Skip to content

Testing Full ML Pipeline

Now that the core Workbench AWS Stack has been deployed. Let's test out Workbench by building a full entire AWS ML Pipeline from start to finish. The script build_ml_pipeline.py uses the Workbench API to quickly and easily build an AWS Modeling Pipeline.

Taste the Awesome

The Workbench "hello world" builds a full AWS ML Pipeline. From S3 to deployed model and endpoint. If you have any troubles at all feel free to contact us at workbench email or on Discord and we're happy to help you for FREE.

  • DataLoader(abalone.csv) --> DataSource
  • DataToFeatureSet Transform --> FeatureSet
  • FeatureSetToModel Transform --> Model
  • ModelToEndpoint Transform --> Endpoint

This script will take a LONG TiME to run, most of the time is waiting on AWS to finalize FeatureGroups, train Models or deploy Endpoints.

❯ python build_ml_pipeline.py
<lot of building ML pipeline outputs>
After the script completes you will see that it's built out an AWS ML Pipeline and testing artifacts.

Run the Workbench Dashboard (Local)

Dashboard AWS Stack

Deploying the Dashboard Stack is straight-forward and provides a robust AWS Web Server with Load Balancer, Elastic Container Service, VPC Networks, etc. (see AWS Dashboard Stack)

For testing it's nice to run the Dashboard locally, but for longterm use the Workbench Dashboard should be deployed as an AWS Stack. The deployed Stack allows everyone in the company to use, view, and interact with the AWS Machine Learning Artifacts created with Workbench.

cd workbench/application/aws_dashboard
./dashboard
This will open a browser to http://localhost:8000

workbench_new_light

Workbench Dashboard: AWS Pipelines in a Whole New Light!

Success

Congratulations: Workbench is now deployed to your AWS Account. Deploying the AWS Stack only needs to be done once. Now that this is complete your developers can simply pip install workbench and start using the API.

If you ran into any issues with this procedure please contact us via Discord or email workbench@supercowpowers.com and the SCP team will provide free setup and support for new Workbench users.