Handwritten-Digit-Recognition

This project creates a Deep Neural Network with 4 hidden layers (1500 nodes each) to train and test against the MNIST Handwritten Digit dataset. The project also uses Convolution Neural Networks with 2 hidden layers against the MNIST dataset Lastly, this project uses Recurrent Neural Networks approach on the same MNIST data

Getting Started

This project uses the Tensorflow Library in Python and uses Deep Neural Networks and Convolution Neural Networks

Prerequisites

1. Tensorflow 1.0
2. Python 2.7
3. NVIDIA CUDA GPU (Optional)

Installing

To install Tensorflow

pip install tensorflow

Running the tests

To run the code

python deep-net.py
python cnn.py
python rnn.py

Built With

• Tensorflow - The Deep Learning Library

Authors

• Jahin Majumdar - (https://github.com/jahin07)