Getting Started

Get started using ClearML by installing the ClearML Python Package, running a script with ClearML, and then choosing a ClearML backend infrastructure offering: the ClearML Hosted Service, or your own self-hosted ClearML Server.

To get started using ClearML:

  1. Install the ClearML Python Package.

    pip install clearml
    
  2. Add two lines of code to your script. When the script runs, see the results in the ClearML demo server https://demoapp.demo.clear.ml/dashboard. The demo server is open to anyone and resets periodically (removing all data).

    from clearml import Task
    task = Task.init(project_name="my project", task_name="my task")
    
  3. Choose a ClearML backend infrastructure, ClearML Server.

    • Signup for a free ClearML Hosted Service account.

      The ClearML Hosted Service is the ClearML backend infrastructure maintained for you. It includes the web server (and ClearML Web UI) and file server. It is always available. You can work independently, collaborate in teams up to three, and create unlimited projects, experiments, models, and other workflows.

    • Deploy your own open source ClearML Server.

      By deploying your own ClearML Server, you get the ClearML backend infrastructure open source component to use, configure, and modify. It includes your own web server (and ClearML Web UI) and file server, as well as the ClearML Server REST API. You can create as many users as you want to create, and they can all collaborate. Use as much storage space as you choose. Configuration options include web login authentication, sub domains and load balancers, and many configuration settings.

    Both the ClearML Hosted Service and a self-hosted ClearML Server integrate with the ClearML Python Package and ClearML Agent.