Expand description
Integration with scirs2-optim for optimization-driven evaluation
This module provides utilities for integrating metrics with optimization algorithms and learning rate schedulers without circular dependencies.
The integration works by providing configuration types and traits that can be used by external schedulers and optimizers.
Structs§
- Hyper
Parameter - A hyperparameter with its range
- Hyper
Parameter Search Result - A hyperparameter search result
- Hyper
Parameter Tuner - A hyperparameter tuner
- MetricLR
Scheduler - A metric-based learning rate scheduler
- Metric
Optimizer - Adapter for using scirs2-metrics with scirs2-optim
- Scheduler
Bridge - Bridge adapter for external scheduler integration
- Scheduler
Config - Configuration for external scheduler creation
- Scheduler
State - Current state of a metric-based scheduler
Enums§
- Optimization
Mode - Metric optimization mode
Traits§
- Metric
Scheduler Trait - Trait for external scheduler integration