Skip to content

MichaelMathieu/factors-variation

Repository files navigation

factors-variation

Code for the paper Disentangling factors of variation in deep representations using adversarial training by Mathieu, Zhao, Ramesh, Sprechmann and LeCun, NIPS 2016.

This code is a cleaned up version of our research code, and does not cover everything that is in the paper (yet). In particular, the code for the NORB dataset not there yet.

Requirements

This code runs using torch. It is only tested on GPU. The following torch libraries are required: nn, cunn, cudnn, optim, image, nngraph.

The following libraries are optional (but strongly recommended: nnx, display.

We are using preprocessed versions of the dataset. You can download the mnist files here . You need to create a folder data/mnist containing these files.

Usage

We provide two model architectures: small and big. By default, the big architecture is used, it should work in all cases. The small architecture is faster to train and should work for MNIST and Sprites.

MNIST

th train.lua --optimDisc adam

or

th train.lua --modelSize small --optimDisc adam

Sprites

th train.lua --dataset sprites --nFeatures 14 --optimDisc adam

yaleB

th train.lua --dataset yaleBExtended --swap2weight 1 --learningRateGen 0.0002 --learningRateDisc 0.0002

About

Code of the paper `Disentangling factors of variation in deep representations using adversarial training` by Mathieu, Zaho, Ramesh, Sprechmann and LeCun

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages