File tree Expand file tree Collapse file tree
Expand file tree Collapse file tree Original file line number Diff line number Diff line change 11"""
22DiffSVC architecture with WaveNet denoiser and NSF-HiFiGAN vocoder.
33
4- Comparing to v1, this version
4+ Comparing to v1, this version
55- Doesn't need spec stats anymore.
66- Added dilation cycle to WaveNet denoiser.
77- Used the log10 mel spectrogram.
Original file line number Diff line number Diff line change @@ -55,9 +55,11 @@ def launch_gradio(
5555 speaker = gr .Dropdown (
5656 label = "Speaker Name (Used for Multi-Speaker Models)" ,
5757 choices = list (speaker_mapping .keys ()),
58- value = speaker
59- if speaker in speaker_mapping
60- else list (speaker_mapping .keys ())[0 ],
58+ value = (
59+ speaker
60+ if speaker in speaker_mapping
61+ else list (speaker_mapping .keys ())[0 ]
62+ ),
6163 )
6264 else :
6365 speaker = gr .Number (
Original file line number Diff line number Diff line change 44It aims to reduce errors for long utterances and detcect aspiration which is not detected by MFA.
55"""
66
7-
87import librosa
98import numpy as np
109import parselmouth as pm
You can’t perform that action at this time.
0 commit comments