Simon Ott
Simon Ott
Hello, We are using dglke for baseline evaluation of our dataset. However RESCAL outperforms other approaches by quite a lot: more than 20% more hits@1, which seems a little odd....
Closes #427 Dataset contains 8 different subset_id's (different dataset settings), each with a `bigbio` and `source` schema. Furthermore there is an subset called `mediqa_ans_all` which includes all data (articles, sections,...
Closes #502 This is a QnA dataset that supports two languages en and es, so there are two subsets containing the same questions: `head_qa_en` and `head_qa_es`. I implemented also a...
- **Name:** BioMRC - **Description:** BioMRC: A Dataset for Biomedical Machine Reading Comprehension - **Task:** cloze-style MRC / multiple choice QA - **Paper:** https://arxiv.org/abs/2005.06376 - **Data:** https://archive.org/details/biomrc_dataset - **License:** Unknown...
Hello, First of all thanks for your work, very helpful framework! I've got a quick question about hyperparameter search and the uniform initialization interval. In the configs used for the...
Added support for Windows 10 desktop notifications, if knockknock is run in WSL
There seems to be some issue with the dropbox, because I get an error when opening a link for a pretrained model, e.g: https://www.dropbox.com/s/y8uucaajpofct3x/codex-m-lp-transe.pt?dl=1
No pressing issue (I just used one of your releases) but maybe you should know that I hat troubles cloning the repo for some reason: ```bash PS C:\Users\ottsi> git clone...
Is there a reason you omit the validation triples of the inference (test-ind) graph from filtering in your example [here](https://pykeen.readthedocs.io/en/stable/tutorial/inductive_lp.html#full-inductive-lp-example) when doing evaluation on the test set? I would expect...