Optuna example that optimizes a neural network regressor for the wine quality dataset using Keras and records hyperparameters and metrics using MLflow.
In this example, we optimize the learning rate and momentum of a stochastic gradient descent optimizer to minimize the validation mean squared error for the wine quality regression.
You can run this example as follows:
$ python keras_mlflow.py
After the script finishes, run the MLflow UI:
$ mlflow ui
and view the optimization results at http://127.0.0.1:5000.