ncem.train.GridSearchContainer.get_best_model_id

GridSearchContainer.get_best_model_id(subset_hyperparameters: Optional[List[Tuple[str, str]]] = None, metric_select: str = 'r_squared_linreg', partition_select: str = 'test', cv_mode: str = 'mean')[source]

Get best model identifier.

Parameters

  • subset_hyperparameters (list, optional) – List of subset hyperparameters.

  • metric_select (str) – Selected metric.

  • partition_select (str) – Selected parition.

  • cv_mode (str) – cross validation mode.

Returns

Return type

best_model_id

Raises

  • ValueError – If measure, partition or cv_mode not recognized.

  • Warning – If cv_mode max is selected with the following metrics: loss, elbo, mse, mae