reiinakano / tfjs-autoencoder

A denoising autoencoder written in TensorFlow.js

This repository contains a demo written with TensorFlow.js that shows a neural network removing noise from handwritten digits. A denoising autoencoder written in Keras is trained to remove noise from MNIST digits. The Keras model is then ported to TensorFlow.js.

You can run it immediately in your browser by going to https://reiinakano.github.io/tfjs-autoencoder/.

To run it locally, you must install Yarn and run the following command to get all the dependencies.

yarn prep

Then, you can run

yarn start

You can then browse to localhost:9966 to view the application.

This demo was written for the book Deep Learning in the Browser. You can check out the book's main repository here.