Edge features thrown out

Hi, thanks for this implementation!

I was wondering if the pytorch-geometric implementation of this architecture is throwing the edge features out by mistake, as seen here

egnn-pytorch/egnn_pytorch/egnn_pytorch.py

Lines 148 to 151 in 1b8320a

    
           if edge_attr is None: 
        
               edge_attr = torch.cat([edge_attr, rel_dist], dim=-1) 
        
           else: 
        
               edge_attr = rel_dist

Or maybe my understanding is wrong?
Cheers,

@josejimenezluna oh yes, you are correct https://github.com/lucidrains/egnn-pytorch/releases/tag/0.0.14

Closing this 👍

	if edge_attr is None:
	edge_attr = torch.cat([edge_attr, rel_dist], dim=-1)
	else:
	edge_attr = rel_dist