Closed JuliaGast closed 5 months ago
It would be great to have these two properties for each dataset.
num_nodes
num_rels
Otherwise I have to compute it for every method (also fine)
num_rels = len(set(dataset.edge_type))
num_nodes = len(set(dataset.full_data['sources']+ dataset.full_data['destinations']))
Yes, this can be supported in the next iteration.
issue addressed in #8
great thanks. closing this.
It would be great to have these two properties for each dataset.
num_nodes
: number of distinct node ids in train+valid+test setnum_rels
: number of distinct edge_types in train+valid+test setOtherwise I have to compute it for every method (also fine)
num_rels = len(set(dataset.edge_type))
num_nodes = len(set(dataset.full_data['sources']+ dataset.full_data['destinations']))