I want to train vosk model vosk-model-en-us-0.22 from https://alphacephei.com/vosk/models with an addition data of my voice with transcript of 1 hour so that the model gets overfit with my voice and don't get problem recognizing my voice properly. So how can I train it if someone can give me roadmap how to train that as I am beginner in open source.
- How to record my voice in which type i.e. mp3, wav etc.
- How to feed recording dataset and transcript to model.
- What code I use for training it and testing it.
- How to make it better. And any other related query
I tried reading https://alphacephei.com/vosk/lm but could not help myself with it.