search

IBM / dualtkb

https://arxiv.org/abs/2010.14660
Apache License 2.0
17 stars 1 forks source link

Will the dataset be released soon? #1

Open ChunhuaLiu596 opened 3 years ago

ChunhuaLiu596 commented 3 years ago

Hey, I am really interested in this work and want to know whether the data are available now?

pdognin commented 3 years ago

Thanks for your question We discussed a lot about finding a way to release at least some alignments we used between text and triples since we do not own the data.
We performed quite a bit of pre-processing/filtering prior to generating the text,triple pairs, so getting the original alignments will require some work which is hard to allocate right now. In the meantime, we did move on to more curated datasets like WebNLG.