Pole-safe rational neural layer / reference implementation
Beats tuned baseline
Pole-safe rational neural layer — reference implementation
Runnable experiment code, benchmark implementation, machine-readable results and reports for this SynthCore idea.
Language: Python · Runtime: Python 3 · License: MIT
Public artifacts
README.md— 1.7 KBbench_report.json— 3.4 KBpole_safe_experiment.py— 4.3 KBreport.md— 1.6 KBreport_bench_2026-09-03T225017.md— 5.2 KBresults.json— 1.0 KBstage2_bench.py— 5.2 KB