ml deployment-job create

Start a deployment job asynchronously. The job can perform batch scoring, streaming, or other types of batch operations such as solving a Decision Optimization problem.

Syntax

cpd-cli ml deployment-job create \
[--context=<catalog-project-or-space-id>] \
[--cpd-config=<cpd-config-location>] \
[--cpd-scope=<cpd-scope>] \
[--custom=<map<key,value>>] \
[--decision-optimization=<input-output-data-properties>] \
--deployment=<resource-reference> \
[--description=<resource-description>] \
[--hardware-spec=<hardware-specification>] \
[--hybrid-pipeline-hardware-specs=<hybrid-pipeline-hardware-specification>] \
[--jmes-query=<jmespath-query>] \
[--name=<resource-name>] \
[--output=json|yaml|table] \
[--output-file=<output-file-location>] \
--profile=<profile-name> \
[--quiet] \
[--raw-output=true|false] \
[--retention=<job-meta-retention-value>] \
[--scoring=<input-output-data-details>] \
--space-id=<space-identifier> \
[--tags=<tag1,tag2,...>] \
[--verbose]

Arguments

The ml deployment-job create command has no arguments.

Options

Table 1: Command options
Option Description
--context Specify the configuration context
 name.
Status
Optional.
Syntax
--context=<catalog-project-or-space-id>
Default value
Not applicable.
Valid values
A valid configuration context name.
--cpd-config The IBM Software Hub
 configuration location. For example, $HOME/.cpd-cli/config.
Status
Optional.
Syntax
--cpd-config=<cpd-config-location>
Default value
$HOME/.cpd-cli/config
Valid values
A valid IBM® Software Hub configuration location.
--cpd-scope The IBM Software Hub space, project, or catalog scope. For example, cpd://default-context/spaces/7bccdda4-9752-4f37-868e-891de6c48135.
Status
Optional.
Syntax
--cpd-scope=<cpd-scope>
Default value
No default.
Valid values
A valid IBM Software Hub space, project, or catalog scope.
--custom Specify user-defined properties as key-value pairs.
Status
Optional.
Syntax
--custom=<map<key,value>>
Default value
No default.
Valid values
Valid key-value pairs.
--decision-optimization Provide details about the input and output data and other properties that are used for a batch deployment job of a decision optimization problem.
Status
Optional.
Syntax
--decision-optimization=<input-output-data-properties>
Default value
No default.
Valid values
input_data
Use the 'input_data' value to specify the input data for batch processing as part of the job's payload. The 'input_data' value is mutually exclusive with the 'input_data_references' value. When 'input_data' is specified, the processed output of batch deployment job is available in the deployment job response's 'scoring.predictions' parameter.
input_data_references
Use the 'input_data_references' value to specify the details that pertain to the remote source where the input data for batch deployment job is available. The 'input_data_references' value must be used with the 'output_data_references' value. The 'input_data_references' value is mutually exclusive with the 'input_data' property value.
output_data_references
Use the 'output_data_references' value to specify the details that pertain to the remote source where the input data for batch deployment job is available. The 'output_data_references' value must be used with the 'input_data_references' value.
--description Specify a resource
 description.
Status
Optional.
Syntax
--description=<resource-description>
Default value
No default.
Valid values
A valid resource description.
--hardware-spec Specify a hardware specification.
Status
Optional.
Syntax
--hardware-spec=<hardware-specification>
Default value
No default.
Valid values
A valid hardware specification.

--help

-h

 Display command help.
Status
Optional.
Syntax
--help
Default value
No default.
Valid values
Not applicable.
--hybrid-pipeline-hardware-specs Specify a hybrid pipeline hardware specification.
Status
Optional.
Syntax
--hybrid-pipeline-hardware-specs=<hybrid-pipeline-hardware-specification>
Default value
No default.
Valid values
A valid hybrid pipeline hardware specification.
--jmes-query Provide a JMESPath query to customize the output.
Status
Optional.
Syntax
--jmes-query=<jmespath-query>
Default value
No default.
Valid values
A valid JMESPath query.
--name Specify a resource name.
Status
Required.
Syntax
--name=<resource-name>
Default value
No default.
Valid values
A valid resource name.
--output Specify an output format.
Status
Optional.
Syntax
--output=json|yaml|text
Default value
text
Valid values
Valid formats include JSON, YAML, or text (the default format).
--output-file Specify a file path where all output is redirected.
Status
Optional.
Syntax
--output-file=<output-file-location>
Default value
No default.
Valid values
A valid output file path location.
--profile The name of the profile that you created to store information about an instance of IBM Software Hub and your credentials for the instance.
Status
Required.
Syntax
--profile=<profile-name>
Default value
No default.
Valid values

