Open izzajalandoni opened 1 year ago
Not really an issue, but any attempts on recreating FastSpeech 2s? Or plans to recreate it?
Am I right in understanding input of 2s is the output from variance adaptor, right? And not mel spectrogram..?
Not really an issue, but any attempts on recreating FastSpeech 2s? Or plans to recreate it?
Am I right in understanding input of 2s is the output from variance adaptor, right? And not mel spectrogram..?