W3cubDocs

/Ansible 2.9

gcp_mlengine_version – Creates a GCP Version

New in version 2.9.

Synopsis

  • Each version is a trained model deployed in the cloud, ready to handle prediction requests. A model can have multiple versions .

Requirements

The below requirements are needed on the host that executes this module.

  • python >= 2.6
  • requests >= 2.18.4
  • google-auth >= 1.3.0

Parameters

Parameter Choices/Defaults Comments
auth_kind
string / required
    Choices:
  • application
  • machineaccount
  • serviceaccount
The type of credential used.
auto_scaling
dictionary
Automatically scale the number of nodes used to serve the model in response to increases and decreases in traffic. Care should be taken to ramp up traffic according to the model's ability to scale or you will start seeing increases in latency and 429 response codes.
min_nodes
integer
The minimum number of nodes to allocate for this mode.
deployment_uri
string / required
The Cloud Storage location of the trained model used to create the version.
description
string
The description specified for the version when it was created.
env_type
string
Specifies which Ansible environment you're running this module within.
This should not be set unless you know what you're doing.
This only alters the User Agent string for any API requests.
framework
string
The machine learning framework AI Platform uses to train this version of the model.
Some valid choices include: "FRAMEWORK_UNSPECIFIED", "TENSORFLOW", "SCIKIT_LEARN", "XGBOOST"
is_default
boolean
    Choices:
  • no
  • yes
If true, this version will be used to handle prediction requests that do not specify a version.

aliases: default
labels
dictionary
One or more labels that you can add, to organize your model versions.
machine_type
string
The type of machine on which to serve the model. Currently only applies to online prediction service.
Some valid choices include: "mls1-c1-m2", "mls1-c4-m2"
manual_scaling
dictionary
Manually select the number of nodes to use for serving the model. You should generally use autoScaling with an appropriate minNodes instead, but this option is available if you want more predictable billing. Beware that latency and error rates will increase if the traffic exceeds that capability of the system to serve it based on the selected number of nodes.
nodes
integer
The number of nodes to allocate for this model. These nodes are always up, starting from the time the model is deployed.
model
dictionary / required
The model that this version belongs to.
This field represents a link to a Model resource in GCP. It can be specified in two ways. First, you can place a dictionary with key 'name' and value of your resource's name Alternatively, you can add `register: name-of-resource` to a gcp_mlengine_model task and then set this model field to "{{ name-of-resource }}"
name
string / required
The name specified for the version when it was created.
The version name must be unique within the model it is created in.
prediction_class
string
The fully qualified name (module_name.class_name) of a class that implements the Predictor interface described in this reference field. The module containing this class should be included in a package provided to the packageUris field.
project
string
The Google Cloud Platform project to use.
python_version
string
The version of Python used in prediction. If not set, the default version is '2.7'. Python '3.5' is available when runtimeVersion is set to '1.4' and above. Python '2.7' works with all supported runtime versions.
Some valid choices include: "2.7", "3.5"
runtime_version
string
The AI Platform runtime version to use for this deployment.
scopes
list
Array of scopes to be used.
service_account
string
Specifies the service account for resource access control.
service_account_contents
jsonarg
The contents of a Service Account JSON file, either in a dictionary or as a JSON string that represents it.
service_account_email
string
An optional service account email address if machineaccount is selected and the user does not wish to use the default email.
service_account_file
path
The path of a Service Account JSON file if serviceaccount is selected as type.
state
string
    Choices:
  • present
  • absent
Whether the given object should exist in GCP

Notes

Note

  • for authentication, you can set service_account_file using the c(gcp_service_account_file) env variable.
  • for authentication, you can set service_account_contents using the c(GCP_SERVICE_ACCOUNT_CONTENTS) env variable.
  • For authentication, you can set service_account_email using the GCP_SERVICE_ACCOUNT_EMAIL env variable.
  • For authentication, you can set auth_kind using the GCP_AUTH_KIND env variable.
  • For authentication, you can set scopes using the GCP_SCOPES env variable.
  • Environment variables values will only be used if the playbook values are not set.
  • The service_account_email and service_account_file options are mutually exclusive.

Examples

- name: create a model
  gcp_mlengine_model:
    name: model_version
    description: My model
    regions:
    - us-central1
    online_prediction_logging: 'true'
    online_prediction_console_logging: 'true'
    project: "{{ gcp_project }}"
    auth_kind: "{{ gcp_cred_kind }}"
    service_account_file: "{{ gcp_cred_file }}"
    state: present
  register: model

- name: create a version
  gcp_mlengine_version:
    name: "{{ resource_name | replace('-', '_') }}"
    model: "{{ model }}"
    runtime_version: 1.13
    python_version: 3.5
    is_default: 'true'
    deployment_uri: gs://ansible-cloudml-bucket/
    project: test_project
    auth_kind: serviceaccount
    service_account_file: "/tmp/auth.pem"
    state: present

Return Values

Common return values are documented here, the following are the fields unique to this module:

Key Returned Description
autoScaling
complex
success
Automatically scale the number of nodes used to serve the model in response to increases and decreases in traffic. Care should be taken to ramp up traffic according to the model's ability to scale or you will start seeing increases in latency and 429 response codes.

minNodes
integer
success
The minimum number of nodes to allocate for this mode.

createTime
string
success
The time the version was created.

deploymentUri
string
success
The Cloud Storage location of the trained model used to create the version.

description
string
success
The description specified for the version when it was created.

errorMessage
string
success
The details of a failure or cancellation.

framework
string
success
The machine learning framework AI Platform uses to train this version of the model.

isDefault
boolean
success
If true, this version will be used to handle prediction requests that do not specify a version.

labels
dictionary
success
One or more labels that you can add, to organize your model versions.

lastUseTime
string
success
The time the version was last used for prediction.

machineType
string
success
The type of machine on which to serve the model. Currently only applies to online prediction service.

manualScaling
complex
success
Manually select the number of nodes to use for serving the model. You should generally use autoScaling with an appropriate minNodes instead, but this option is available if you want more predictable billing. Beware that latency and error rates will increase if the traffic exceeds that capability of the system to serve it based on the selected number of nodes.

nodes
integer
success
The number of nodes to allocate for this model. These nodes are always up, starting from the time the model is deployed.

model
dictionary
success
The model that this version belongs to.

name
string
success
The name specified for the version when it was created.
The version name must be unique within the model it is created in.

packageUris
list
success
Cloud Storage paths (gs://…) of packages for custom prediction routines or scikit-learn pipelines with custom code.

predictionClass
string
success
The fully qualified name (module_name.class_name) of a class that implements the Predictor interface described in this reference field. The module containing this class should be included in a package provided to the packageUris field.

pythonVersion
string
success
The version of Python used in prediction. If not set, the default version is '2.7'. Python '3.5' is available when runtimeVersion is set to '1.4' and above. Python '2.7' works with all supported runtime versions.

runtimeVersion
string
success
The AI Platform runtime version to use for this deployment.

serviceAccount
string
success
Specifies the service account for resource access control.

state
string
success
The state of a version.



Status

Authors

  • Google Inc. (@googlecloudplatform)

Hint

If you notice any issues in this documentation, you can edit this document to improve it.

© 2012–2018 Michael DeHaan
© 2018–2019 Red Hat, Inc.
Licensed under the GNU General Public License version 3.
https://docs.ansible.com/ansible/2.9/modules/gcp_mlengine_version_module.html