Michael Galkin

That's a great initiative, I can assist here - my laptop is a very basic M1 (8 GB) with mps enabled (torch 1.13 dev from the nightly branch). For the...

> @migalkin - just wanted to confirm with you that the evaluation loops is also failing currently due to an error with Tensor indexing? Generally, running a script with the...

Complete reproduction [of their code would require](https://github.com/LongYu-360/TripleRE-Add-NodePiece/blob/994216dcb1d718318384368dd0135477f852c6a4/TripleRE%2BNodepiece/ogb_wikikg2/model.py#L195-L205) joint Anchor + Relation tokenization 😉

I'd be in favor of merging this model without that special OGB evaluator

Oops, my bad 😬 I was attending the poster session and wrote down a few relevant KGE models. Brought back the original template!

Loading of hyper-relational WD50K datasets works, you can already try it with smth like ```python from pykeen.datasets.wd50k import WD50K_100 def main(): dataset = WD50K_100( create_inverse_triples=True, max_num_qualifier_pairs=5, eager=True ) print(dataset.num_entities) if...

There is a lot of code repetition in: * `parse_structure_MP`, `parse_structure_NOMAD`, `parse_structure_OQDM`, `parse_structure_Carolina`: essentially, they all can be wrapped within one function and, depending on the particular `dataset_name`, the final...

The PR looks much better now! Is can still be polished further though to remove unnecessary code repetitions in other functions (highlighted above^)

Hi, thanks for the questions, let's clarify those: > The train graph is available during training, and during evaluation/test, the inference graph is available. Or is the training graph still...