/STT

Speech to Text service

Primary LanguagePython

Speech to Text

app frontend

Small app for Speech To Text model development.

  • Easy to upload audio file
  • Execute model seamlessly on input
  • Output clear and nice for user

Technologies used:

Download models

Download the models from coqui before starting: https://coqui.ai/models. Then change the settings.py file with the correct path to model and scorer.

English

wget https://github.com/coqui-ai/STT/releases/download/v0.9.3/coqui-stt-0.9.3-models.pbmm -P models/english
wget https://github.com/coqui-ai/STT/releases/download/v0.9.3/coqui-stt-0.9.3-models.scorer -P models/english

French

  • wget https://github.com/coqui-ai/STT-models/releases/download/french/commonvoice-fr/v0.6/model.pbmm -P models/french_common
wget https://github.com/coqui-ai/STT-models/releases/download/french/commonvoice-fr/v0.6/fr-cvfr-2-prune-kenlm.scorer -P models/french_common
  • wget https://github.com/coqui-ai/STT-models/releases/download/french/jaco-assistant/v0.0.1/model.pbmm -P models/french_jaco
wget https://github.com/coqui-ai/STT-models/releases/download/french/jaco-assistant/v0.0.1/kenlm_fr.scorer -P models/french_jaco

Start app

Run following command to start app:

docker-compose up -d

You can then access the app at http://localhost:8501.

Some test audio files are in the audio folder.