Can Voicefixer directly generate audio from the mel - spectrogram in png format?

[shuichuanqhs]

Hello, author. First of all, thank you for your excellent work. I wonder if Voicefixer can directly generate audio from the mel - spectrogram in png format?