TaskFactorCard
Controlled_Studies scenarios · Implementation: mlsysim/engine/scenario_definitions/controlled_studies.py.
Construct this scenario with keyword inputs. The fields below show the types and defaults declared by this class.
One authored task factorization with a separate quality observation.
Notes
Equation: duration = samples * ops_per_sample * seconds_per_op. Assumptions: All samples share the supplied work and realized execution rate. Excludes: Training dynamics and a predicted quality response.
References
Derivation: Multiplying sample count, work per sample, and seconds per operation gives the declared host duration; quality is supplied independently.
Examples
>>> from mlsysim import Scenarios
>>> Scenarios.ControlledTaskFactors.cards["unverified"].quality_pct is None
TrueAttributes
| Name | Description |
|---|---|
| factor | |
| label | |
| model_config | |
| ops_per_sample | |
| quality_pct | |
| samples | |
| seconds_per_op |
label
Type: str.
Declared default: required.
samples
Type: int.
Declared default: required.
ops_per_sample
Type: Quantity.
Declared default: required.
seconds_per_op
Type: Quantity.
Declared default: required.
quality_pct
Type: Optional[float].
Declared default: required.
factor
Type: str.
Declared default: required.