Skip to content

basyirin-dev/sigmaflow-pde

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

4 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

σFlow-PDE

License: MIT Python

A drop-in H-Bar training engine that escapes the σ-trap in neural PDE solvers via live σ/δ/α ODE integration, autonomous phase curriculum, and auto-falsification.


Overview

Neural PDE solvers (FNO, DeepONet, PINNs) are notorious for getting stuck in low-frequency, mean-predicting solutions — the σ-trap. σFlow-PDE introduces a training-time framework that:

  • Live σ/δ/α ODE integration – Continuously evolves spectral coefficients during training to escape local spectral minima.
  • Autonomous phase curriculum – Adaptively schedules training phases based on spectral convergence diagnostics.
  • Auto-falsification – Automatically detects and rejects models that fail spectral consistency checks, ensuring robust generalization.

Topics

physics-ml · neural-operators · compositional-generalization · training-dynamics · pde-solver · h-bar-framework · ood-generalization · fno · deeponet · reproducible-ml

License

Distributed under the MIT License. See LICENSE for more information.

About

σFlow-PDE: A drop-in H-Bar training engine that escapes the σ-trap in neural PDE solvers via live σ/δ/α ODE integration, autonomous phase curriculum, and auto-falsification.

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages