Skip to article frontmatterSkip to article content
Site not loading correctly?

This may be due to an incorrect BASE_URL configuration. See the MyST Documentation for reference.

RiemannianStaircaseParams

RiemannianStaircaseParams configures the Burer–Monteiro rank schedule, SDP certificate verification, saddle escape, and the augmented-Lagrangian local solver used at every staircase level.

Open In Colab

Parameter groups

  • pMin and pMax bound the Burer–Monteiro factorization rank.

  • alpha scales the new column used to escape a certified saddle direction.

  • eta is the PSD certificate tolerance.

  • verificationMethod selects sparse Spectra verification or dense Eigen verification.

  • maxSpectraIters, numLanczosVectors, and spectraTol tune the sparse eigensolver fallback.

  • getAlmParams() and setAlmParams() copy the inner augmented-Lagrangian parameters across the Python boundary.

rank range: 2 to 5
inner iterations: 80
verification: VerificationMethod.Spectra

Rot2 matrix problems require pMin >= 2; Rot3 requires pMin >= 3. DenseEigen is useful for small validation problems, while Spectra avoids a dense cubic eigendecomposition and is the normal scalable choice.