Neptune documentation

What is Neptune?

Neptune an open, experiment-centric data science platform built for experts by experts. Neptune’s goal is to enable top data scientists develop the most accurate machine learning models that help drive organizations’ competitive edge, while keeping the experimentation process organized, efficient and easy to share with different stakeholders.

Register here and give it a try!

What can I use Neptune for?

Unconstrained machine learning experimentation

Neptune offers number of features centered around experimentation (training, evaluating, testing) with machine learning models of any kind. Check tutorials to learn more about our experiment-focused approach.

Neptune allows you to freely choose your Python libraries. It simply mean that you work with your favorite tools while leveraging full neptune’s potential. Check get started tutorial that will guide you - step-by-step - how to adapt your code to fully integrate it with Neptune (in practise: few additional lines of code).

Neptune is light-weight solution that let’s you easily start playing with it in your organization.

Real-time model training tracking

Each experiment, once initiated, is immediately available in the project dashboard, which allows you to see its details, track training progress and compare it to all remaining experiments.

Run your computations in the cloud automatically

For model training and experimentation purposes Neptune leverages Google Cloud Platform (GCP) end-to-end. This allows you to access large variety of machines - including preemptible instances - without any administration effort.

Tracking project progress

Either you are ML researcher or department head, you care about project progress. Neptune facilitates transparency, by introducing feature to compare experiment, see models improvements over time. Power Users can check experiments API, to learn how to build customized comparisons.

Is Neptune a tool for me?

Neptune is designed with two particular groups in mind:

  1. Hands-on data scientists, who train machine learning models and run experiments. With neptune you will find extensive experimentation and model training more fun and productive.
  2. Lead data scientists, who need easy-to-use place for her projects’ health-check. She also cares about her team and wants to deliver best possible tooling while allowing teammate to work in most effective way.

What next?

Depending on your needs, you can go to:

Happy training :)