Abstract: ML deployment is not the end — it is where ML models start to materialize impact and value to the business and people. We need monitoring and testing to ensure ML models behave as expected.
What you’ll Learn: A set of tests and metrics that ML practitioners should care about post ML deployment.
Important Note:
- All tests chosen are non-parametric, which means they don't assume normality in data distributions
- For non-Databricks users, you will need to configure your own MLflow tracking server to view the MLflow UI
Files in this repository:
- Slides in PDF
- Notebooks in both .dbc, .html, and ipynb formats
- .dbc format is available to be uploaded to the Databricks environment
- .html format can be open in any browser once downloaded, without needing access to Databricks
- Download and uncompress the .zip file and you will see the individual notebooks in .html extension
- The mlops2021 folder contains the individual notebooks in .ipynb format
For the code to work, you need to create a directory named mlops2021 and place all the individual notebooks underneath, shown in the screenshot:
