End-to-End Engine Horsepower Prediction with Deep Learning Regression Part-III
Introduction
In the previous blogs we explored data pre-processing and model building, in this article we will explore model tracking using MLFlow.
MLflow is an open-source platform designed to manage the end-to-end machine learning lifecycle. It was developed by Databricks and has become widely used in the machine learning community. MLFlow provides tools and libraries to track experiments, package code, share models, and manage the deployment of machine learning models.
Key components of MLflow:
- Tracking
- Projects
- Models
- Model Registry
Benefits of using MLflow:
- Reproducibility
- Model Management
- Flexibility
Overall, MLFlow provides a comprehensive framework to streamline the machine learning development process, making it easier to track experiments, reproduce results, and deploy models efficiently in real-world applications.
Implementation Of MLflow:
First install the mlflow in your local system.
After installation connect it to the backend server to get the url.
mlflow.set_tracking_uri("sqlite:///mlflow.db")
mlflow.set_experiment("fuel efficiency")
Log the parameters in the model with mlflow for tracking:
Change the parameters and observe the changes in the mlflow page.
We can get the above page for every run after changing the parameters.
After some runs of the model, we can easily track the parameters which are need to be changed for the improvement of the model.
In the next article we will explore model deployment using Flask.
References:
https://www.kaggle.com/code/prince381/predicting-the-miles-per-gallon-mpg-of-cars
https://mlflow.org/docs/latest/tracking.html
By Rayapureddi Subhash