Retrain a model programmatically with python

Question

I have created a Scikit-learn model with python and I have ran the experiment. Now , I need to retrain the model with new data through python but its mentioned in the documentation that we need to create a batch interference pipeline to retrain a model. If I create a batch interference pipeline, will I be able to deploy it? Any help will be appreciated. Thanks.

Answer 1

Batch Inference Pipeline is useful for making predictions on large data and you can publish the pipeline to your workspace. To retrain your model programmatically, you can simply run your experiment again, then save and register the model to use at the deployment stage. Please review Azure ML Pipelines for more details on how to optimize your workflow.