Hey team, I have now deployed my kedro-azureml-mlf...
# questions
g
Hey team, I have now deployed my kedro-azureml-mlflow project. This project creates a machine learning regression model that given 21 numerical inputs, predicts 1 numerical output. The 1st pipeline includes the data-processing, data-science (training and testing) steps. I want now to create a 2nd pipeline for the real-time inference pipeline with kedro and test it with the using the azure ml endpoint test gui (see the attached). This pipeline I want to include pre-processing steps and the predictions with the .pck versioned file after training of the model. Is there any tutorial/benchmark I can check to learn how to do that?
y
Have you tried this tutorial? https://github.com/Galileo-Galilei/kedro-mlflow-tutorial It's not azure specific but it explains how to serve an inference pipeline (including preprocessing and post processing) to predict in real time