The ALTER_MODEL_PROFILE procedure is used to create or update model parameters at the provider/model level. If the corresponding record does not exist, it is created; if it exists, it is updated.
Note
This process is supported starting from V4.6.2.1.
Syntax
PROCEDURE alter_model_profile(
IN model VARCHAR(512),
IN json_params JSON);
Parameters
Parameter |
Description |
Type |
Value range |
Nullable |
|---|---|---|---|---|
| model | The full model name, in the format of provider/model_name. provider/model By the first / Split, model_name Some parts can continue to include /. |
VARCHAR(512) | Up to 512 bytes | NO |
| json_params | Model parameters, which must be a JSON object. | JSON | See the following table. | NO |
The json_params top-level fields are as follows:
field |
Type |
Whether required |
Description |
|---|---|---|---|
model_config |
JSON OBJECT | No | Model API call parameters. |
run_config |
JSON OBJECT | No | OceanBase AI Database runtime options. |
Currently supported fields of run_config:
field |
Type |
Description |
|---|---|---|
batch_size |
INT | The batch processing size. |
max_image_size |
INT | Maximum number of bytes for image input. |
min_concurrency |
INT | Adaptive concurrency lower limit, default: 10. |
max_concurrency |
INT | Adaptive concurrency cap, default: 100. |
Examples
CALL DBMS_AI_SERVICE.ALTER_MODEL_PROFILE('aliyun/qwen-plus', '{
"model_config": {
"max_tokens": 4096,
"temperature": 0,
"enable_thinking": false
},
"run_config": {
"batch_size": 16,
"max_image_size": 4194304,
"min_concurrency": 10,
"max_concurrency": 100
}
}');