_defaults#
- ns_sdn.engine._core._defaults = <module 'ns_sdn.engine._core._defaults' from '/home/runner/work/ns-sdn/ns-sdn/src/ns_sdn/engine/_core/_defaults.py'>[source]#
Package-wide numeric constants and estimator defaults.
Every magic number resolves here. Changing one changes fitted output, so these are implementation detail rather than public API; the public constructors expose the ones a user may legitimately move (the Huber threshold, the schedule, the active-weight threshold) as arguments with these as their defaults. The training defaults are the two-stage schedule of the IEEE-ORSS DGS10 run.
Example
>>> _STAGE_A[1] + _STAGE_B[1] 3000 >>> _FREQUENCY_MIN < _FREQUENCY_FLOOR * 100 and _TINY < _AMPLITUDE_FLOOR True