Platform For AI:Quick Start for Elastic Algorithm Service (EAS)

Prerequisites

• Active Alibaba Cloud account with root permissions

• PAI service activated in your workspace

• Basic familiarity with cloud computing concepts

Use your root account to activate PAI and create a workspace. Log on to the PAI console, select a region in the top-left corner, and then complete the one-click authorization and product activation.

Billing

This topic uses public resources to create a model service. The billing method is pay-as-you-go. For more information about billing rules, see EAS billing.

Preparations

To deploy a model service, you typically need to prepare model files and code files, such as web interfaces. If the Alibaba Cloud Image for the platform does not meet your deployment requirements, you must also build your own runtime image.

# Download the model
from modelscope import snapshot_download
model_dir = snapshot_download('Qwen/Qwen3-0.6B')

from flask import Flask

app = Flask(__name__)

@app.route('/hello/model')
def hello_world():
    # You can call the model here to get results.
    return 'Hello World'

if __name__ == '__main__':
    app.run(host='0.0.0.0', port=8000)

Service deployment

1. Log on to the PAI console. Select a region on the top of the page. Then, select the desired workspace and click Elastic Algorithm Service (EAS).

2. Click Deploy Service. In the Custom Model Deployment section, click Custom Deployment.

3. Configure the deployment parameters. Configure the key parameters as follows and keep the default values for the other parameters.

   • Deployment Method: Image-based Deployment.

   • Image Configuration: In the Alibaba Cloud Image list, select vllm:0.8.5.post1-mows0.2.1.

   • Mount Storage: This topic stores the model file in OSS at the path oss://examplebucket/models/Qwen/Qwen3-0___6B. Therefore, select OSS and configure it as follows.

     • Uri: The OSS path where the model is located. Set this to oss://examplebucket/models/.

     • Mount Path: The destination path in the service instance where the file is mounted, such as /mnt/data/.

   • Command: The Alibaba Cloud Image has a default startup command associated with it. You can modify it as needed. For this example, change it to vllm serve /mnt/data/Qwen/Qwen3-0___6B.

   • Resource Type: Select Public Resources. Set Resource Specification to ecs.gn7i-c16g1.4xlarge. If you want to use other resource types, see Resource configurations.

4. Click Deploy. The service deployment takes about 5 minutes. When the Service Status changes to Running, the service is successfully deployed.

   Tip: Monitor the deployment progress in the deployment logs for real-time status updates.

Online debugging

After the service is deployed, you can use the online debugging feature to test whether the service is running correctly. You can configure the request method, request path, and request body based on your specific model service.

The online debugging method for the service deployed in this topic is as follows:

1. On the Inference Service tab, click the destination service to go to the service overview page. Switch to the Online Debugging tab.

2. In the Online Debugging Request Parameters section of the debugging page, set the request parameters and click Send Request. The request parameters are as follows:

   • Chat interface: Append /v1/chat/completions to the existing URL.

   • Headers: Add a request header. Set the key to Content-Type and the value to application/json.

     

   • Body:

     {
       "model": "/mnt/data/Qwen/Qwen3-0___6B",
       "messages": [
         {
           "role": "user",
           "content": "Hello!"
         }
       ],
       "max_tokens": 1024
     }

3. The response is shown in the following figure.

   

   Validation Checklist:

   • ✓ HTTP 200 status code received

   • ✓ Response contains generated text

   • ✓ Response time < 5 seconds for simple queries

Service invocation

curl http://16********.cn-hangzhou.pai-eas.aliyuncs.com/api/predict/****/v1/chat/completions \
-H "Content-Type: application/json" \
-H "Authorization: *********5ZTM1ZDczg5OT**********" \
-X POST \
-d '{
  "model": "/mnt/data/Qwen/Qwen3-0___6B",
  "messages": [
    {
      "role": "user",
      "content": "Hello!"
    }
  ],
  "max_tokens": 1024
}' 

import requests

# Replace with the actual endpoint.
url = 'http://16********.cn-hangzhou.pai-eas.aliyuncs.com/api/predict/***/v1/chat/completions'
# For the header, set the value of Authorization to the actual token.
headers = {
    "Content-Type": "application/json",
    "Authorization": "*********5ZTM1ZDczg5OT**********",
}
# Construct the service request based on the data format required by the specific model.
data = {
  "model": "/mnt/data/Qwen/Qwen3-0___6B",
  "messages": [
    {
      "role": "user",
      "content": "Hello!"
    }
  ],
  "max_tokens": 1024
}
# Send the request.
resp = requests.post(url, json=data, headers=headers)
print(resp)
print(resp.content)

Stop or delete the service

This topic uses public resources to create the EAS service, which is billed on a pay-as-you-go basis. When you no longer need the service, stop or delete it to avoid further charges.

Management Tips: For temporary pauses, use "Stop" to preserve configuration. For permanent removal, use "Delete" to free all resources.

References

• To improve LLM service efficiency, see Deploy an LLM intelligent router.

• For more information about EAS features, see EAS overview.