You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Thanks for a great library! While using it for Yiddish, I noticed that some of the transliterations do not conform to the YIVO romanization standard.
To pinpoint what kind of errors uroman is making, I conducted a romanization experiment using the data from Saleva (2020) and another library for Yiddish romanization called yiddish.
Would it be possible to implement the -l yid flag such that the output conforms to the YIVO romanization standard?
As far as I'm aware, it's by far the most used romanization format for Yiddish.
Thanks!
The text was updated successfully, but these errors were encountered:
Thanks for a great library! While using it for Yiddish, I noticed that some of the transliterations do not conform to the YIVO romanization standard.
To pinpoint what kind of errors
uroman
is making, I conducted a romanization experiment using the data from Saleva (2020) and another library for Yiddish romanization calledyiddish
.Here are some benchmark numbers using accuracy and mean F1 score as defined in Proceedings of the Seventh Named Entities Workshop:
The diffs for what uroman gets wrong can be found here. Many seem to be
i/y
mismatches as well as Hebrew expansion errors:Would it be possible to implement the
-l yid
flag such that the output conforms to the YIVO romanization standard?As far as I'm aware, it's by far the most used romanization format for Yiddish.
Thanks!
The text was updated successfully, but these errors were encountered: