Skip to content

ipazc/lstm_autoencoder

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

5 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

lstm_autoencoder

A time serie can be described with an LSTM Autoencoder. Usually, LSTM's are required to have fixed timesteps in order for the decoder part of the autoencoder to know beforehand how many timesteps should produce. However, this version of LSTM Autoencoder allows to describe timeseries based on random samples with unfixed timesteps.

In this LSTM autoencoder version, the decoder part is capable of producing, from an encoded version, as many timesteps as desired, serving the purposes of also predicting future steps.

Installation

It is required keras, tensorflow under the hood, pandas for the example and pyfolder for save/load of the trained model. They can be installed with pip:

pip3 install -r requirements.txt

Tensorflow is not included in the requirements.txt, so it must be manually installed:

pip3 install tensorflow

Usage

There is an example in a example.py

Releases

No releases published

Packages

No packages published

Languages