Note: This is a repo that can be shared to our customers. This means it's NOT OK to include Microsoft confidential content. All discussions should be appropriate for a public audience.
MLOps Model Factory is a platform and an end to end workflow that supports generating multiple models and used for deployment to any target.
- Supports generation of multiple ML Models through a single platform and repo
- MLOps pipeline for Data preparation, transformation, Model Training, evaluation, scoring and registration
- Based on Azure ML SDK v2 1.4
- Option to package ML Models in Docker Images
The idea of this platform and end to end workflow is to provide a minimum number of scripts to implement an environment to train and test multiple ML Models using Azure ML SDK v2 and Azure DevOps.
The workflow contains the following folders/files:
-
devops: the folder contains Azure DevOps related files (yaml files to define Builds).
-
docs: documentation.
-
src: source code that is not related to Azure ML directly. This is typically data science related code.
-
mlops: scripts that are related to Azure ML.
-
mlops/nyc-taxi: a fake pipeline with some basic code to build a model
-
mlops/london-taxi: a fake pipeline with some basic code to build another model
-
test: a folder with dummy test to write unit tests for the build
-
model: Model related files and dependencies
-
.amlignore: using this file we are removing all the folders and files that are not supposed to be in Azure ML compute.
The workflow contains the following documents:
- docs/how_to_setup.md: explain how to configure the workflow.
Information about how to setup the repo is in the following document.
Developers and Data scientists can use the execute-command in the notebooks to try out the commands in the AML compute from their local machine.