Benchmark#
We provide overview of the methods used to generate our benchmark dataset here. More details on each dynamical class within the hierarchy can be found in:
Dataset Summary#
Although we open-sourced our data generating script, we preprocess thousands of graphs with tens of million of samples with the following distribution.
NOTE: We provide several scripts to generate benchmark data for simple, coupled and climate models. Have a look at
scripts/README.md
for more information on how to run them.