Skip to content

Latest commit

 

History

History
 
 

vanishing_gradients

Vanishing Gradients.

This demo helps to visualize the problem of vanishing gradients (particularly w.r.t to use of different activation functions) using TensorFlow.js and Plotly.js.

Features

  • Analyse various activation functions by plotting them.
  • Train a fully connected neural network using custom parameters.
  • Visualize the vanishing gradients in the nn-architecture printed.
  • Opacity of links between different nodes represent magnitude of positive and negative gradients separately.
  • Visualise accuracy and Loss curves.

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 vanishing_gradients
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

References