Data Production Scripts?

I was wondering if you can provide the scripts for producing this data, since the only scripts available here are just for evaluation? I would like to scale up the complexity of the data and see how models fare

Start by inspecting the repository's existing evaluation scripts and determine how they relate to the data used by the models. Done means providing scripts that produce the data and support increasing its complexity so model performance can be evaluated at larger scales.

Written by the indexing model from the issue text.