Updated on 2024-05-30 GMT+08:00

Creating a Training Job Configuration

Function

This API is used to create a training job configuration.

URI

POST /v1/{project_id}/training-job-configs

Table 1 describes the required parameters.
Table 1 Parameters

Parameter

Mandatory

Type

Description

project_id

Yes

String

Project ID. For details about how to obtain a project ID, see Obtaining a Project ID and Name.

Request Body

Table 2 describes the request parameters.
Table 2 Parameters

Parameter

Mandatory

Type

Description

config_name

Yes

String

Name of a training job configuration. The value must contain 1 to 64 characters consisting of only digits, letters, underscores (_), and hyphens (-).

config_desc

No

String

Description of a training job configuration. The value must contain 0 to 256 characters. By default, this parameter is left blank.

worker_server_num

Yes

Integer

Number of workers in a training job. Obtain the maximum value from Querying Job Resource Specifications.

app_url

Yes

String

Code directory of a training job, for example, /usr/app/. This parameter must be used together with boot_file_url. After setting model_id, you do not need to set app_url or boot_file_url, and engine_id.

boot_file_url

Yes

String

Boot file of a training job, which needs to be stored in the code directory, for example, /usr/app/boot.py. This parameter must be used together with app_url. After setting model_id, you do not need to set app_url or boot_file_url, and engine_id.

model_id

Yes

Long

Model ID of a training job. Obtain model_id by calling the API described in Querying a Built-in Algorithm. After setting model_id, you do not need to set app_url or boot_file_url, and engine_id.

parameter

No

Array<Object>

Running parameters of a training job. It is a collection of label-value pairs. For details, see the sample request. This parameter is a container environment variable when a training job uses a custom image. For details, see Table 4.

spec_id

Yes

Long

ID of the resource specifications selected for a training job. Obtain the ID by calling the API described in Querying Job Resource Specifications.

data_url

No

String

OBS URL of the dataset required by a training job, for example, /usr/data/.

This parameter cannot be used together with data_source or dataset_id and dataset_version_id. However, one of the parameters must exist.

dataset_id

No

String

Dataset ID of a training job. This parameter must be used together with dataset_version_id, but cannot be used together with data_url or data_source.

dataset_version_id

No

String

Dataset version ID of a training job. This parameter must be used together with dataset_id, but cannot be used together with data_url or data_source.

data_source

No

JSON Array

Dataset of a training job. This parameter cannot be used together with data_url or dataset_id and dataset_version_id. For details, see Table 3.

engine_id

Yes

Long

ID of the engine selected for a training job. The default value is 1. After setting model_id, you do not need to set app_url or boot_file_url, and engine_id. Obtain the ID by calling the API described in Querying Job Engine Specifications.

train_url

No

String

OBS URL of the output file of a training job. By default, this parameter is left blank. Example value: /usr/train/

log_url

No

String

OBS URL of the logs of a training job. By default, this parameter is left blank. Example value: /usr/train/

user_image_url

No

String

SWR URL of a custom image used by a training job. Example value: 100.125.5.235:20202/jobmng/custom-cpu-base:1.0

user_command

No

String

Boot command used to start the container of a custom image of a training job. The format is bash /home/work/run_train.sh python /home/work/user-job-dir/app/train.py {python_file_parameter}. The boot script run_train.sh must be invoked to initialize variables, such as the AK/SK. The run_train.sh script is followed by python to ensure that the Python files can be executed in the initialized variable environment. run_train.sh is used to start Python.

Table 3 data_source parameters

Parameter

Mandatory

Type

Description

dataset_id

No

String

Dataset ID of a training job. This parameter must be used together with dataset_version_id, but cannot be used together with data_url.

dataset_version

No

String

Dataset version ID of a training job. This parameter must be used together with dataset_id, but cannot be used together with data_url.

type

No

String

Dataset type. The value can be obs or dataset. obs and dataset cannot be used at the same time.

data_url

No

String

OBS bucket path. This parameter cannot be used together with dataset_id or dataset_version.

Table 4 parameter parameters

Parameter

Mandatory

Type

Description

label

No

String

Parameter name.

value

No

String

Parameter value.

Response Body

Table 5 describes the response parameters.
Table 5 Parameters

Parameter

Type

Description

is_success

Boolean

Whether the request is successful

error_message

String

Error message of a failed API call.

This parameter is not included when the API call succeeds.

error_code

String

Error code of a failed API call. For details, see Error Codes.

This parameter is not included when the API call succeeds.

Sample Request

The following shows how to create a training job configuration whose name is testConfig and description is This is config.
POST    https://endpoint/v1/{project_id}/training-job-configs
{
    "config_name": "testConfig",
    "config_desc": "This is config",
    "worker_server_num": 1,
    "app_url": "/usr/app/",
    "boot_file_url": "/usr/app/boot.py",
    "parameter": [
        {
            "label": "learning_rate",
            "value": "0.01"
        },
        {
            "label": "batch_size",
            "value": "32"
        }
    ],
    "spec_id": 1,
    "dataset_id": "38277e62-9e59-48f4-8d89-c8cf41622c24",
    "dataset_version_id": "2ff0d6ba-c480-45ae-be41-09a8369bfc90",
    "engine_id": 1,
    "train_url": "/usr/train/",
    "log_url": "/usr/log/",
    "model_id": 1
}

Sample Response

  • Successful response
    {
        "is_success": true
    }
  • Failed response
    {
        "is_success": false,
        "error_message": "Error string",
        "error_code": "ModelArts.0105"
    }

Status Code

For details about the status code, see Status Code.