How to feed train、val、test data to the dgl model respectively?

here is my question

when applying a DL task, a common way is deviding the whole dataset into three parts, one for trainning, one for evaluate and one for test.
I know there got API in dgl which can help me to devide the dataset,but I don’t know how to feed the model for trainning(or val,or test) with part of the dataset after I input all the graph structure information to create a graph

so I want know if there is an easy way to achieve this goal when using dgl


See my reply at the issue.