Exactly don't grasp each node of mask of role

[tteon]

Hi there
I’m now studying GNN for recommender system
especially , i stuck in data preprocessing section which is actually train_mask … , etc. part

my question is this.

1. is it equation at ML modeling splitting train , val , test role ? if it correct then ratio is still maintain?

2. I maybe solving abovementitoned things how can i handling them at code treatment? any reference recommend plz

thanks a lot your attention

[mufeili]

1. train_mask is a binary tensor indicating the nodes in the training set. The split can vary depending on the dataset.
2. You just need to randomly select indices separately for the training, validation and test set.