Creating an MLflow Deploy instance

The MLflow Deploy service provides the ability to automatically package ML models into Docker containers and make them available via REST API to solve real-time service tasks.

The service is integrated with the Cloud ML Platform components: JupyterHub and MLflow.

Creating MLflow Deploy instances is available both through your VK Cloud personal account and via MLflow Client.

1. Create a JupyterHub instance.
2. Create an MLflow instance.

1. Go to your VK Cloud personal account.

2. Go to ML Platform.

3. In the MLflow Deploy Instance block, click the Create Instance button.

4. Set up the instance configuration:

   • Instance name: a name of the instance. It also sets the OS hostname parameter.
   • Virtual machine category: a category of the preset VM configurations. More details in the review of the Cloud Servers service.
   • Virtual machine type: a preset VM configuration (CPU and RAM).
   • Availability zone: the data center where the instance will be launched.
   • Disk size: the VM disk size in GB.
   • Disk type: the VM disk type.
   • MLflow instance: the MLflow instance which will be connected with the MLflow Deploy instance.

5. Click the Next Step button.

6. Set up the network:

   • Network: select an existing network or create a new one.
   • Virtual machine key: a key for decrypting the administrator password. Select an existing key or create a new one.

7. Click the Create Instance button.