Resource Specifications
Overview
The Resource Specifications feature allows platform administrators to pre-define compute resource specification templates, including CPU, memory, disk, and GPU configurations. When users create container instances (such as Notebooks, inference instances, or fine-tuning instances), they can directly select a pre-defined specification instead of manually filling in resource parameters, simplifying the workflow and standardizing resource usage.
Use Cases
- Standardized resource allocation: Define standard specifications for different workload types (e.g., development, training, inference)
- Simplified user experience: Users don't need to understand low-level resource configuration details — just select a specification
- Cost control: Limit the maximum resources users can request through pre-defined specifications
Creating Resource Specifications
- Log in as a platform administrator.
- Navigate to Platform Management → Resource Specifications.
- Click the Create Specification button.
- Fill in the specification details:
| Parameter | Description |
|---|---|
| Name | Custom name for easy identification (e.g., "Dev-4C8G", "Training-A100-80G") |
| CPU | CPU core count configuration |
| Memory | Memory size configuration (GiB) |
| GPU Model | Optional; specify a GPU model (e.g., A100, V100) |
| GPU Count | Optional; number of GPU cards |
| GPU Memory | Optional; VRAM allocation per GPU card |
- Click OK to complete creation.
Managing Specifications
In the specification management list, you can perform the following operations on existing specifications:
- View: View detailed configuration of a specification
- Edit: Modify resource configuration parameters
- Delete: Remove specifications that are no longer needed
Note
Using Specifications
After specifications are created, users can select them in the following scenarios:
- When creating a Notebook development instance, select a specification in the Resource Configuration step
- When creating a dedicated model inference instance, select a compute specification
- When creating a model fine-tuning instance, select a GPU resource specification
- When creating an application Space, select a cloud resource specification
After selecting a specification, the system automatically populates the corresponding CPU, memory, GPU, and other resource configuration parameters.