imyaman

Hello Google Cloud AutoML Vision & AutoML Vision Edge Customer,

We’re writing to let you know that to prepare for the General Availability (GA) release of Cloud AutoML Vision and AutoML Vision Edge expected later this year, a Beta Refresh release will be rolled out between September 14, 2019 and September 29, 2019. As part of the Beta Refresh release, your Image Classification datasets will be migrated, and will be accessible through a new integrated UI.

Additionally, effective the GA release date, the pricing for all Cloud models in AutoML Vision will have the same structure as current Object Detection pricing. This will ensure easier to understand and consistent pricing across all AutoML Vision features.

If you are an ML Kit customer that uses AutoML Vision Edge exclusively via ML Kit, please note that the ML Kit console and client libraries will remain in beta. If you have not used Cloud AutoML Vision APIs or Console directly in the Google Cloud Console, the following changes won't impact your usage or pricing and you may stop reading now.

Between September 14 and September 29, we will transparently migrate your Image Classification datasets to the environment shared with Object Detection. Once your data has been migrated, your existing UI will forward to a new integrated UI available at https://console.cloud.google.com/vision/dashboard.

The following features will not be available in the new UI:

• Integrated Human labeling; alternatively, you can use Cloud AI Platform Data Labeling Service.
• Automated deployment of models; you will need to manually deploy and set the number of nodes on which your model(s) are deployed.

Furthermore, the ability to resume training in the new UI after pausing will only be applicable to models trained in the new UI.

From the GA release date, pricing for Cloud models will be charged using node hour as the unit, and corresponding free tiers will receive free credits once per billing account lifetime rather than every month.

The new pricing for Cloud models effective the GA release date is summarized below. The node hours represent an hour of processing on different node types of GPU or TPU. We use 10 nodes in parallel for training of Cloud models.

After the migration is completed on September 29, please review your projects' Models page by clicking the project links below, to ensure that only active models are deployed. To avoid charges for deployed models which you do not require, please remove their deployment.

We have identified that your projects listed below have active Cloud models and will be counted towards the lifetime credit of the free tier of their respective billing accounts:

• sys5 (sys5-f9f83)

If you have any questions or need assistance removing deployments you do not require, please join the cloud-vision-discuss Google group.

Thank you for being a valued AutoML Vision customer.