SparseConnectivityTracer.jl, A. Hill, G. Dalle
• SparseMatrixColorings.jl, G. Dalle, A. Montoison
• sparsediffax, G. Dalle

which in turn stands on the shoulders of giants — notably Andreas Griewank, Andrea Walther, and Assefaw Gebremedhin.