Installation

You will likely need a few extra packages and libraries to run Karen's core routines. Check out the details below to jumpstart your Karen-experience.

Install Dependencies

sudo apt-get install \
  festival \
  festvox-us-slt-hts  \
  libfann2 \
  python3-fann2 \
  libportaudio2 \
  libasound2-dev

pip3 install karen

NOTICE - Karen is under development against Python 3. She is not compatible with Python 2 so be sure to use pip3 or python3 if appropriate (and install the related binaries if needed).

Install via Download (Alternative)

Make sure you see the requirements.txt for other python libraries that are required. The PIP method is recommended as it will automatically include these dependencies.

cd /path/to/karen
python3 setup.py install

Get the Mozilla DeepSpeech Models

To download the speech models you can use the script below inside a Python shell or visit the DeepSpeech page:

import karen
karen.download_models(version="0.9.3", model_type="pbmm", include_scorer=False)

NOTE: The version number is optional and if ommitted the process will attempt to determine your currently installed version of deepspeech. Also, you will need to specify model_type="tflite" if you are running on Raspberry Pi as the pbmm models are not compatible with the Arm architecture.

  • Once you're finished make sure to read about Starting Up.

Help & Support

Installation instructions and documentation is available at https://projectkaren.ai