pykeen / benchmarking

📊 Results from the reproducibility and benchmarking studies presented in "Bringing Light Into the Dark: A Large-scale Evaluation of Knowledge Graph Embedding Models Under a Unified Framework" (http://arxiv.org/abs/2006.13365)
MIT License
35 stars 4 forks source link

Wrong directory structure #10

Closed cthoyt closed 4 years ago

cthoyt commented 4 years ago

https://github.com/mali-git/pykeen_experimental_results/tree/master/ablation/results/adadelta

this doesn't make sense. Only model folders should be in results

@mali-git could you please fix the organization here

@migalkin @mberr @lvermue idk if you put this here but are you also aware of the way this is all organized?

mali-git commented 4 years ago

yes, I will fix that. There was an issue because we used an unfavourable hyper-paramerter value for adadelta, that's why we re-run the adadelta experiments and kept the results separate in order to not mix them up with the exisitng adadelta results. I will clean up today the whole folder.

mali-git commented 4 years ago

I refactored the results diretory.

cthoyt commented 4 years ago

Is there a specific commit we can reference and close this issue?

mali-git commented 4 years ago

Yes, 14f2eb97103dff61435ee793bf6d61f6452de7c7

cthoyt commented 4 years ago

Thanks