I was wondering if it would be at all possible to access the test predictions during the training phase of the link prediction pipeline to better understand the types of predictions the model is getting right and wrong. So I would like to be able to see the set of nodes, test prediction, and actual label (0 or 1). In the logs I can see some of the training metrics but not the actual test predictions themselves. Please let me know if this would be possible within the pipeline. Thank you
I can add the feature as a roadmap candidate, and then it might be included in a subsequent release of the library. We have a lot of things we want to do for upcoming releases so cannot promise we'll get to this in the near future however. It depends on how it will be prioritized internally.