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
There are many other examples of this behaviour. By turning on debugging it looks like it could be an issue with the tokenizer or token features. I have very little knowledge of Japanese or how this translation works so I can't provide much more insight.
Thanks!
The text was updated successfully, but these errors were encountered:
Hi! In every case I've seen where translation fails, the output is equal to the input apart from some non-Katakana characters which do get translated properly (the 空港 for airport locations).
I was hoping for Romaji representations of these location names.
Here are the full debug outputs for the above examples:
I haven't looked into performance on Hiragana much as I'm dealing exclusively with locations, which to my knowledge are usually written in either Kanji or Katakana.
Here are some examples of translations that work as I had hoped:
Translation fails for some locations which have names that have been translated from another language into Japanese.
For example:
アボッツフォード空港 (Abbotsford Airport)
アンガルシー (Anglesey)
コンスタンツァ (Constanta)
エスペランス空港 (Esperance Airport)
There are many other examples of this behaviour. By turning on debugging it looks like it could be an issue with the tokenizer or token features. I have very little knowledge of Japanese or how this translation works so I can't provide much more insight.
Thanks!
The text was updated successfully, but these errors were encountered: