Galileo-Galilei/kedro-mlflow-tutorial

Use MlflowMetricsDataSet instead of hardcoding metric logging

Galileo-Galilei opened this issue · 1 comments

Galileo-Galilei commented

This code demonstrates bad practices which is against the goal of this tutorial:

https://github.com/Galileo-Galilei/kedro-mlflow-tutorial/blob/main/src/kedro_mlflow_tutorial/pipelines/ml_app/nodes_ml/train_model.py#L38-L50

Metrics should ot be logged directly but through a dataset.