Add Model Assets

Overview

ModelOp Center is designed to be agnostic to the type of model, the platform on which the model runs, and the data platforms with which the model interacts during various portions of the model lifecycle.

Types of Model Assets that may be used during the Model Life Cycle Include:

  • Trained model artifacts

  • Requirements (Dependencies) lists

  • Data assets

  • Schemas

 

How to add Assets

During Model Import

Please see Register a Model to get details about adding assets during model import.

Via ModelOp UI

Upload a local File

  1. Go to the Model Details page by clicking on the model from the Models page

    Go to the Model Details Page
  2. Go to the Assets page by clicking on Assets

    Go to Assets Page
  3. Click on Upload File which should open a dialog box to select a local file.

  4. The file to be uploaded will show up on the page. Depending on the file size, you will see 2 options: Embed and Upload. The Embed option stores the file in the ModelOp Center database while the Upload option will store the file in ModelOp Center controlled S3 bucket or HDFS store depending on the model type. Click the option best suited for the use case.

  5. Once the Asset has been added, select the appropriate asset role. Optionally, add tags to the asset if required by the Model governance process.

Upload a File from a Remote Location

This option, at the moment, only supports copying assets from the remote S3 bucket. The option is set by the organization during the install. To confirm if the file will be copied to the ModelOp Center controlled bucket, please see the step 1 through 4.

  1. Go to the Model Details page by clicking on the model from the Models page

  2. Go to the Assets page by clicking on Assets

  3. Click on the Add By URL button

  4. Enter the URL for the asset in the URL field of the dialog box. You will see a note mentioning if the file will be copied to the ModelOp Center controller S3 bucket

  5. Select the type to be S3 (which should be by default selected based on the URL)

  6. Select the region for the S3 location and click Add Asset button

  7. Once the Asset has been added, select the appropriate asset role. Optionally, add tags to the asset if required by the Model governance process.

Steps to automatically copy from an existing s3 location to a MOC-controlled location

Provide a File by Reference

  1. Go to the Model Details page by clicking on the model from the Models page

  2. Go to the Assets page by clicking on Assets

  3. Click on the Add By URL button

  4. Enter the URL for the asset in the URL field of the dialog box.

  5. Select the type to be S3 or HDFS depending on where the asset is

  6. This step is only required if you select type to be S3 on previous step. Select the region for the S3 location and click Add Asset button

  7. Once the Asset has been added, select the appropriate asset role. Optionally, add tags to the asset if required by the Model governance process.

Provide a SQL Asset

Please see Integrate with SQL Databases to understand how ModelOp Center integrates with SQL databases. Once you get familiar with how ModelOp Center handles SQL Assets, follow these steps to add SQL assets to a given model.

Via CLI

Please see the asset command for adding assets using the CLI.

Data Assets

ModelOp Center “Assets” that support various data technologies:

  • S3 Files

  • HDFS Files

  • SQL Asset

 

Summary of example activities where a User or an MLC may use a Data Asset:

Action

Data Required

Available via:

MOC Asset Examples [ASSET_ROLE : EXAMPLE}

Add an Attachment (to a StoredModel)

All types

CLI: s3 & embedded assets only

UI: all

API: all

Applicable to S3, HDFS, and SQL Asset

Run a Metrics Job

(e.g. “Back Test” using labeled data)

Evaluation Data

CLI, UI, API, MLC

S3:

Test_Data: s3://<model_base>/EvaluationData.csv

HDFS:

Test_Data: hdfs://<model_base>/EvaluationData.csv

SQL Asset:

Test_Data: SELECT * FROM <read_only_Eval_table> WHERE <conditions>

Run a Data Drift Job

•Training/Baseline Data

•Comparator Data

CLI: would trigger the API

API, MLC

S3:

Training_Data: s3://<model_base>/TrainingData.csv

Compare_Data: s3://<model_base>/ComparatorData.csv

HDFS:

Training_Data: hdfs://<model_base>/TrainingData.csv

Compare_Data: hdfs://<model_base>/ComparatorData.csv

SQL:

Training_Data: SELECT * FROM <read_only_Training_table> WHERE <conditions>

Compare_Data: Training: SELECT * FROM <read_only_Comparator_table> WHERE <conditions> … may need to have tag to specify input vs. output comparator data

Run a Model Concept Drift Job

•Training/Baseline Data

•Comparator Data

CLI: would trigger the API

API, MLC

S3:

Training_Data: s3://<model_base>/TrainingData.csv

Compare_Data: s3://<model_base>/ComparatorData.csv

HDFS:

Training_Data: hdfs://<model_base>/TrainingData.csv

Compare_Data: hdfs://<model_base>/ComparatorData.csv

SQL:

Training_Data: SELECT * FROM <read_only_Training_table> WHERE <conditions>

Compare_Data: Training: SELECT * FROM <read_only_Comparator_table> WHERE <conditions> … may need to have tag to specify input vs. output comparator data

Run a Bias Detection Job

Evaluation Data

CLI, UI, API, MLC

S3:

Test_Data: s3://<model_base>/EvaluationData.csv

HDFS:

Test_Data: hdfs://<model_base>/EvaluationData.csv

SQL Asset:

Test_Data: SELECT * FROM <read_only_Eval_table> WHERE <conditions>

Run a Training Job

•Training Data

MLC, UI, API

S3:

Training_Data: s3://<model_base>/TrainingData.csv

HDFS:

Training_Data: hdfs://<model_base>/TrainingData.csv

SQL Asset:

Training_Data: SELECT * FROM <read_only_Training_table> WHERE <conditions>

 

Other Assets

Trained Model Artifacts

ModelOp Center supports Trained Model Artifacts stored in S3 bucket or HDFS (does not support as SQL Asset). When adding asset that is used as Trained Model Artifact, please select the asset role to be “Weights File” or “Model Binary File” depending on the use case.

 

Schemas

ModelOp Center runtime supports data schemas which helps to enforce specif type of on different fields. The schema can be used for input data and/or output data. When uploading a schema, select “Model Schema” asset role.

 

Requirements

ModelOp runtime comes with basic pre-installed libraries for different model and languages. If the model requires additional libraries to be installed by the ModelOp Runtime, it supports defining model library requirements in requirements.txt file. The ModelOp runtime will check the requirements file before running the model and install the missing libraries. When uploading the requirements file, please select the asset role of “Requirements”.

 

 

 

Next Article: Model Life Cycle Management: Overview >