GraphSage on unseen nodes

[girwin1]

I have trained a graphsage model on a graph. I now want to get the embeddings for new unseen nodes.
So I pass to the model the node features and the edge index but then i get this error:

features = torch.FloatTensor(feature_list)
edges = torch.Tensor([src, dst]).type(torch. int64)
model(features, edges)

IndexError: Dimension out of range (expected to be in range of [-1, 0], but got -2)

Any ideas?

Also, the node has edges to other unseen nodes. Should these edges be included in the data passed to the model?

Thank you for any help!

[rusty1s]

What is the final shape of features and edges?

[girwin1]

Features is [9] as i have 1 node with 9 features and edges is [2,7]

[rusty1s]

How can you only have 1 node, but 7 edges? It looks like you are missing features of neighbors.

[rusty1s]

Also, the node has edges to other unseen nodes. Should these edges be included in the data passed to the model?

Ideally, yes.