Copy and back up your Custom Vision projects

After you've created and trained a Custom Vision project, you may want to copy your project to another resource. If your app or business depends on a Custom Vision project, we recommend you copy your model to another Custom Vision account in another region. Then if a regional outage occurs, you can access your project in the region where it was copied.

The ExportProject and ImportProject APIs enable this scenario by allowing you to copy projects from one Custom Vision account into others. This guide shows you how to use these REST APIs with cURL. You can also use an HTTP request service, like the REST Client for Visual Studio Code, to issue the requests.

Tip

For an example of this scenario using the Python client library, see the Move Custom Vision Project repository on GitHub.

Tip

If you want to move your Custom Vision project to Azure AI Vision Image Analysis 4.0, see the Migration guide. To compare the two services, see the Comparison page.

Prerequisites

  • Two Azure AI Custom Vision resources. If you don't have them, go to the Azure portal and create a new Custom Vision resource.
  • The training keys and endpoint URLs of your Custom Vision resources. You can find these values on the resource's Overview tab on the Azure portal.
  • A created Custom Vision project. See Build a classifier for instructions on how to do this.

Process overview

The process for copying a project consists of the following steps:

  1. First, you get the ID of the project in your source account you want to copy.
  2. Then you call the ExportProject API using the project ID and the training key of your source account. You'll get a temporary token string.
  3. Then you call the ImportProject API using the token string and the training key of your target account. The project will then be listed under your target account.

Get the project ID

First call GetProjects to see a list of your existing Custom Vision projects and their IDs. Use the training key and endpoint of your source account.

curl -v -X GET "{endpoint}/customvision/v3.3/Training/projects"
-H "Training-key: {training key}"

You'll get a 200\OK response with a list of projects and their metadata in the body. The "id" value is the string to copy for the next steps.

[
  {
    "id": "00000000-0000-0000-0000-000000000000",
    "name": "string",
    "description": "string",
    "settings": {
      "domainId": "00000000-0000-0000-0000-000000000000",
      "classificationType": "Multiclass",
      "targetExportPlatforms": [
        "CoreML"
      ],
      "useNegativeSet": true,
      "detectionParameters": "string",
      "imageProcessingSettings": {
        "augmentationMethods": {}
      }
    },
    "created": "string",
    "lastModified": "string",
    "thumbnailUri": "string",
    "drModeEnabled": true,
    "status": "Succeeded"
  }
]

Export the project

Call ExportProject using the project ID and your source training key and endpoint.

curl -v -X GET "{endpoint}/customvision/v3.3/Training/projects/{projectId}/export"
-H "Training-key: {training key}"

You'll get a 200/OK response with metadata about the exported project and a reference string "token". Copy the value of the token.

{
  "iterationCount": 0,
  "imageCount": 0,
  "tagCount": 0,
  "regionCount": 0,
  "estimatedImportTimeInMS": 0,
  "token": "string"
}

Tip

If you get an "Invalid Token" error when you import your project, it could be that the token URL string isn't web encoded. You can encode the token using a URL Encoder.

Import the project

Call ImportProject using your target training key and endpoint, along with the reference token. You can also give your project a name in its new account.

curl -v -G -X POST "{endpoint}/customvision/v3.3/Training/projects/import"
--data-urlencode "token={token}" --data-urlencode "name={name}"
-H "Training-key: {training key}" -H "Content-Length: 0"

You'll get a 200/OK response with metadata about your newly imported project.

{
  "id": "00000000-0000-0000-0000-000000000000",
  "name": "string",
  "description": "string",
  "settings": {
    "domainId": "00000000-0000-0000-0000-000000000000",
    "classificationType": "Multiclass",
    "targetExportPlatforms": [
      "CoreML"
    ],
    "useNegativeSet": true,
    "detectionParameters": "string",
    "imageProcessingSettings": {
      "augmentationMethods": {}
    }
  },
  "created": "string",
  "lastModified": "string",
  "thumbnailUri": "string",
  "drModeEnabled": true,
  "status": "Succeeded"
}

Next steps

In this guide, you learned how to copy and move a project between Custom Vision resources. Next, explore the API reference docs to see what else you can do with Custom Vision.