tabben.evaluators
tabben.evaluators¶
- tabben.evaluators.get_metrics(task: str, *, classes: int = 2, _namespace=<class 'tabben.evaluators.metrics.ScikitLearnMetrics'>)¶
Return the specific metric implementations given a dataset’s task (and number of classes if classification).
- Parameters
task (str) – Task of the dataset
classes (int, default=2) – Number of classes if the dataset has a classification task
_namespace – Namespace containing the specific implementations of the metrics (this is an internal parameter that is used to switch backends)
- Returns
Set of metrics/functions that should be used to evaluate models on this benchmark
- Return type
sequence of callable