Migrating Custom Vision Trial to Azure Subscription
Microsoft has been providing trial of custom vision AI that provides capabilities to train the system to detect object or classify them as required. However, the trial works only for a month and transferring it to the Azure Subscription is quite a hassle.
Here, we are going to migrate the trial subscription to the Azure Subscription.
This is a project where Seonjin Jeon and her team is working on detecting the handrail in images for the visually impaired people to help them navigate more easily.
Here, Seonjin has a Custom Vision project hosted in Trial Account. Since her trial is expiring soon, she wants to migrate it to another azure subscription.
When You click on the Handrail Project, you get a screen as below. You need to click on the Settings icon on the upper right corner.
Clicking the Settings on upper right corner, in the screen above, you will come to the below screen. This screen shows your training key, project ID and Endpoints.
On the bottom left corner, you can see the “Move to Azure” Button. Clicking this button will show following screen.
Like the message says, you can not move projects hosted in locations other than south central US. This wont be a problem because last time I checked you can create custom vision AI resources only in South Central US.
Ideally this should show my resources from my full subscription. But it did not. So, Click on move to different Azure Directory. When you check that button, you get additional text boxes. The drop down will show you a target directory. Select your target directory there. In the Reource Id, you need to copy the resource ID from the Azure.
If you login to the Azure, you can get your resource ID as below.
Copy the resource ID and paste it in the window below.
If things work fine you are greeted by the below screen.
Wait for couple of minutes. And you can see the screen below where your trial account is migrated to a full one.
Your Resources from Trial Subscription is migrated to Azure Now. Go Drink a Beer !!
I am really thankful to Seonjin for sharing me the screen shots from her project and allowing me to use them in the blogpost.