sparse-kappa documentation

sparse-kappa is a sparse matrix condition number estimation library with PyTorch acceleration, covering both numerical algorithms and optional GNN-based prediction workflows.

Getting started

• Installation
  • Prerequisites
  • Check Your System
  • Installing PyTorch
  • Installing Sparse Kappa
  • Verify Installation
  • Common Issues
  • Next Steps
• Quick Start Guide
  • Basic Usage
  • Common Workflows
  • Method Selection Guide
  • Performance Tips
  • Next Steps

User guide

• User Guide
  • Understanding Condition Numbers
  • Choosing Methods
  • Parameter Tuning
  • Best Practices
  • Common Pitfalls
  • Advanced Usage
• Methods and Algorithms
  • 1-Norm Methods
  • 2-Norm Methods
  • PyTorch Wrapper Methods
  • Algorithm Comparison
  • Convergence Theory
  • Numerical Stability
  • Practical Recommendations
• Algorithm Reference
  • 1-Norm Algorithms
  • 2-Norm Algorithms
  • Convergence Analysis
  • Numerical Stability
  • References
• Examples
  • Basic 2-norm estimation
  • Compare multiple methods
  • 1-norm with LU solver
  • GNN training workflow
• Performance guide
  • Method selection
  • Practical tips
  • Benchmark template
• Troubleshooting
  • Import errors
  • Convergence issues
  • Very large condition numbers
  • GPU / device mismatch

API reference

• API reference
  • Main API
  • Solver API
  • 1-norm methods API
  • 2-norm methods API
  • GNN API

Project

• Contributing
• Changelog

Indices and tables

• Index

• Module Index

• Search Page