Skip to content

Latest commit

 

History

History
24 lines (22 loc) · 705 Bytes

README.md

File metadata and controls

24 lines (22 loc) · 705 Bytes

Handwritten Digit Recognition

This is a simple implementation of a basic CNN to recognize hand written digits.

Installation

$ git clone https://github.com/shubham1172/MNISTDigitRecoginition
$ cd MNISTDigitRecognition
$ pip3 install -r requirements.txt

Running

To train,

$ python3 train.py

To start the interactive GUI,

$ python3 play.py

Data

I have used the MNIST dataset for training the CNN. It contains a training set of 60,000 examples, and a test set of 10,000 examples of handwritten digits.

Results

screenshot
The screenshot shows a sample prediction for the digit '2'.