alphacep / vosk-api

Offline speech recognition API for Android, iOS, Raspberry Pi and servers with Python, Java, C# and Node
Apache License 2.0
8.15k stars 1.12k forks source link

Spanish recognition is missing accents #1663

Open VicJ4r4 opened 1 day ago

VicJ4r4 commented 1 day ago

Thanks for reading. English is not my native language. Is anyone here who speaks Spanish? I'm using the vosk-model-small-es-0.42 Spanish model in a Debian based distro. The problem is that it doesn't properly detect the accent marks above the vowels. For example, when I say something like:

aquel artículo de metodología científica ayudará tanto a los jóvenes estudiantes como a los viejos académicos

I get a result like:

aquel artculo de metodologa cientfica ayudar tanto a los jvenes estudiantes como a los viejos acadmicos

What I usually have to do is edit it with the spelling checker of word processors...

nshmyrev commented 1 day ago

Please provide an audio sample for the above