Improving Neural Ordinary Differential Equations with Nesterov's Accelerated Gradient Method

This is the official implementation of Improving Neural Ordinary Differential Equations with Nesterov's Accelerated Gradient Method (NeurIPS 2022).

Main requirements

The code base in in Python. The following packages are required:

• torch
• torchvision
• torchdiffeq
• tqdm
• imageio
• einops These packages can be installed with the following command:

pip install torch torchvision torchdiffeq tqdm imageio einops

Additionally, if you want to run the visualization code, the numpy, pandas, matplotlib are also required. These are well-supported package, so if there's any problem with the installation process, please refer to the official installation guide.

Details for each experiments can be found in

MNIST

mnist/README.md

CIFAR

cifar/README.md

Point Cloud

point-cloud/README.md

Walker 2D

walker2d/README.md

Silverbox Initialization

README-sb.md