pt-sdae

PyTorch implementation of a version of the Stacked Denoising AutoEncoder (note this implementation is unofficial). Compatible with PyTorch 1.0.0 and Python 3.6 or 3.7 with or without CUDA.

Examples

An example using MNIST data can be found in the examples/mnist/mnist.py which achieves around 80% accuracy using k-Means on the encoded values.

Here is an example confusion matrix, true labels on y-axis and predicted labels on the x-axis.

Usage

This is distributed as a Python package ptsdae and can be installed with python setup.py install. The PyTorch nn.Module class representing the SDAE is StackedDenoisingAutoEncoder in ptsdae.sdae, while the pretrain and train functions from ptsdae.model are used to train the autoencoder.

Currently this code is used in a PyTorch implementation of DEC, see https://github.com/vlukiyanov/pt-dec.