logWMSE, an audio quality metric & loss function with support for digital silence target. Useful for training and evaluating audio source separation systems.
audio python music machine-learning ai deep-learning torch sound-processing loss-functions bss mss audio-separation music-source-separation speech-denoising audio-quality audio-source-separation sound-separation audio-denoising sound-source-separation audio-quality-assessment
-
Updated
Apr 24, 2025 - Python