Welcome to the Digit Recognizer project, a comprehensive implementation of a handwritten digit recognition system from scratch using Python and PyTorch. The core of this project is a Convolutional Neural Network (CNN) model trained on the MNIST dataset, a widely used benchmark for image classification tasks. Our model employs the popular ResNet architecture, which is known for its exceptional performance and scalability.
The project originated as a part of the mid-term project for the Artificial Intelligence course at the esteemed State University of Surabaya. It offers a real-world application of machine learning and web development, seamlessly blending the two domains.
-
State-of-the-Art Accuracy: Our model has been meticulously trained and tested on the MNIST dataset, achieving an impressive accuracy of 99.01% on the test set. This high accuracy demonstrates the model's ability to recognize handwritten digits with exceptional precision.
-
Web User Interface: We've gone a step further by developing a user-friendly web interface to showcase the digit recognition capabilities. This allows users to draw or input handwritten digits and witness the model's recognition accuracy in real-time.
-
Seamless Deployment: The model and web interface are seamlessly integrated using Flask, a powerful web framework for Python. You can run the web application locally on your machine, offering an interactive experience.
Follow these steps to set up and run the Digit Recognizer project on your local machine:
-
Clone the Repository: Begin by cloning this repository to your local machine.
-
Navigate to the Web App Directory: Change your working directory to the "web-app" folder within the cloned repository.
cd web-app -
Install Dependencies: To ensure all necessary dependencies are installed, run the following commands. This will take care of both frontend and backend dependencies.
npm install && npm run setup -
Run the Backend
npm run be
-
Run the Frontend
npm run fe
-
Access the Web Interface: Open your web browser and visit http://localhost:3000 to start using the Digit Recognizer.