Configuration

Once you've completed basic installation and setup, you can further configure Neptune for your workflow.

Info

The topics in this section are focused on the Neptune API (Python client library) and how it works underneath the hood.

• For how to use Neptune in your model-training code, see Adding Neptune to your code.
• For onboarding your team, see Onboarding guide for teams.

Environment variables

The most important environment variable to use is NEPTUNE_API_TOKEN.

Neptune API tokens can belong to user accounts or service accounts. Regardless of your setup, make sure to save any API tokens as environment variables.

For the full list, see Environment variables.

Behavior of Neptune client library

You can set the following:

• Logging mode
• Logging level
• Location of .neptune data folder

We also show how to set up:

• Error handling
• Callbacks for when data synchronization lags

Learn more

• Best practices: Complex setups – how Neptune works in distributed or parallel computing.
• Integrations – setup together with specific ML frameworks and other tools.
• Management – user and data management in the Neptune app itself.
• Self-hosted – configuration of self-hosted solutions.

Got issues or questions?

To get dedicated support or leave feedback, email us at support@neptune.ai.

You can also check out the self-hosting FAQ and other ways to get help.