TF Serving Example

This project enables the execution of various artificial intelligence models and stores the prediction results in a database.

Prediction results are saved based on the logged-in user and the model used. The database schema is illustrated below:

How to Run the Project

1. Build Spring Boot

You need to compile Spring Boot without database test.

cd backend
mvn clean install -DskipTests

2. Build and Start the Project

Navigate to the project directory and execute the following command:

docker-compose up --build

If you don't use ARM processor you should use docker-compose-amd.yml file.

3. Load the Database Backup

Step 1: Open a new terminal and connect to the PostgreSQL container. Then, create a backup folder:

docker exec -it postgres bash
mkdir /backup

Step 2: In another terminal, copy the backup file to the PostgreSQL container:

docker cp database/backup/backup.sql postgres:/backup/backup.sql

Step 3: Switch back to the PostgreSQL terminal and restore the backup:

psql -U postgres -d tf_serving -f /backup/backup.sql

Screenshots