Skip to main content

Training Cluster

The Training Cluster provides a dedicated environment for running training workloads with predefined allocations of RAM, CPU, and GPU resources. Pricing for a Training Cluster is fixed and does not vary based on the percentage of resource utilization. Additionally, creating and running deployments within a Training Cluster does not incur any extra charges.

Create a Training Cluster​

To create a new Training Cluster:

  1. From the sidebar, select Training Cluster.
  2. Click the Create Cluster button.
  3. Enter a Cluster Name.
    • A random name (based on the current timestamp) is provided by default, but you can modify it as needed.
  4. Select the desired Cluster Configuration plan.
    • You can choose between an E2E-provided Cluster or your own Private Cluster, depending on your requirements.
    • Configuration details include:
      • Number of Nodes
      • GPU Memory
      • RAM
      • CPU
      • GPU
      • Price
  5. After selecting a plan, click Create to launch your Training Cluster.

Once the cluster is created, it will appear in the Training Cluster list, showing its configuration and current status.

Manage Training Cluster​

  • Upgrade Plan – Select the Update Plan option under the Actions to modify your existing cluster plan or configuration.
  • Terminate Cluster – Select the Terminate Cluster option under the Actions to delete your Training Cluster when it is no longer required.

Overview​

In the Overview section, you can view detailed information about your Training Cluster, including:

  • Cluster Name
  • Number of Nodes
  • Plan Name
  • Cluster Node Configuration, displaying the allocated GPU, CPU, and RAM resources.

This view provides a quick summary of your cluster’s setup and resource distribution.

Monitoring​

The Monitoring section provides real-time insights into your cluster’s performance metrics.
You can track the following parameters for each node within the Training Cluster:

  • Disk Usage
  • Memory Usage
  • GPU Utilization
  • GPU Memory Utilization
  • GPU Temperature
  • GPU Power Usage
  • CPU Utilization
  • Memory Utilization
  • Disk Total Read Bytes
  • Disk Total Write Bytes

These metrics help ensure optimal utilization and health of your Training Cluster resources.