Unsupervised Graph Embeddings

I’m trying to do graph embedding using pytorch geometric but i can’t manage to do it in an unsupervised way. For example, the SAGE model is supposed to be trained unsupervised but all i could find was supervised implemention. I saw hat Node2Vec was trained in an unsupervised way but i’d like to apply it to the GNN models.
Thank you for the help!

1 Like