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

Querying a Built-in Algorithm

Function

This API is used to obtain the details about a built-in model.

URI

GET /v1/{project_id}/built-in-algorithms

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 Query parameters

Parameter

Mandatory

Type

Description

per_page

No

Integer

Number of job parameters displayed on each page. The value range is [1, 100]. Default value: 10

page

No

Integer

Index of the page to be queried. Default value: 1

sortBy

No

String

Sorting mode of the query. The value can be engine, model_name, model_precision, model_usage, model_precision, model_size, create_time, or parameter. Default value: engine

order

No

String

Sorting order. Options:

  • asc: ascending order
  • desc: descending order. The default value is desc.

search_content

No

String

Search content, for example, a parameter name. By default, this parameter is left blank.

Response Body

Table 3 describes the response parameters.
Table 3 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.

model_total_count

Integer

Number of models

models

Array<Object>

Model parameter list. For details, see Table 4.

Table 4 models structure data

Parameter

Type

Description

model_id

Integer

Model ID

model_name

String

Model name

model_usage

Integer

Model usage. Options:

  • 1: image classification
  • 2: object class and location
  • 3: image semantic segmentation
  • 4: natural language processing
  • 5: image embedding

model_precision

String

Model precision

model_size

Long

Model size, in bytes

model_train_dataset

String

Model training dataset

model_dataset_format

String

Dataset format required by a model

model_description_url

String

URL of the model description

parameter

String

Running parameters of a model. This parameter is a container environment variable when a training job uses a custom image. For details, see the sample request.

create_time

Long

Time when a model is created

engine_id

Long

Engine ID of a model

engine_name

String

Engine name of a model

engine_version

String

Engine version of a model

Table 5 parameter parameters

Parameter

Type

Description

label

String

Parameter name

value

String

Parameter value

required

Boolean

Whether a parameter is mandatory

Sample Request

The following shows how to obtain the algorithm whose name contains configname.

GET https://endpoint//v1/{project_id}/built-in-algorithms?per_page=10&page=1&sortBy=engine&order=asc&search_content=model

Sample Response

  • Successful response
    {
        "models": [
            {
                "model_id": 4,
                "model_name": "ResNet_v2_50",
                "model_usage": 1,
                "model_precision": "75.55%(top1), 92.6%(top5)",
                "model_size": 102503801,
                "model_train_dataset": "ImageNet, 1,000 classes for image classification",
                "model_dataset_format": "shape: [H>=32, W>=32, C>=1]; type: int8",
                "model_description_url": "https://github.com/apache/incubator-mxnet/blob/master/example/image-classification/symbols/resnet.py",
                "parameter": "[{\"label\":\"batch_size\",\"value\":\"4\",\"placeholder_cn\":\"Total number of training images updated each time\",\"placeholder_en\":\"\",\"required\":true},{\"label\":\"lr\",\"value\":\"0.0001\",\"placeholder_cn\":\"Learning rate\",\"placeholder_en\":\"\",\"required\":true},{\"label\":\"save_frequency\",\"value\":\"1\",\"placeholder_cn\":\"Interval for saving the model, indicating that the model is saved every N epochs\",\"placeholder_en\":\"\",\"required\":true},{\"label\":\"num_classes\",\"value\":\"\",\"placeholder_cn\":\"Total number of image classes in training\",\"placeholder_en\":\"\",\"required\":true},{\"label\":\"num_epoch\",\"value\":\"10\",\"placeholder_cn\":\"Number of training epochs\",\"placeholder_en\":\"\",\"required\":true}]",
                "create_time": 1522218780025,
                "engine_id": 501,
                "engine_name": "MXNet",
                "engine_version": "MXNet-1.2.1-python2.7"
            },
            {
                "model_id": 5,
                "model_name": "Faster_RCNN_ResNet_v2_101",
                "model_usage": 2,
                "model_precision": "80.05%(mAP)",
                "model_size": 190936449,
                "model_train_dataset": "PASCAL VOC2007, 20 classes for object detection",
                "model_dataset_format": "shape: [H, W, C==3]; type: int8",
                "model_description_url": "https://github.com/apache/incubator-mxnet/tree/master/example/rcnn",
                "parameter": "[{\"label\":\"lr\",\"value\":\"0.0001\",\"placeholder_cn\":\"Learning rate\",\"placeholder_en\":\"\",\"required\":true},{\"label\":\"eval_frequence\",\"value\":\"1\",\"placeholder_cn\":\"Frequency for validating the model. By default, validation is performed every epoch.\",\"placeholder_en\":\"\",\"required\":true},{\"label\":\"mom\",\"value\":\"0.9\",\"placeholder_cn\":\"Momentum of the training network\",\"placeholder_en\":\"\",\"required\":true},{\"label\":\"wd\",\"value\":\"0.0005\",\"placeholder_cn\":\"Weight decay coefficient\",\"placeholder_en\":\"\",\"required\":true},{\"label\":\"num_classes\",\"value\":\"\",\"placeholder_cn\":\"Total number of image classes in training. The value must plus 1 because there is a background class.\",\"placeholder_en\":\"\",\"required\":true}]",
                "create_time": 1525313224596,
                "engine_id": 501,
                "engine_name": "MXNet",
                "engine_version": "MXNet-1.2.1-python2.7"
            }
        ],
        "model_total_count": 41,
        "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.