ROSE Standard and Custom Metrics¶

ROSE supports different Machine Learning (ML) Metrics such as RMSE, MAE, and F2score and many more.

Standard Metrics¶

For a full list of the supported metrics please refer to the following link ROSE Standard Metrics

Custom Metrics¶

ROSE allows the user to define additional metrics if not supported by default. To define a custom metric, you can do the following:

import the operator for the custom metric:

from rose.metrics import GREATER_THAN_THRESHOLD

Now define your @acl.as_stop_criterion with additional args operator:

# Defining the stop criterion with a metric
@acl.as_stop_criterion(metric_name='custom_metric',
                       operator=GREATER_THAN_THRESHOLD, threshold=0.8)
def check_metric(*args):
    return Task(executable=f'python3 check_custom_metric.py')

In this way, ROSE will understand the relation between the custom metric and the target threshold value.