MNIST-CNN

Handwritten Digit Recognition using a Convolutional Neural Network.

This repository contains a CNN implemented in TensorFlow.js.

Features

• Train the CNN using custom parameters.
• Visualise Training and Loss curves.
• Test the model using custom digits drawn by user on a canvas
• Visualise layer activations.
• Play around with a 3D model of the CNN.

Dependencies

These dependencies are required to be installed.

• Node.js version 8.9 or higher
• Yarn

How to build an example

cd into the directory

If you are using yarn:

cd MNIST
yarn
yarn watch

Details

The package contains two scripts:

yarn watch: starts a local development HTTP server which watches the filesystem for changes so you can edit the code (JS or HTML) and see changes when you refresh the page immediately.

yarn build: generates a dist/ folder which contains the build artifacts and can be used for deployment.

Contributors

• Vipul Bansal
• Yash Vardhan Sharma
• Aaryan Garg

References

• tfjs-examples