rajatkhatri7 / handwritten-digit-predictor

This is visual representation of Neural Network classification of hand written digits.

Handwritten-Digit-predictor

This is visual representation of Neural Network classification of hand written digits.

Introduction

Neural networks are core of deeplearning.Unlike machine learniing where we have to provide structured data to the model ,this is not the case in neural networks . There are layers of nodes called Neurons which takes the input , no need of preprocessing, and gives their output to the neuron of other layer so each layer outputs the more and more accurate output so leads to better prediction.

---

Streamlit is a python library used to deploy machine learning on web app easily. Visit https://www.streamlit.io/ to explore it. Its very useful and easy to use.

Dataset

To train the neural network I'm using famous MNIST hand written digits dataset. Dataset consist of 70000 hand written digits from which 60k images are used to train the model and rest are for testing purpose .

How to run this demo

pip install --upgrade streamlit opencv-python
streamlit run https://raw.githubusercontent.com/Rajatkhatri7/Handwritten-Digit-predictor/master/app.py

Contribute

To contribute to this project and make it more advance feel free to mail me at [email protected]