Deploying a New Model¶
You can deploy a model either from the Model Gallery or the Model Deployment page. Below are the configuration parameters explained:
-
Select Model: Choose the model you want to deploy (e.g., DeepSeek-R1). Use the dropdown menu to quickly find a model that matches your business needs and task scenarios.
-
Model Service Name: Must meet the following requirements:
- Length: Between 2 and 64 characters
- Allowed characters: Lowercase letters, numbers, and hyphens (
-) - Naming rule: Must start and end with a lowercase letter or number
- Examples:
text-gen-service,model-01
-
Region:
- Select the deployment region for your service (e.g., “Shanghai Zone 7”)
- Choose the region based on your business coverage and latency requirements
-
Instance Count:
- Configure the number of instances to deploy (default: 1)
- Note: More instances improve concurrency but also increase cost