The name of the profile that you created.
--quiet Suppress verbose messages.
Status
Optional.
Syntax
--quiet
Default value
No default.
Valid values
Not applicable.
--raw-output When set to true, single values are not surrounded by quotation marks in
 JSON output mode.
Status
Optional.
Syntax
--raw-output=true|false
Default value
false
Valid values
false
Single values in JSON output mode are surrounded by quotation marks.
true
Single values in JSON output mode are not surrounded by quotation marks.
--retention Specify a value to define the number of days to retain the job meta. Job meta is automatically deleted after the specified number of days. The value '-1' sets the meta to never auto-delete.
Status
Optional.
Syntax
--retention=<job-meta-retention-value>
Default value
-1
Valid values
Any positive integer value, or -1 to signify to never auto-delete.
--scoring Specify details about the input and output data and other properties that are used for a model’s batch deployment job (Python Function or Python Scripts).
Status
Optional.
Syntax
--scoring=<input-output-data-details>
Default value
No default.
Valid values
input_data
Use the 'input_data' value to specify the input data for batch processing as part of the job's payload. The 'input_data' value is mutually exclusive with the 'input_data_references' value. When 'input_data' is specified, the processed output of batch deployment job is available in the deployment job response's 'scoring.predictions' parameter. The ‘input_data' value is not supported for batch deployment of Python Scripts.
input_data_references
Use the 'input_data_references' value to specify the details that pertain to the remote source where the input data for batch deployment job is available. The 'input_data_references' value must be used with the 'output_data_references' value. The 'input_data_references' value is mutually exclusive with the 'input_data' property value. The 'input_data_references' value is not supported for batch deployment job of Spark models and Python Functions.
output_data_references
Use the 'output_data_references' value to specify the details that pertain to the remote source where the input data for batch deployment job is available. The 'output_data_references' value must be used with the 'input_data_references' value. The 'output_data_references' value is not supported for batch deployment job of Spark models and Python Functions.
--space-id Specify a space identifier.
Status
Optional.
Syntax
--space-id=<space-identifier>
Default value
No default.
Valid values
A valid space identifier.
--tags Specify the data asset tags. Multiple tags can be specified.
Status
Optional.
Syntax
--tags=<tag1,tag2,...>
Default value
No default.
Valid values
A valid list of comma-separated data asset tags.
--verbose Logs include more detailed messages.
Status
Optional.
Syntax
--verbose
Default value
No default.
Valid values
Not applicable.

Examples

Note: The following examples use environment variables. Use a script to create environment variables with the correct values for your environment. You can add the listed environment variables to the installation variables script. For more information, see Setting up installation environment variables.
Specify the environment variables, the deployment job name, and the scoring parameters in a script file.
Replace the following properties with your data.
deployment_job_name = 'script_batch_deployment_job'

deployment = {
    'id': deployment_id
}
deployment_json = json.dumps(deployment)

scoring = {
    "input_data_references": [
      {
        "type": "data_asset",
        "connection": {},
        "location": {
          "href": "/v2/assets/783d8fc5-ae2c-47d0-a311-f8890dfa1ce0?space_id=d9bfa660-1be7-46ab-aa53-e9010a634bba"
        }
      }
    ],
    "output_data_reference": {
      "type": "data_asset",
      "connection": {},
      "location": {
        "href": "/v2/assets/c7de9cb4-0ec5-41fa-8b94-3e89eb2cb795?space_id=d9bfa660-1be7-46ab-aa53-e9010a634bba"
      }
    }
}
scoring_json = json.dumps(scoring)
space_id = <The space identifier>
Create a deployment job.
cpd-cli ml deployment-job create \
--deployment=${deployment_json} \
--name=${deployment_job_name} \
--output=json \
--scoring=${scoring_json} \
--space-id==${space_id}