This folder contains scripts that help gather the most important files from your already-trained model with Kaldi, and use these files into the online decoder for recognising a single file.
Right now the most common/used/famous API over Kaldi's online decoding
environment is Alpha Cephei's Vosk API. See prep_vosk.sh and
test_vosk_simple.py scripts, which by the way should be executed in this very
same order.
There's also a script to run the Portuguese model from Vosk directly into
Kaldi. See run_vosk_on_kaldi.sh for that.
Grupo FalaBrasil (2021) - https://ufpafalabrasil.gitlab.io/
Universidade Federal do Pará (UFPA) - https://portal.ufpa.br/
Cassio Batista - https://cassota.gitlab.io/