Version:

Models + Analytics

A model is a mathematical or programmatical representation of a real-world process. AAW currently supports three types of models: TensorFlow, RAPIDS, and Blackbox.

  • Point to + Add New Model then click either New TensorFlow or Import Blackbox to begin the model setup/import process.
  • Type into Filter to filter down the models
  • Click show to display any archived models.
  • Click refresh to refresh the table
  • Click export to export the table's values as JSON or CSV
  • Click an existing model to display additional actions. Actions vary depending on the type of model:
    • NEURAL_NET models:
      • Click View Selection to open the Model Details page
      • Click Fit Featureset to begin fitting the featureset to the dataset
      • Click Train to start training the model
      • Click Terminate Training to stop training the model
      • Click Deploy to deploy the model. Provide a Name and optional Description, then select the Mode and # of Replicas to use for the deployment. If Continuous or Batch , select a Source Table and Output Table. Click Deploy.
      • Click Export Entity to export the model as a JSON object
      • Click Archive to archive the model; it will be removed from the list of models and will be no longer useable
      • Click Description / Configuration to review summary information for the model
    • BLACKBOX models:
      • Click View Selection to open the Model Details page
      • Click Deploy to deploy the model
      • Click Export Entity to export the model as a JSON object
      • Click Archive to archive the model; it will be removed from the list of models and will be no longer useable
      • Click Description / Configuration to review summary information for the model
../_images/aaw_ui_models.png

Details

The Model Details page provides a detailed look at a given model, including configuration information, featureset, training and test datasets, and any deployments. Available actions depend on the type of model being detailed.

  • NEURAL_NET models:
    • Click Back to return to the Models page
    • Click Fit Featureset to begin fitting the featureset to the dataset
    • Click Train to start training the model
    • Click Terminate Training to stop training the model
    • Click Deploy to deploy the model. Provide a Name and optional Description, then select the Mode and # of Replicas to use for the deployment. If Continuous or Batch , select a Source Table and Output Table. Click Deploy.
    • Click Export Entity to export the model as a JSON object
    • Click Archive to archive the model; it will be removed from the list of models and will be no longer useable
    • If a lambda function is present in the Features table, click Lambda to view the lambda function code
    • If the model has been deployed previously, click the deployments' name in the Model Deployments table to open the Deployment Details
  • BLACKBOX models:
    • Click Back to return to the Models page
    • Click Deploy to deploy the model
    • Click Export Entity to export the model as a JSON object
    • Click Archive to archive the model; it will be removed from the list of models and will be no longer useable
    • If the model has been deployed previously, click the deployments' name in the Model Deployments table to open the Deployment Details
../_images/aaw_ui_model_details.png

New TensorFlow Model

To create a new TensorFlow model:

  1. Provide a Model Name.
  2. Optionally, provide a Model Description.
  3. Optionally, for Model Template, click Search to search for and select a template. Several Training Parameters will be automatically created.
  4. For Featureset, click Search to search for and select an existing featureset. In the Search window, click + Add New Featureset to add a new featureset and use it. See New Featureset for more details.
  5. For Training Dataset, click Search to search for and select an existing dataset. In the Search window, click + Add New Dataset to add a new dataset and use it. See New Dataset for more details.
  6. For Testing Dataset, click Search to search for and select an existing dataset. In the Search window, click + Add New Dataset to add a new dataset and use it. See New Dataset for more details.
  7. For Training Parameters, click one of the following:
    • + Boolean creates a boolean parameter. Provide a Parameter Name and click the slider as necessary. Slide it to the right for true; left for false.
    • + Text creates a text parameter. Provide a Parameter Name and Parameter Value.
    • + Number creates a number parameter. Provide a Parameter Name and Parameter Value.
    • + JSON creates a JSON parameter. Provide a Parameter Name and click Edit JSON to provide a JSON Parameter Value.
    • Click the trashcan icon to remove a parameter.
  8. Click Create.

New RAPIDS Model

To create a new RAPIDS model:

  1. Provide a Model Name.
  2. Optionally, provide a Model Description.
  3. Optionally, for Model Template, click Search to search for and select a template.
  4. For Dataset, click Search to search for and select an existing dataset. In the Search window, click + Add New Dataset to add a new dataset and use it. See New Dataset for more details.
  5. For Dataset Features select one or more features from the selected dataset.
  6. For Dataset Label, select a label from the list of features in the selected dataset.
  7. Set the Training Percentage for the label to the desired amount.
  8. Click Create.

Import Blackbox Model

To import a Blackbox model:

  1. Provide a Model Name.
  2. Optionally, provide a Model Description.
  3. Provide a Blackbox Docker Container repository URI or opt create a Blackbox model container.
    • If a Blackbox model has already been created using the Kinetica Blackbox SDK and published to a Docker repository, provide the URI in the following format: <repo-name>/<image-name>:<tag>, e.g., kinetica/kinetica-blackbox-quickstart:7.0.1
    • If opting to create a Blackbox model container automatically using AAW, click + New Container.
      1. Provide the Docker Repository Name.
      2. Optionally, provide the Docker Repository Description.
      3. Optionally, upload a Requirements File.
      4. Upload the Module File (must be a Python file).
      5. Provide the Module Function name from the Module File.
      6. Select a Docker credential.
      7. Click Create.
  4. Provide a Module and Function from the model
  5. For Input Columns:
    1. Click Add Input Column to create input columns.
    2. Provide a Column name and Type.
  6. For Output Columns:
    1. Click Add Output Column to create output columns.
    2. Provide a Column name and Type.
  7. Click Create.