Demo for Clean Speech Testset

Audio Reconstruction from Mel-spectrogram using different models (Compare with GAN).

Ground Truth
Vocos
BigVGAN
RFWave