Deploying a recommender system on the cloud
How to train and deploy a recommender system on AWS with custom containers
This project covers the end-to-end process of training and deploying a recommender system on AWS, following a typical use case such as predicting movie ratings from users.
File structure
- /data: Input datasets with users, movies and ratings information.
- /img: Images for the project.
- /model: Directory where the recommender model is saved after training.
- /src: Sourcecode of the algorithm and API configuration files. More details in src/README.md
- build_and_push.sh: Script provided by AWS to automatically build and push the Docker image to Elastic Container Registry (ECR).
- Dockerfile: Text file which contains the set of commands for building the Docker image.
Article here: Medium: Movie recommendations on the cloud