Inaccurate language detection for short phrases like greetings
#1
by lasalH - opened
I have been using the facebook/fasttext-language-identification library for language detection in my project, and I have noticed that it fails to accurately detect the language for short phrases such as "Hi" or "Hello". Furthermore, this issue is persistent in other languages as well. I would appreciate any assistance or guidance in resolving this issue, as accurate language detection is critical for the success of my project. Thank you.
Same issue
In my opinion, despite the advantage of supporting more languages, this model seems to be less accurate on average than the previous one.