Some stands on the shoulders of TPUs that do 1 century worth of compute in a day. Other stands on the shoulders of C or Python libraries.
Also, is engineering "just a glorified wrapper around science"?, is physics "just a glorified wrapper around maths"?, and is litterature "just a glorified wrapper around words"?
I'm greatly happy on seeing libraries that work with my mother language on HN :-)
I guess people think that well... 'glorified wrappers' aren't worth discussion on HN because there's too many NLP libraries on English? There aren't a lot of libraries around Korean, especially practical ones so I'm happy to see KoLNPy on HN.
This is very cool. While I do not speak Korean, I can appreciate the part-of-speech tagging and the other features, as that's required in any NLP framework. I'm wondering, is there a way to generate a syntax tree using this framework? Just looking at the docs now.
Surprised but happy to see this here. I maintain a few Python wrappers for MeCab and was working on Korean support when I found this - it was better than I whatever I was doing so I linked to it instead.
8 comments
[ 4.2 ms ] story [ 25.2 ms ] threadSome stands on the shoulders of TPUs that do 1 century worth of compute in a day. Other stands on the shoulders of C or Python libraries.
Also, is engineering "just a glorified wrapper around science"?, is physics "just a glorified wrapper around maths"?, and is litterature "just a glorified wrapper around words"?
What's with the condescension culture?
I guess people think that well... 'glorified wrappers' aren't worth discussion on HN because there's too many NLP libraries on English? There aren't a lot of libraries around Korean, especially practical ones so I'm happy to see KoLNPy on HN.