🫀 Personal Key Indicator of Heart Disease
Application

End-to-End · Continuous Machine Learning · Github CI/CD

     

     

---

Application Working

Personal Key Indicator of Heart Disease Application is an End-to-End Continuous Machine Learning Project. The backbone of this project is Dvc, Mlflow, GitHub CI/CD, and CML. Dvc stores the dataset and trained model in Azure blob, and Mlflow tracks all the training, logs artifacts, and metrics, and versionize the updated model in the model registry. Whenever a pull request comes into the experiment branch, the GitHub action will pull the dataset from the Azure blob and retrain the model, then posts the CML report as a pr comment. If the pull request merges into the experiment branch, the GitHub action will pull the dataset, mlflow track the training, and Dvc pushes the latest model to the Azure blob. At the same time, GitHub Bot pushes the latest model meta commits to the GitHub and sends CML reports as the latest push comment. Whenever a code change occurs in the experiment branch, Streamlit Application fetches and updates the app and the model.

demo.mp4

Github Bot pull request comment after running Github Action on a pull request to the Experiment branch

After merging the pull request