Welcome to Neptune!

Neptune is a light-weight experiment management tool that helps you keep track of your machine learning experiments.

Use Neptune to log hyperparameters and output metrics from your runs, then visualize and compare results. Automatically transform tracked data into a knowledge repository, then share and discuss your work with colleagues.

  • Neptune fits in any workflow, ranging from data exploration and analysis, decision science to machine learning and deep learning.

  • Neptune works with common technologies in the data science domain: Python, Jupyter Notebooks, and R, to mention a few.

  • It has convenient integrations with many Python libraries for machine learning and deep learning such as: Keras, PyTorch Lightning, XGBoost, Matplotlib.

  • It integrates with other tracking tools such as MLflow and TensorBoard or Sacred.

  • It works seamlessly with your infrastructure, be it AWS, GCP, Kubernetes, Azure, or on-prem machines.

  • The Neptune Python Library is an open source package that allows you to integrate your Python scripts with Neptune. Once you have integrated with Neptune, you can:

    • Create and track experiments

    • Manage and run experiments

    • Fetching experiment and project data

Get Started

  • New user? Register and climb aboard.

  • Registered already? Log in here, then click Getting Started and follow the onboarding instructions:

    Get Started Onboarding

  • Take a look at Neptune Project starter code in our sample project.

Track, Organize, Collaborate

image

The Neptune workflow comprises three iterative phases:

  • Track all objects in the data science or machine learning project. It can be model training curves, visualizations, input data, calculated features and so on. The snippet below presents an example of integration with Python code.

    import neptune

neptune.init('shared/onboarding', api_token='ANONYMOUS')
neptune.create_experiment()

neptune.append_tag('minimal-example')
n = 117
for i in range(1, n):
    neptune.send_metric('iteration', i)
    neptune.send_metric('loss', 1/i**0.5)
    neptune.set_property('n_iterations', n)

    Note

    The api_token belongs to the public user Neptuner. After running the code, your experiment will appear on the experiments dashboard.

    For more information, see Experiment Tracking.

  • Organize the structure of your project:

    • Code

    • Notebooks

    • Experiment results

    • Model weights

    • Meeting notes

    • Reports

    Everything is in one place, accessible from the application or programmatically. Neptune exposes a Query API, that allows users to access their Neptune data right from the Python code.

    For more information, see Experiments View.

  • Collaborate in the team:

    • Share your experiments

    • Compare results

    • Comment and communicate your work

    • Use widgets and mentions to show your progress

    • Speak your language in our data-science focused interactive wiki!

      image

    For more information, see Collaborating in Neptune.

More Resources

In addition to this documentation set, check out the following resources:

Spread the Love

Go ahead and mention us on social media!

  • Twitter: Tweet us. Our handle is @neptune.ai.

  • Product feedback: File an issue or suggest a feature or improvement in our GitHub repo.