Edge classification boilerplate

I have a graph with following properties:

Graph(num_nodes=100, num_edges=17144,
      ndata_schemes={'feat': Scheme(shape=(2,), dtype=torch.float64), 'label': Scheme(shape=(), dtype=torch.int64), 'train_mask': Scheme(shape=(), dtype=torch.bool), 'val_mask': Scheme(shape=(), dtype=torch.bool), 'test_mask': Scheme(shape=(), dtype=torch.bool)}
      edata_schemes={'weight': Scheme(shape=(18,), dtype=torch.float64)})

I want to perform edge classification on this graph, so following functions are needed:

class GCN():
    def __init__(self):

    def forward(self):

def train():

def test():

can anyone please provide some starting point to complete these functions

Will passing in concatenated features [src, edge, dst] in network same as node classification one do the job? though will need some help with the theory if this is the correct way

Hi, we have a user guide chapter talking about how to implement an edge classification model step-by-step. 5.2 Edge Classification/Regression — DGL 0.6.1 documentation . Please check it out.

1 Like

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.