SpeechMOS
SpeechMOS allows effective subjective speech score prediction in minimal lines of code using MOS systems. It evaluates audio naturalness through a pre-trained UTMOS model, suitable for TTS model assessments. Built on PyTorch, it necessitates no extra library imports, supporting Python 3.8 with torch and torchaudio compatibility. It supplies precise quality scores from various models, aiding audio analytics and research.
