Open coding-sharks opened 8 months ago
I am interested in “How to use mamba to generate audio”. One of amazing things is the long sequence attention, i wanna know whether mamba can be used in TTS, so that it does not need the Vocoder. Maybe it will create a more "End-to-End" TTS?
a related experiment can be found here
I am interested in “How to use mamba to generate audio”. One of amazing things is the long sequence attention, i wanna know whether mamba can be used in TTS, so that it does not need the Vocoder. Maybe it will create a more "End-to-End" TTS?