Solve Ax=b linear systems with NumPy, JAX, PyTorch, and Lineax backends.
npx skills add https://github.com/yonesuke/skills --skill linear-solver
Please help me install this Agent Skill. Skill: linear-solver Source: https://github.com/yonesuke/skills/tree/main/linear_solver Command: npx skills add https://github.com/yonesuke/skills --skill linear-solver
Solves Ax=b using a variety of backends (NumPy, JAX, PyTorch, and Lineax).
Select a backend (NumPy, JAX, PyTorch, or Lineax) and run the provided example to solve Ax=b.
High-intent search queries and answers about installing and using this skill.
To solve linear systems Ax=b, you can use this tool to select JAX, PyTorch, NumPy, or Lineax as your backend. It exposes direct and iterative solvers, allowing you to compute solutions across different hardware accelerators like GPUs or TPUs seamlessly.
Yes, you can use iterative solvers like CG and GMRES for sparse matrices. The tool also supports direct methods including LU, Cholesky, QR, and SVD, alongside TDMA for structured systems, providing flexibility for various matrix properties.
The best way to compare solver performance is using this unified interface, which standardizes direct and iterative methods across NumPy and JAX. You can rapidly prototype solver strategies and benchmark execution on CPU versus GPU or TPU backends.
You do not need Lineax to solve structured linear systems unless you select it as your backend. The tool requires standard Python libraries and supports NumPy, JAX, and PyTorch, offering TDMA and other solvers for structured matrices independently.
Use direct solvers like LU, Cholesky, QR, or SVD for smaller or dense systems requiring high numerical stability. Switch to iterative methods like CG or GMRES for large, sparse systems where memory efficiency and approximation speed are prioritized.
CUexterAutomate distributed PyTorch training with unified APIs and automatic device placement.
hxk622Execute MATLAB and GNU Octave scripts for numerical computing and visualization.
xjtulycPerform matrix decompositions and solve dense or sparse linear systems with NumPy and SciPy.
ZurybrOptimize largest eigenvalue computation for small dense matrices via Cython and LAPACK.
