PyTorch implementations of a generalized ResNet-inspired network architecture which allows for broader experimentation.
Start up your python environment and install the python packages stored in requirements.txt:
pip3 install -r requirements.txt
Run the default network (RK4 scheme using a doubleSymLayer on the CIFAR-10 dataset):
python3 RKNet.py
The concepts behind the networks implemented by this toolbox are detailed in:
Lars Ruthotto and Eldad Haber (2018). Deep Neural Networks Motivated by Partial Differential Equations. arXiv.org.
Eldad Haber and Lars Ruthotto (2017). Stable architectures for deep neural networks. Inverse Problems, 34(1).
Bo Chang, Lili Meng, Eldad Haber, Lars Ruthotto, David Begert, and Elliot Holtham (2018). Reversible architectures for arbitrarily deep residual neural networks. Presented at the Thirty-Second AAAI Conference on Artificial Intelligence.
This material is in part based upon work supported by the US Israel National Science Foundation Grant Number 2018209 and the US National Science Foundation under Grant Number DMS-1751636. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.