Curvlinops Logo

Getting started

  • How to use Curvlinops

Curvlinops

  • Linear operators
  • Spectral density approximation
  • Trace approximation
  • Diagonal approximation
  • Frobenius norm approximation
  • Code samples

Internals

  • Internals
Curvlinops
  • Curvlinops
  • View page source

Curvlinops

Installation

pip install curvlinops-for-pytorch

Getting started

  • How to use Curvlinops

Curvlinops

  • Linear operators
    • Hessian
    • Generalized Gauss-Newton
    • Fisher (approximate)
    • Uncentered gradient covariance (empirical Fisher)
    • Jacobians
    • Inverses
    • Sub-matrices
  • Spectral density approximation
    • lanczos_approximate_spectrum()
    • lanczos_approximate_log_spectrum()
    • LanczosApproximateSpectrumCached
  • Trace approximation
    • hutchinson_trace()
    • hutchpp_trace()
    • xtrace()
  • Diagonal approximation
    • hutchinson_diag()
    • xdiag()
  • Frobenius norm approximation
    • hutchinson_squared_fro
  • Code samples

Internals

  • Internals
    • GGN-related
      • loss_hessian_matrix_sqrt()
    • EKFAC-related
Next

© Copyright 2022, F. Dangel, L. Tatzel, R. Eschenhagen.

Built with Sphinx using a theme provided by Read the Docs.