capspeech/eval/README.md

CapSpeech Evaluation Tools

Get Start

Install dependicies:

conda create -n capeval python=3.9
conda activate capeval
pip install -r requirements.txt
pip install git+https://github.com/sarulab-speech/UTMOSv2.git

For ASR, we need:

conda install ffmpeg

Evaluate pitch, monotony, speed, age, gender

RUN:

python base_eval.py

Evaluate UTMOSv2

RUN:

python mos_eval.py

Evaluate ASR Results

RUN:

python asr_eval.py

Evaluate emotion, accent

RUN:

cd src/example/
python categorized_emotion.py
python dialect_world_dialect.py

Please refer to Vox-profile for more evaluation tools.