An ML classifier web app to classify bank notes whether they are fake or not.
Dataset Link: https://archive.ics.uci.edu/ml/datasets/banknote+authentication
The model is based on the following predictor variables:
- Variance of the image
- Skewness of the image
- Curtosis of the image
- Entropy of the image
The following dependencies are required to be installed before running the application:
- Docker
- Pipenv
Step 1: Make sure the prerequisites are installed
Step 2: Clone the repo and initialize the pipenv shell environment
pipenv shell
Step 3: Install the required dependencies from the Pipfile.lock
pipenv install
Step 4: Run the postgres database docker and start the uvicorn server
make run-db start-server
Step 5: Start the streamlit server in a separate terminal
make run-streamlit
demo.1.mp4
There are a whole bunch of things that can be added as feature to the project, I welcome all to raise issues or pull requests but please understand the project was made as a personal portfolio project and not a full-fledged one.