Open lsorber opened 5 years ago
Agree. There are also a lot domain-specific words missing from Wikipedia like e.g. webdav, kubernetes, etc. Are there any plans to consider corpuses with domain specific words as a basis for numberbatch? 'webdav' for example is included in the GloVe vectors. I can't find it in numberbatch.
There seems to be a numberbatch-20180108 floating around.
Referenced here: http://blog.conceptnet.io/posts/2018/distinguishing-attributes-using-conceptnet/ And here: https://github.com/LuminosoInsight/semeval-discriminatt/blob/master/discriminatt/classifier.py
Looks like there's a new version out [1], kudos to the authors for the new release!
What's the chances of a newer version now that COVID/etc terms are running rampant?
It's been a while since the last update, and I was wondering if there are any new versions planned. Thanks for the great work so far!