Training without speaker information

alphacep / vosk-api

Offline speech recognition API for Android, iOS, Raspberry Pi and servers with Python, Java, C# and Node

Apache License 2.0

7.38k stars 1.04k forks source link

Training without speaker information #1435

Open itsmeju58 opened 10 months ago

itsmeju58 commented 10 months ago

Hi! I'm trying to train the model according to the instructions for colab with my own dataset. However, I have no information about the speakers. Is it possible to assign one speaker to all recordings, or will this seriously affect accuracy? Or maybe somehow I can disable the use of information about the speakers in the training file? Or is there any other way to solve this problem?

nshmyrev commented 10 months ago

It is better to assign unique speaker to every recording. Just make speaker id the same as utterance id.