myshell-ai
Synthesised Audio is nothing like Reference Audio
I have the Elon voice that the webui opens with. The webui runs okay and produces a new audio. But nothing like elon's voice. I get the following error. I found another post stating 'The following output is a normal information indicating that the checkpoint has been successfully loaded'
https://github.com/myshell-ai/OpenVoice/issues/37
(.venv) D:\techy\talkingHeads\OpenVoice> d: && cd d:\techy\talkingHeads\OpenVoice && cmd /C "d:\techy\talkingHeads\OpenVoice.venv\Scripts\python.exe c:\Users\User.vscode\extensions\ms-python.debugpy-2024.8.0-win32-x64\bundled\libs\debugpy\adapter/../..\debugpy\launcher 60881 -- D:\techy\talkingHeads\OpenVoice\openvoice_app.py "
Loaded checkpoint 'checkpoints/base_speakers/EN/checkpoint.pth'
missing/unexpected keys: [] []
Loaded checkpoint 'checkpoints/base_speakers/ZH/checkpoint.pth'
missing/unexpected keys: [] []
Loaded checkpoint 'checkpoints/converter/checkpoint.pth'
missing/unexpected keys: [] []
d:\techy\talkingHeads\OpenVoice.venv\lib\site-packages\gradio\components\dropdown.py:103: UserWarning: The max_choices parameter is ignored when multiselect is False.
warnings.warn(
Running on local URL: http://127.0.0.1:7861
To create a public link, set share=True in launch().
New to coding. What am I missing here please.
I'm seeing the same and I'm new to neither coding nor python coding. I have found that the missing keys print is unimportant here, but I'm running my own script based of the third example and I'm getting audio that sounds like someone very unconvincingly trying to fake the voice rather than a clone of the voice. I've even tried to grab a few of the reference recordings and it still doesn't sound anything like the output that is given with the reference recordings.
