Using multiple integrations at once#
Neptune's integrations work independently of each other. This means that you can use multiple integrations together in the same script.
Example
If you're doing hyperparameter optimization using Optuna on scikit-learn models, you can use the following together:
- The Optuna integration, to log all the relevant Optuna Study and Trial metadata.
- The scikit-learn integration, to log the scikit-learn model metadata.
- If you use Airflow to schedule the above setup, you can add the Airflow integration to automatically log Airflow metadata.
We provide an example notebook and script to demonstrate how these integrations can work together in practice.
Note: Since the integrations work independently, it is necessary to organize the logged metadata in a way that makes it easy for you to search for the information you need, without it being overwritten by another integration.
The concepts remain the same irrespective of the integration you are using.
(Notebook) Neptune + Optuna + scikit-learn#
This Jupyter notebook shows how you can use Neptune's Optuna and scikit-learn integrations together.
(Script) Neptune + Optuna + scikit-learn + Airflow#
This script demonstrates a setup similar to the above, with Airflow scheduling on top.
To view or download the raw notebooks and scripts, visit the Neptune examples repo.