HOWTO - install AzureML SDK v1+v2 and ESML accelerator library
HOWTO - Supported use cases & Accelerated use cases
- Configure the lake_settings
- See section "3 Getting started: Notebooks" in: HOWTO - install AzureML SDK v1+v2 and ESML accelerator library
- Run the 3 notebooks of your choice to genereate the Azure ML Pipelines:
- Example: If you want to work with Databricks and pyspark, for batch deployment: 1 + 2b + 3b
- Example: If you want to use Azure ML compute and AutoML, for online deployment AND/OR batch deployment: 1 + 2a + 3b and/or 3c
Output:
- 2 Azure Machine Learning pipelines for: Training and Inference
- 1 Online endpoint
- Configure the inline python parameters in the file
21-train_in_2_gold_train_pipeline.py
Parameters to configure in 21-train_in_2_gold_train_pipeline.py
at line 49, 50
advanced_mode = False # ADVANCED MODE (DatabricksSteps also) + Manual ML (or AutoML if defined in Databricks notebook)
use_automl = True # SIMPLE MODE + AutoMLStep (if True)Manual ML Step (if False)
- Import Pipeline from from template - Template location
- Point the pipeline to your project and models branch, such as "project001_M11_dev_branch"
- Configure the Variables in Azure Devops / Github Actions