unstable-singularity-detector

no cloud dependencies — everything runs locally (CLI, Docker, or Gradio). GitHub: https://github.com/Flamehaven/unstable-singularity-detector Paper inspiration: https://arxiv.org/abs/2509.14185 Would love feedback from folks working on scientific computing, PDE solvers, numerical methods, or ML-based verification. I’m around for questions.

ML. --- ### Who This Is For - Researchers studying singularities in PDEs - Practitioners working with PINNs - Anyone frustrated by non-reproducible ML papers - Students learning scientific machine learning --- ### Built For Reproducibility - 119 comprehensive tests (all passing) - GPU-accelerated with PyTorch + CUDA - Automated validation CI/CD - Professional documentation + limitations clearly stated - Not just "research code"—designed to actually work in other people's hands. --- ### Current Implementation Status - Lambda prediction formulas (formula-based validation) - Funnel inference (secant method) - Enhanced Gauss-Newton optimizer (test problem verified) - PINN solver framework - Multi-stage training (framework complete, full validation pending) - Full 3D Navier-Stokes solver (future work) - Transparency matters: This shows exactly what works today vs. what's roadmap. --- ### Links GitHub: https://github.com/Flamehaven/unstable-singularity-detector Paper: https://arxiv.org/abs/2509.14185 Documentation: README includes limitations, validation methodology, troubleshooting --- ### Important Disclaimer This is an independent re-implementation inspired by DeepMind's published methodology. It is NOT affiliated with, endorsed by, or in collaboration with DeepMind. Results are validated against published formulas, not direct comparison with DeepMind's unpublished experimental data. --- If you've ever wished "I wish I could just run their code," this is for you. Contributions, issues, and feedback welcome. Let's make frontier research more accessible together.