Importing and Deploying your First BYOM Model¶
This tutorial covers basic tasks that are related to BYOM Model training, evaluation, approval and deployment within multiple environments.
In this tutorial, you will complete the following tasks:
Define a PMML Model¶
Property | Value |
---|---|
Name | PMML BYOM |
Description | My First PMML BYOM Model |
Format | PMML |
Link to BYOM | Optionally allows you to specify the external link to BYOM version. |
To define a BYOM of type PMML:
From List of Projects, open the Demo project.
The List of Models displays for the Demo project.
Click the Define BYOM Model button.
The Define BYOM Model dialog displays.
In the Basic tab, set the properties:
Property
Value
Name
PMML BYOM
Description
My First PMML BYOM Model
Format
PMML
Link to BYOM
Optionally allows you to specify the external link to BYOM version.
In the Advanced tab, set the properties:
Property
Value
Configuration
An optional property that allows adding further details about the BYOM model.
Automation
An optional property that allows adding automation configuration about the BYOM model.
Click Save to save the defined BYOM model.
Note: Clicking Save & Import Versions will move you directly towards the Import Versions dialog. For details, see Import a Model Version.
The BYOM model adds to the List of Models.
See Also
Import a BYOM Model Version¶
To import a model version for the defined BYOM model:
From the List of Models, select PMML BYOM and click the View Model Versions icon.
The PMML BYOM model versions page displays.
Click the Import Model button. The Import Model Version dialog displays.
Set the properties:
Property
Value
External ID
12345
Link to BYOM Model Version
Format
PMML
Tags
PMML, BYOM
Enable Model Evaluation
Checked
Upload Files
model.pmml, evaluation.py, requirements.txt
Note: Deselecting the Enable Model Evaluation option does not allow the model version evaluation. For details, see Model Evaluation in AnalyticOps User Guide.
Click Import. The import process progress displays in the dialog and the imported model displays in the model versions list.
See Also
Evaluate BYOM Model Version¶
To initiate the evaluation of an imported PMML model version, two methods can be followed.
First Method
From the PMML BYOM model versions, select the Imported version of the model. The Actions icon displays on the top.
Click the Actions icon to see the menu. Select Evaluate.
Second Method
From the PMML BYOM model versions, select the View Events icon for the Imported version of the model. The Model version lifecycle page displays.
Click the Evaluate button. The Evaluate Model Version dialog displays.
In the Basic tab, set the properties:
Property
Value
Dataset Template
PIMA Diabetes
Dataset
PIMA Diabetes Evaluate
Connection
Vantage Demo
In the Advanced tab, set the properties:
Property
Value
Engine
Docker Image
Resource Template
S Standard
Click Evaluate. The model version evaluation progress displays on the screen as below.
Note: For more information about Evaluation job progress, see Model Lifecycle in AnalyticOps User Guide.
Click Done when the evaluation progress completes. The model version status is changed to Evaluated.
The Evaluation Details display on the model version lifecycle page.
See Also
Approve BYOM Model Version¶
To initiate the approval of an evaluated PMML BYOM model version, two methods can be followed.
First Method
From the PMML BYOM model versions, select the Evaluated version of the model. The Actions icon displays on the top.
Click the Actions icon to see the menu. Select Approve.
Comment: Screen for Approval flow from Model versions list will display here. The feature is not available
Second Method
From the PMML BYOM model versions, select the View Events icon for the Evaluated version of the model. The Model version lifecycle page displays.
Click the Approve button. The Approve Model Version dialog displays.
Insert Comments and click Approve. The model version status is changed to Approved.
The Approval Details display on the model version lifecycle page.
Note: For more information about Approval and Rejection of model version, see Model Lifecycle in AnalyticOps User Guide.
See Also
Deploy BYOM Model Version as Batch¶
To deploy an approved PMML BYOM model version, two methods can be followed.
First Method
From the PMML BYOM model versions, select the Approved version of the model. The Actions icon displays on the top.
Click the Actions icon to see the menu. Select Deploy.
Second Method
From the PMML BYOM model versions, select the View Events icon for the Deployed version of the model. The Model version lifecycle page displays.
Click the Deploy button. The Deploy Model Version dialog displays.
In the Basic tab, set the properties:
Property
Value
Engine Type
Batch
Dataset Template
PIMA Diabetes
Connextion
Vantage Demo
Schedule
Hourly
Hours
1
Mins
0
In the Advanced tab, set the properties:
Property
Value
Engine
Docker Image
Resource Template
S Standard
Click Deploy. The model version deployment progress displays on the screen as below.
Note: For more information about Deployment job progress, see Model Lifecycle in AnalyticOps User Guide.
Click Done when the deployment progress completes. The model version status remains as Deployed.
The Deployment Details display on the model version lifecycle page.
See Also
Deploy BYOM Model Version in Vantage¶
To deploy an approved or deployed PMML BYOM model version, two methods can be followed.
First Method
From the PMML BYOM model versions, select the Deployed version of the model. The Actions icon displays on the top.
Click the Actions icon to see the menu. Select Deploy.
Second Method
From the PMML BYOM Model model versions, select the View Events icon for the Deployed version of the model. The Model version lifecycle page displays.
Click the Deploy button. The Deploy Model Version dialog displays.
In the Basic tab, set the properties:
Property
Value
Engine Type
In-Vantage
Connection
Vantage Demo
Database
AOA_DEMO
In the Advanced tab, set the properties:
Property
Value
Engine
Docker Image
Resource Template
S Standard
Click Deploy. The model version deployment progress displays on the screen as below.
Note: For more information about Deployment job progress, see Model Lifecycle in AnalyticOps User Guide.
Click Done when the deployment progress completes.
The model version status remains as Deployed.The Deployment Details display on the model version lifecycle page.
See Also
Retire All Deployments¶
To retire the active deployments for PMML BYOM model version, two methods can be followed.
First Method
From the PMML Diabetes Prediction model versions, select the Deployed version of the model. The Actions icon displays on the top.
Click the Actions icon to see the menu. Select Retire.
Second Method
From the PMML BYOM Model model versions, select the View Events icon for the Deployed version of the model. The Model version lifecycle page displays.
Click the Retire button. The Retire Model Version dialog displays.
Retire Batch Deployment¶
Select the Batch Deployment ID in the Retire Model Version dialog.
Click Retire. The model version retirement progress displays on the screen as below.
Comment: Screenshot needs to update as Retiring a BYOM is not working at the moment: https://github.com/ThinkBigAnalytics/AoaCoreService/issues/574
Note: For more information about Retiring job logs, see “Retire Job Progress” in AnalyticOps User Guide.
Click Done when the retirement progress completes. The model version status is Deployed as there are still active deployments running for the model version.
Comment: Screenshot needs to update as Retiring a BYOM is not working at the moment: https://github.com/ThinkBigAnalytics/AoaCoreService/issues/574
The Retirement Details display on the model version lifecycle page.
Comment: Screenshot needs to update as Retiring a BYOM is not working at the moment: https://github.com/ThinkBigAnalytics/AoaCoreService/issues/574
Retire Vantage Deployment¶
Select the Vantage Deployment ID in the Retire Model Version dialog.
Comment: Screenshot needs to update as Retiring a BYOM is not working at the moment: https://github.com/ThinkBigAnalytics/AoaCoreService/issues/574
Click Retire.
The model version retirement progress displays on the screen as below.Note: For more information about Retiring job progress, see Model Lifecycle in AnalyticOps User Guide.
Click Done when the retirement progress completes. The model version status changes to Retired since there is no active deployment for the version.
The Retirement Details display on the model version lifecycle page.
See Also