All Products
Search
Document Center

Platform For AI:Custom images

Last Updated:Mar 11, 2026

Register container images from ACR or build new images in PAI for training jobs.

Prerequisites

  • Alibaba Cloud account: Activate PAI and create a workspace.

  • ACR setup: For Enterprise Edition images, create an ACR Enterprise Edition instance with a namespace and repository.

  • Permissions:

Methods

PAI provides two methods for adding custom images:

Method

Use case

Complexity

Import existing image

Reuse existing images

Medium—manual ACR setup required

Build new image

Create images from scratch in PAI

Low—one-stop workflow

  • Import existing image: Build an image locally or on ECS, push to ACR, and register in PAI.

  • Build new image: Configure ACR destination, runtime, and build settings in PAI console. PAI builds, pushes, and registers the image automatically.

Import an existing image

Register an image pushed to ACR.

Procedure

  1. Open the PAI console and select your workspace.

  2. Navigate to AI Asset Management > Images.

  3. On the Custom Image tab, click Register Image, and select Import Existing Image.

  4. Configure parameters:

Register Image dialog - Import Existing Image

Parameter

Description

Image Type

Enterprise Edition Instance / Namespace / Repository / Version / Custom Domain Name

Select an instance, namespace, repository, version, and custom domain name. To create these resources, go to the Container Registry console.

Note
  • Enterprise Edition Instance and Custom Domain Name are required only for Enterprise Edition images.

  • If permission errors occur, grant AliyunContainerRegistryReadOnlyAccess to your RAM user. See Manage RAM user permissions.

Visibility

  • Visible to Me: Only you and workspace administrators can see this image.

  • Visible to Current Workspace: All workspace members can see this image.

Chip Type

Select CPU or GPU to match your target hardware.

  1. Click Confirm to register the image.

Build a new image

Build a custom image in PAI. Only Enterprise Edition images can be built.

Procedure

  1. Open the PAI console and select your workspace.

  2. Navigate to AI Asset Management > Images.

  3. On the Custom Image tab, click Register Image, and select Build New Image.

  4. Configure parameters in Basic Information:

Register Image dialog - Build New Image

Parameter

Description

Basic Information

Image Type

Only Enterprise Edition images can be built. See What is Container Registry (ACR)?.

Enterprise Edition Instance / Namespace / Repository / Custom Domain Name

Select an instance, namespace, repository, and custom domain name. To create these resources, go to the Container Registry console.

Note

If permission errors occur, grant AliyunContainerRegistryReadOnlyAccess to your RAM user. See Manage RAM user permissions.

Visibility

  • Visible to Me: Only you and workspace administrators can see this image.

  • Visible to Current Workspace: All workspace members can see this image.

Chip Type

Select CPU or GPU to match your target hardware.

Runtime Information

Resources

Compute resources for the build. Use higher specifications for Dockerfiles with compilation steps.

VPC Configuration / Security Group

VPC bound to your Enterprise Edition instance and security group.

Note

For public internet access to download dependencies, create an Internet NAT gateway, attach an EIP, and configure SNAT rules. See Use a dedicated gateway to improve public network access speed.

Build Configuration

Build Method

  • Based on Existing Image: Start from an official, custom, or third-party image. Install dependencies with pip2, pip3, yum, or apt. For private repositories, provide authentication credentials.

  • Based on Dockerfile: Write custom Dockerfile instructions. For syntax details, see Dockerfile reference.

  1. Configure Runtime Information.

  2. Configure Build Configuration and select build method.

  3. Click Confirm to start the build.

Monitor build jobs

Track progress and troubleshoot build issues.

  1. On the Custom Image tab, click View Jobs.

  2. Locate your build and check the status.

    View Jobs interface showing build job list

  3. Click View Log to see the build output and debug issues.

  4. Click View Monitoring to see the resource utilization for the build job.

    Job details page with View Log and View Monitoring

Use custom images

Use registered custom images for training jobs.

  1. On the Custom Image tab, locate your image and copy the registry address.

    Custom Image tab showing registered images

  2. Use this address when creating training jobs. See Create a training job.

Reference