theislab/trVAE

Conditional out-of-distribution prediction

trVAE

*Conditional out-of-distribution generation for unpaired data using transfer VAE (Bioinformatics, 2020).

Note: We have upgraded trVAE to a faster and more efficient implementation. Please refer to Here

Introduction

A Keras (tensorflow < 2.0) implementation of trVAE (transfer Variational Autoencoder) .

trVAE can be used for style transfer in images, predicting perturbations responses and batch-removal for single-cell RNA-seq.

• For pytorch implementation check Here

Getting Started

Installation

Before installing trVAE package, we suggest you to create a new Python 3.6 (or 3.7) virtual env (or conda env) with the following steps:

1. Installing virtualenv

pip install virtualenv

2. Create a virtual with Python 3.6

virtualenv trvae-env --python=python3.6 

3. trVAE package installation

To install the latest version from PyPI, simply use the following bash script:

pip install trvae

or install the development version via pip:

pip install git+https://github.com/theislab/trvae.git

or you can first install flit and clone this repository:

git clone https://github.com/theislab/trVAE
cd trVAE
pip install -r requirements
python setup.py install 

Examples

• For perturbation prediction and batch-removal check this example from Haber et al.

Reproducing paper results:

In order to reproduce paper results visit here.

Reference

If you found trVAE useful please consider citing the published manuscript.