Skip to content

clementw168/mixture-density-net

Repository files navigation

A reimplementation of Mixture Density Networks

This repository reimplements the mixture density network (MDN) model proposed by Bishop (1994) in PyTorch. The MDN model is a neural network that can predict multiple possible outputs for a given input. Contrary to traditional neural networks for regression trained with mean squared error, the MDN model is trained with maximum likelihood estimation with a Gaussian mixture distribution.

The project evaluates the capability of the MDN model to predict multiple possible outputs for a given input. Three datasets were considered :

  • a toy ill-posed sinusoidal inverse problem,
  • robot kinematics,
  • MNIST.

Our report is available here.

You can also checkout our poster.

poster

About

An implementation of Mixture density networks in Pytorch with some custom applications

Resources

Stars

Watchers

Forks

Releases

Packages

Used by

Contributors

Languages