Creating a Container Instance
A guide for creating a container instance on the AI computing platform for algorithm development and model fine-tuning.
Introduction
Container instances are often used for algorithm development and model fine-tuning. If there is a small amount of training data, you can apply for a single-card or 8-card instance. It provides a local data disk and associated file storage. You can use Jupyter for algorithm development and fine-tuning, and output the results to the mounted shared file storage. After use, download the results and release the container instance.
Prerequisites
Before creating a container instance, ensure the following prerequisites are met:
- ** Management Console** account and password are obtained.
- Personal real-name authentication is completed and the account balance is greater than 0 yuan.
Procedure
Follow these steps to create a container instance:
-
Log in to the Management Console.
-
In the top navigation bar, click:
- Products and Services > AI Computing Platform > AI Computing Platform to go to its overview page.
-
In the left navigation bar, select Container Instances. The container instance list page will display by default.
-
Click Create Container Instance. On the Create Container Instance page, configure various parameters and click OK.
Parameter Configuration:
| Parameter | Description |
|---|---|
| Instance Name | User-defined name for easy identification. |
| Resource | Configure the following depending on the resource type: |
| - Billing Mode | Default is pay-as-you-go. |
| - Resource Type | Choose from high-speed training, shared GPU, or CPU computing types. Configure GPU model, CPU model, memory, etc. |
| Storage and Data (optional) | Select the user directory where the dataset is located and the corresponding mount directory. |
| Mirror | Select an image for the instance: public, custom, or private image address. A password-protected image will require the user to input a username and password. |
-
Return to the Container Instances page and wait for the container instance to be created. Once successfully created, the instance will show a status of
running(Running). -
Log in to the container instance using web connection or Jupyter.
Conclusion
You have successfully created and accessed a container instance for algorithm development and model fine-tuning.