AI engine usage examples - - Alibaba Cloud Documentation Center

The Lindorm AI engine provides a set of RESTful APIs. You can use curl commands to perform operations such as model management and inference. This topic shows how to use curl to call these APIs.

Prerequisites

You have enabled the AI engine. For more information, see Enable the AI engine.
You have added the IP address of your client to the allowlist of your Lindorm instance. For more information, see Configure an allowlist.

Parameters

Note

The following parameters apply to all examples in this topic.

Parameter	Description	How to obtain
url	The endpoint for connecting to the AI engine.	On the Database Connection page, click the AI Engine tab to obtain the endpoint, username, and password.
username	The username and password to access the AI engine.
password	The username and password to access the AI engine.

Connection example

curl -X GET "http://ld-bp1hq5192030n****-proxy-ai-vpc.lindorm.aliyuncs.com:9002/v1/ai/models/list" \
-H "x-ld-ak: test" \
-H "x-ld-sk: test"

Important

9002 is the port number and is required.

Model management

Create a model

Create a model named bge_m3_model in the AI engine.

curl -X POST "http://<url>/v1/ai/models/create" \
-H "x-ld-ak: <username>" \
-H "x-ld-sk: <password>" \
-H "Content-Type: application/json" \
-d '{
    "model_name": "bge_m3_model",
    "model_path": "huggingface://BAAI/bge-m3",
    "task": "FEATURE_EXTRACTION",
    "algorithm": "BGE_M3"
}'

Response:

{"code": 0,"msg": "SUCCESS","data": null,"success": true}

List all models

curl -X GET "http://<url>/v1/ai/models/list" \
-H "x-ld-ak: <username>" \
-H "x-ld-sk: <password>"

Response:

{
  "code": 0,
  "msg": "SUCCESS",
  "data": {
    "models": [{
      "name": "bge_m3_model",
      "status": "READY",
      "created_time": "...",
      "updated_time": "...",
      ...
    }, {
      "name": "bge_model",
      "status": "READY",
      ...
    }]
  },
  "success": true
  "request_id":"..."
}

Get model details

curl -X GET "http://<url>/v1/ai/models/bge_m3_model/status" \
-H "x-ld-ak: <username>" \
-H "x-ld-sk: <password>"

Response:

{
  "code": 0,
  "msg": "SUCCESS",
  "data": {
    "name": "bge_m3_model",
    "status": "READY",
    "task_type":"FEATURE_EXTRACTION",
    "algorithm":"BGE_M3",
    "settings": "...",
    "error":"...",
    "progress": "...",
    ...
  },
  "success": true
}

Delete a model

Delete the bge_m3_model model.

curl -X POST "http://<url>/v1/ai/models/bge_m3_model/drop" \
-H "x-ld-ak: <username>" \
-H "x-ld-sk: <password>"

Response:

{"code": 0,"msg": "SUCCESS","data": null,"success": true}

Model inference

Run inference with an existing model

curl -X POST "http://<url>/v1/ai/models/bge_m3_model/infer" \
  -H "x-ld-ak: <username>" \
  -H "x-ld-sk: <password>" \
  -H "Content-Type: application/json" \
  -d '{"input":["你好","我们"]}' \

This returns the following inference result:

{
  "code": 0,
  "msg": "SUCCESS",
  "data": [[0.027204733341932297,0.004229982383549213, ...], 
  [-0.05367295444011688,0.022600287571549416, ...]],
  "success": true
}

Parameter adjustment

Adjust the parameters for a deployed model.

curl -X POST "http://<url>/v1/ai/models/bge_m3_model/update_config" \
  -H "x-ld-ak: <username>" \
  -H "x-ld-sk: <password>" \
  -H "Content-Type: application/json" \
  -d '{"instance_count": "2"}' \

A successful adjustment returns the following response:

{"code": 0,"msg": "SUCCESS","data": null,"success": true}