Manage pipeline images¶

Images are reusable container environments for pipelines. An image carries the third-party Python packages a pipeline needs (for example, scikit-learn, pandas, or scipy) and can be shared across pipelines and runs. Images live on the Images tab, alongside the pipeline list.

The following table describes the columns on the Images list.

Expand the for an image in the list to rebuild or delete it.

Review an image¶

Open an image to review package dependencies, build status, and build logs.

The following table describes the fields on the image page.

Create an image¶

On the Images tab, click + Add image to define a container environment that pipelines can share.

1. Enter a Name and, optionally, a Description.

2. In the YAML editor, list the Python packages to pip-install. Uncomment pythonVersion or gpu only if a different Python version or a GPU image is required.

   image.yaml

   packages:
     - "scikit-learn>=1.3"
     - "numpy>=1.26"
   # pythonVersion: "3.12"
   # gpu: false
   

3. Click Create.

The image builds in the background and appears in the Images list. A run cannot start until Status is Ready.

Next steps¶

After managing images, attach one to a new pipeline or run a pipeline with an existing image.

• Add a pipeline: Select an existing image or create a new one when adding a pipeline.
• Run and schedule pipelines: Choose an image when starting a run.
• View and manage pipelines: Open a pipeline that uses the image.