Can I use DGL graph to train on DGL-KE?

DGL-KE
#1

Hi Team,

again I am quite new to this so

I have a graph that I created in dgl module and I want to train this graph in DGL-KE (using TransE) how should I go about doing it, is it possible?

G = nx.Graph()
G.add_nodes_from([(0, {“name_”:“Biryani”, “type_”:“Recipe”}),
(1, {“name_”:“Best Lemonade”, “type_”:“Recipe”})])

G.add_nodes_from([(2, {"name_":"blueberries", "type_":"Ingredient"}),
                (3, {"name_":"granulated sugar", "type_":"Ingredient"}),
                (4, {"name_":"vanilla yogurt", "type_":"Ingredient"}),
                (5, {"name_":"lemon juice", "type_":"Ingredient"}),
                (6, {"name_":"peppercorns", "type_":"Ingredient"})])

G.add_edges_from([(0,3),(0,4),(0,5),(1,4),(1,3),(1,6)])

thanks a lot
import dgl
g = dgl.from_networkx(G)

#2

You can look at this branch: GitHub - awslabs/dgl-ke at ng. This is feature is under developmemnt.