Updated on 2024-08-14 GMT+08:00

Parameter Overview

You can use ModelStep to create a model registration phase. The following is an example of defining a ModelStep.

Table 1 ModelStep

Parameter

Description

Mandatory

Data Type

name

Name of a model registration phase. The name contains a maximum of 64 characters, including only letters, digits, underscores (_), and hyphens (-). It must start with a letter and must be unique in a workflow.

Yes

str

inputs

Inputs of the model registration phase

No

ModelInput or ModelInput list

outputs

Outputs of the model registration phase

Yes

ModelOutput or ModelOutput list

title

Title for frontend display

No

str

description

Description of the model registration phase

No

str

policy

Phase execution policy

No

StepPolicy

depend_steps

Dependency phases

No

Step or step list

Table 2 ModelInput

Parameter

Description

Mandatory

Data Type

name

Input name of the model registration phase. The name can contain a maximum of 64 characters, including only letters, digits, underscores (_), and hyphens (-), and must start with a letter. The input name of a step must be unique.

Yes

str

data

Input data object of the model registration phase

Yes

OBS, SWR, or subscribed model object. Currently, only OBSPath, SWRImage, OBSConsumption, OBSPlaceholder, SWRImagePlaceholder, DataConsumptionSelector, and GalleryModel are supported.

Table 3 ModelOutput

Parameter

Description

Mandatory

Data Type

name

Output name of the model registration phase. The name can contain a maximum of 64 characters, including only letters, digits, underscores (_), and hyphens (-), and must start with a letter. The output name of a step must be unique.

Yes

str

model_config

Configurations for model registration

Yes

ModelConfig

Table 4 ModelConfig

Parameter

Description

Mandatory

Data Type

model_type

Model type. Supported types: TensorFlow, MXNet, Caffe, Spark_MLlib, Scikit_Learn, XGBoost, Image, PyTorch, Template, and Custom. The default value is TensorFlow.

Yes

str

model_name

Model name. Enter 1 to 64 characters. Only letters, digits, hyphens (-), and underscores (_) are allowed.

No

str, Placeholder

model_version

Model version in the format of Digit.Digit.Digit. The value range of the digits is [1, 99]. If this parameter is left blank, the version number automatically increases.

CAUTION:

No part of the version number can start with 0. For example, 01.01.01 is not allowed.

No

str, Placeholder

runtime

Model runtime environment. The options of runtime are the same as those of model_type.

No

str, Placeholder

description

Model description that consists of 1 to 100 characters. The following special characters cannot be contained: &!'"<>=

No

str

execution_code

OBS path for storing the execution code. By default, this parameter is left blank. The name of the execution code file is fixed to customize_service.py.

The inference code file must be stored in the model directory. This parameter is left blank. The system can automatically identify the inference code in the model directory.

No

str

dependencies

Package required for the inference code and model. By default, this parameter is left blank. It is read from the configuration file.

No

str

model_metrics

Model precision, which is read from the configuration file

No

str

apis

All apis input and output parameters of a model (optional), which are parsed from the configuration file

No

str

initial_config

Model configuration information

No

dict

template

Template configuration items. This parameter is mandatory when model_type is set to Template.

No

Template

dynamic_load_mode

Dynamic loading mode. Currently, only Single is supported.

No

str, Placeholder

prebuild

Whether the model is prebuilt. The default value is False.

No

bool, Placeholder

install_type

Model installation type. The value can be real_time, edge, batch. If this parameter is left blank, all types are supported by default.

No

list[str]

Table 5 Template

Parameter

Description

Mandatory

Data Type

template_id

ID of the used template. The template has a built-in input and output mode.

Yes

str, Placeholder

infer_format

Input and output mode ID. When this parameter is used, the input and output mode built in the template does not take effect.

No

str, Placeholder

template_inputs

Template input configuration, specifying the source path for configuring a model

Yes

list of TemplateInputs object

Table 6 TemplateInputs

Parameter

Description

Mandatory

Data Type

input_id

Input item ID, which is obtained from the template details

Yes

str, Placeholder

input

Template input path, which can be an OBS file path or OBS directory path. When you use a template with multiple input items to create a model, if the target paths input_properties specified in the template are the same, the OBS directory or OBS file name entered here must be unique to prevent files from being overwritten.

Yes

str, Placeholder, Storage