Skip to main content

Quick Start Guide

Prerequisites​

  • TIR Account: Ensure you have access to a TIR project.
  • Pipeline YAML File: A valid Argo Workflow or Kubeflow Pipelines YAML template.
  • (Optional) Docker Image: If running custom containers, have your image pushed to the E2E Container Registry.

Create Your First Pipeline​

  1. Navigate to Pipelines:

    • In your TIR project dashboard, click Pipelines > Pipeline in the sidebar.
    • You will be directed to the Manage Pipelines page.

  2. Start Pipeline Creation:

    • If this is your first pipeline, click GET STARTED to create a default hello-world pipeline.
    • Otherwise, click CREATE PIPELINE to create a custom pipeline.

  3. Choose Creation Mode:

    • Create a new pipeline — Upload a .yaml file, provide a name and description, then click UPLOAD.
    • Create a new pipeline version under an existing pipeline — Select the target pipeline, upload a .yaml file, and click UPLOAD.
    warning

    Avoid introducing additional nodes or commands in your YAML that could conflict with existing node definitions.

  4. Pipeline Created:

    • After successful creation, you will see the pipeline details page with pipeline versions.
    • From here you can create runs or view existing runs.

Create a Run​

  1. Navigate to Runs:

    • Go to Pipelines > Run in the sidebar, or click the Create Run icon from a specific pipeline.

  2. Select an Experiment:

    • Choose an existing experiment or create a new one. An experiment groups related pipelines and their run histories.

  3. Configure Run Parameters:

    • Select the pipeline version to execute.
    • Configure any run parameters defined in your YAML template.
    • Click Next.

  4. Select Resources:

    • Choose a resource plan (CPU or GPU) for your run execution.
    • Click FINISH to create the run.

  5. View Results:

    • The created run will appear in the Run section.
    • Click the run name to view execution details and logs.

API Reference​

For programmatic access to Pipelines, refer to our comprehensive API documentation:

Pipeline API Documentation

Create, manage, and execute Pipelines, Runs, and Scheduled Runs using the TIR API. Includes endpoints for pipeline upload, version management, experiment tracking, and run scheduling.

Next Steps​

  • Features — Pipeline versioning, scheduling, Docker execution, common templates, and monitoring.
  • Pricing — Resource plans and billing details.