How to use mini-batch dataset?

[taylorlulu]

I found that the dataset Amazon Berkeley 3D models dataset (abo-3dmodels.tar) which is 154Gb and the Google Scanned Objects dataset is too large for my computer. I wonder if there's any way I can just run the code on YCB objects and Replica-CAD dataset dataset.

How can I edit the configuration files?
I've tried the beechwood_0_int which says

I would appreciate a lot for your help.