Could the developers at Facebook schedule a new training on the new data (so much has appeared on the web in 5 years!) to release an updated model?
I think a lot of people would benefit from an update. I'm sure accuracy would improve because the number of documents even on Wikipedia and Tatoeba has increased tremendously.
I don't know the exact date when the
lid.176.bin
was released, but according to the web archive (https://web.archive.org/web/20180104060303/https://fasttext.cc/docs/en/language-identification.html) it's been over 5 years.Could the developers at Facebook schedule a new training on the new data (so much has appeared on the web in 5 years!) to release an updated model?
I think a lot of people would benefit from an update. I'm sure accuracy would improve because the number of documents even on Wikipedia and Tatoeba has increased tremendously.