How to Choose a Docker Template

  1. Select a GPU:

    • In the Rent & Run module, after selecting a GPU, you will be prompted to choose a Docker template. This template will define the environment in which your task will run.

  2. Browse Available Templates:

    • The Grafi Cloud Console offers a variety of pre-configured templates based on popular AI and machine learning frameworks. Templates are categorized by use case, such as:

      • AI Model Training: Includes TensorFlow, PyTorch, and Keras.

      • Data Processing: Includes OpenCV, Pandas, and NumPy.

      • Rendering: Includes Blender and other rendering tools.

  3. Select the Appropriate Template:

    • Choose the template that best fits your task. For instance, if you are training an AI model, select a template with TensorFlow or PyTorch pre-installed.

  4. Deploy Your Application:

    • Once the template is selected, the system automatically pulls the Docker image and prepares the environment on the GPU. You can now upload your data or application script and run your task.

Custom Docker Templates:

  • Create Your Own: If none of the available templates fit your specific needs, Grafi Cloud allows you to create and upload a custom Docker template. This gives you full control over the environment configuration, ensuring that all necessary dependencies are included.

  • Upload to Grafi Cloud: Once you’ve created a custom Docker image, you can upload it to Grafi Cloud for use in future tasks.

Last updated