Our Software Packages

Many of our publications include source code releases. This page lists more extensive, purpose-built software packages.

 

BackPACK — Get more out of your backward pass

Written by Felix Dangel and Frederik Kunstner, BackPACK is a library built on top of PyTorch to make it easy to extract more information from a backward pass. 

 

Install with pip install backpack-for-pytorch

DeepOBS — A Deep Learning Optimizer Benchmark Suite

Written by Frank Schneider and Lukas Balles, DeepOBS provides a benchmark suite for deep learning optimization methods. It aims to help researchers streamline the process of developing new optimizers, by offering a collection of test problems (dataset and architecture) of varying complexity, baselines for a variety of optimizers, and even an automated pipeline for generating comparison plots in papers.

 

Install with pip install deepobs

Entropy Search — Information-efficient global optimization

Originally written by Philipp Hennig and Christian Schuler, later expanded by Edgar Klenske, Robert Eisele and others, Entropy Search is a global optimization / experimental framework designed to extract information about the location of the extremum instead of trying to collect low function values. 

 

The software package was originally released in 2012 and, although repeatedly updated afterward, is not maintained any longer. However, the open-source code has found its way into third-party packages, such as emukit.