Data for paper "SizeShiftReg: a Regularization Method for Improving Size-Generalization in Graph Neural Networks", by Davide Buffelli, Pietro Liò, and Fabio Vandin, accepted at NeurIPS 2022.
Additional file 4. The additional file (S3.xlsx) provides tables examining the hyperparameter sensitivity used to evaluate Cellograph's robustness on the organoid dataset described in the main text.