ml model create
Important: IBM Cloud Pak® for Data
Version 4.7 will reach end of support (EOS) on 31 July, 2025. For more information, see the Discontinuance of service announcement for IBM Cloud Pak for Data Version 4.X.
Upgrade to IBM Software Hub Version 5.1 before IBM Cloud Pak for Data Version 4.7 reaches end of support. For more information, see Upgrading IBM Software Hub in the IBM Software Hub Version 5.1 documentation.
Create a model with the defined payload. A model represents a machine learning model asset.
Syntax
cpd-cli ml model create \
[--async] \
[--content-location=<attachment-details>] \
[--context=<catalog-project-or-space-id>] \
[--cpd-config=<cpd-config-location>] \
[--cpd-scope=<cpd-scope>] \
[--custom=<map<key,value>>] \
[--description=<resource-description>] \
[--domain=<model-domain-name>] \
[--hyper-parameters=<model-training-hyper-parameters>] \
[--jmes-query=<jmespath-query>] \
[--label-column=<label-column>] \
[--metrics=<operation-metrics>] \
[--model-definition=<model-definition>] \
--name=<resource-name> \
[--output=json|yaml|table] \
[--output-file=<output-file-location>] \
[--pipeline=<resource-reference>] \
--profile=<cpd-configuration-profile-name> \
[--project-id=<cpd-project-id>] \
[--quiet] \
[--raw-output=true|false] \
[--schemas=<expected-data-schemas>] \
[--size=<model-size>] \
--software-spec=<software-specification> \
[--space-id=<space-identifier>] \
[--tags=<tag1,tag2,...>] \
[--test-data-references=<holdout-or-test-data-set>] \
[--training-data-references=<model-training-data>] \
[--transformed-label-column=<transformed-label-column-name>] \
--type=<model-type> \
[--user-defined-objects=<objects-defined-by-model>] \
[--verbose]Arguments
The ml model create command has no
arguments.
Options
| Option | Description |
|---|---|
--async |
Run the command asynchronously. By
default, processing finishes before the command runs.
|
--content-location |
Specify details about the
attachments that are uploaded with the model.
|
--context |
Specify the configuration context
name.
|
--cpd-config |
The Cloud Pak for Data
configuration location. For example,
$HOME/.cpd-cli/config.
|
--cpd-scope |
The Cloud Pak for Data space,
project, or catalog scope. For example,
cpd://default-context/spaces/7bccdda4-9752-4f37-868e-891de6c48135.
|
--custom |
Specify user-defined properties as
key-value pairs.
|
--description |
Specify a
resource
description.
|
--domain |
Provide a domain name for the
model. For example, sentiment, entity, visual-recognition, finance, retail, real
estate.
|
|
|
Display command
help.
|
--hyper-parameters |
Specify the hyper parameters that
are used for training the model.
|
--jmes-query |
Provide a JMESPath query to
customize the output.
|
--label-column |
Specify the label
column.
|
--metrics |
Specify metrics that can be
returned by an operation.
|
--model-definition |
Specify the model definition. The
'software_spec' is a reference to a software specification. The 'hardware_spec' is a reference to a
hardware specification.
|
--name |
Specify a
resource name.
|
--output |
Specify an output format.
Valid formats include json, yaml, or text (the default
format).
|
--output-file |
Specify a file path where all
output is redirected.
|
--pipeline |
Specify a reference to a resource.
The 'hardware_spec' is a reference to the hardware specification. The
'hybrid_pipeline_hardware_specs' reference is used only when training a hybrid pipeline. A hybrid
pipeline is used to specify the compute requirements for each pipeline
node.
|
--profile |
The name of the profile that you
created to store information about an instance of Cloud Pak for Data and your credentials for the
instance.
|
--project-id |
Specify a Cloud Pak
for Data project instance that contains the resource.
|
--quiet |
Suppress verbose
messages.
|
--raw-output |
When set to true, single values
in
JSON output mode are not surrounded by quotes.
|
--schemas |
Specify the expected data
schemas. Schemas that are defined here take precedence over any schemas that are provided in the
data references.
|
--size |
Specify the model size. This
setting is used by scoring to record the model size.
|
--software-spec |
Specify a software specification
(SoftwareSpecRel object).
|
--space-id |
Specify a space
identifier.
|
--tags |
Specify tags that can be used
when searching for resources.
|
--test-data-references |
Specify the holdout or test data
sets.
|
--training-data-references |
Specify the training data that was
used to create the model.
|
--transformed-label-column |
Specify the name of the label
column (as seen by the estimator), which might be transformed by the previous pipeline transformers.
The name is not necessarily the same column as the 'label_column' in the initial data
set.
|
--type |
Specify the model
type. For information on supported
model types, see Supported machine learning tools, libraries, frameworks, and software
specifications.
|
--user-defined-objects |
Specify defined objects that are
referenced by the model. For any defined model class or function, the name (as referenced in the
model) must be specified as the 'key' and a fully qualified class or function name must be specified
as the 'value'. This is applicable for 'Tensorflow 2.X' models serialized in 'H5' format that uses
the 'tf.keras' API.
|
--verbose |
Logs include more detailed
messages.
|
Table 1: Command options