How to use? 
Run this Python command: 
python main.py --model <model_name> --duration <time_in_seconds> --sample_rate <sample_rate_in_hz> 

We have 3 models: 
- soniox (takes number of seconds as a duration for a prerecorded audio and sample_rate of 8000Hz) [defualt=3 seconds]
- soniox_microphone 
- picovoice 

Example: 
python main.py --model soniox --duration 5 --sample_rate 16000