training and validation datasets missing #2
Description
It's very nice that this repository provides a small toy dataset to test inference on.
I know that the entire DUD-E dataset can be downloaded from the internet. In the DenseFS paper, the dataset has been processed to perform ligand-protein docking. To make the paper reproducible, it would be very helpful if the original training dataset (DUD-E, with protein-ligands docked) and validation datasets (a customised subset of ChEMBL and MUV) would be made accessible as well.