SelectionBudgetScenario
Data scenarios · Implementation: mlsysim/engine/scenario_definitions/data.py.
Construct this scenario with keyword inputs. The fields below show the types and defaults declared by this class.
Selection that takes selection_time and leaves a subset that trains in subset_time, against full training in full_time (the selection inequality in hours).
Notes
Equation: total_time is the worked quantity; each property below states its equation and units.
Assumptions: Selection and subset training run sequentially, with all compared durations supplied.
Excludes: Accuracy equivalence, overlapping stages, and any price per unit time.
References
Derivation: The equations below follow the stated counts, geometric relations, or accounting definitions; scenario inputs are supplied assumptions or observations, not fitted predictions.
Examples
Execute the registered worked inputs; this is a modeled calculation.
>>> from mlsysim import Scenarios
>>> case = Scenarios.ExpensiveSelectionPitfall
>>> value = case.total_time
>>> print(f"{value:.3f~P}")
12.000 hAttributes
| Name | Description |
|---|---|
| breakeven_selection | Selection duration at equality with full-training time; strict savings require less. |
| full_time | |
| kept | |
| net_saving | Time saved against full training; negative means selection costs more. |
| overhead_ratio | Dimensionless selection duration divided by subset-training duration. |
| pays | Whether selection plus subset training is faster than full training. |
| pool | |
| selection_share | Dimensionless selection duration divided by full-training duration. |
| selection_time | |
| subset_time | |
| total_time | Combined selection and subset-training duration. |
selection_time
Type: Quantity.
Declared default: required.
subset_time
Type: Quantity.
Declared default: required.
full_time
Type: Quantity.
Declared default: required.
pool
Type: Optional[float].
Declared default: None.
kept
Type: Optional[float].
Declared default: None.
total_time
engine.scenario_types.SelectionBudgetScenario.total_time
Combined selection and subset-training duration.
Equation: return self.selection_time + self.subset_time
net_saving
engine.scenario_types.SelectionBudgetScenario.net_saving
Time saved against full training; negative means selection costs more.
Equation: return self.full_time - self.total_time
pays
engine.scenario_types.SelectionBudgetScenario.pays
Whether selection plus subset training is faster than full training.
Equation: return self.total_time < self.full_time
breakeven_selection
engine.scenario_types.SelectionBudgetScenario.breakeven_selection
Selection duration at equality with full-training time; strict savings require less.
Longest selection that still saves time.
Equation: return self.full_time - self.subset_time
overhead_ratio
engine.scenario_types.SelectionBudgetScenario.overhead_ratio
Dimensionless selection duration divided by subset-training duration.
Selection time over subset training time.
Equation: return (self.selection_time / self.subset_time).to(’’).magnitude