This repository integrates Pykaldi Speech Recognition Toolkit with a text classifier to obtain the sentiment of the user's command (locate, describe, no_op).
- Create an environment using Virtualenv or Conda
- Clone the Pykaldi repository from https://github.com/pykaldi/pykaldi.git and build
from source by following its
README
installation instructions - In the
zamia
directory, run./models.sh
to download the zamia model - Set the source to the kaldi installation by running
source path.sh
- Run
./decode_live_smart_test.py
to start sentiment analysis