report: An empirical study of neural ordinal differential equations

The ODE kernel is based on torchdiffeq, for the details please refer to the README of that Repo.

Run

python run_image_classification.py

To compare NeuralODE with ResNet on MNIST / CIFAR10.

Some running examples:

python run_image_classification.py --epochs 200 --dataset cifar10 --batch-size 512 --double "" --network odenet --gpu 1 --lr 0.1 --momentum 0.95 --adjoint 1 --method midpoint

Some running examples:

python run_ode_fitting.py --model spiralNN --viz --adjoint

Model and code are based on the paper: Deep Multi-Output Forecasting

Running examples:

python multi-output-glucose-forecasting/run.py

For now visdom can fetch all the csv files following the particular format and plot them.

Go to the Visdom folder then execute the following commands:

visdom -port XXXXX

python visdom_pull_server.py -port XXXXX