End-to-End ML Project

Student Exam Performance

This repo includes an end-to-end ML project. The focus is mainly on deployment rather than the difficulty of the dataset.

The dataset is Student exam performance from Kaggles and the prediction type is regression.

This repo was cloned from my repo ML-end-to-end-azure and only the deployment part was changed from Microsoft Azure to AWS.

Run locally

• From the project directory run:

  python app.py

• On your browser navigate to http://127.0.0.1:8080/predictdata

You should see this:

Fig.1

Run docker

• Build:

  • docker build -t mle .

• Check the built exists:

  • docker images

• Run:

  • docker run -d --name <CONTAINER_NAME> -p 8080:8080 mle

• navigate to http://127.0.0.1:8080/predictdata

You should see a page shown in Fig.1

• Stop the container

  • docker stop mycontainer_mle

Deployment on AWS EC2

This model was successfully deployed on AWS EC2. To see the deployment, go to Github Actions.

To replecitae the deployment, refer to instructions