The winning configuration, its final metric, and the total epoch-budget actually spent across every trial/rung (the exit-gate's own "reduces total training
compute" measure).
More...
#include <successive_halving.hpp>
The winning configuration, its final metric, and the total epoch-budget actually spent across every trial/rung (the exit-gate's own "reduces total training
compute" measure).
◆ best_configuration
| Configuration pulsatrix::SuccessiveHalvingResult::best_configuration |
◆ best_metric
| double pulsatrix::SuccessiveHalvingResult::best_metric |
◆ total_epochs_trained
| size_t pulsatrix::SuccessiveHalvingResult::total_epochs_trained |
The documentation for this struct was generated from the following file: