Retrieval-based-Voice-Conversion-WebUI

Retrieval-based-Voice-Conversion-WebUI copied to clipboard

would it be possible to add a third route to a model C to combine with models A and B? I tried to modify the section in infer-web.py and managed...

Mixomo

When training a model is finished, I have numerous pth files, each one saved at intervals of 25 epochs, plus one with no numbers in the name. DO I need...

urasam2

pamzerbhu

During handling of the above exception, another exception occurred: Traceback (most recent call last): File "E:\Nero\Sound\Retrieval-based-Voice-Conversion-WebUI-main\trainset_preprocess_pipeline_print.py", line 79, in pipeline audio = load_audio(path, self.sr) File "E:\Nero\Sound\Retrieval-based-Voice-Conversion-WebUI-main\my_utils.py", line 19, in load_audio...

crygvr

OutOfMemoryError

1

When I try training it on a 1050, I get the following error, before it fails: ``` Traceback (most recent call last): File "C:\ProgramData\anaconda3\envs\rvc\lib\multiprocessing\process.py", line 314, in _bootstrap self.run() File...

AyliasTheCoder

Another exception occurred while processing the above exception: Backtracking (the most recent call): File "E:\Nero\Sound\Retrieval-based-Voice-Conversion-WebUI-main\trainset_preprocess_pipeline_print.py ", line 79, in the pipeline audio = load_audio(path, self.sr ) File "E:\Nero\Sound\Retrieval-based-Voice-Conversion-WebUI-main\my_utils.py ", line...

crygvr

Can RVC, like training in other fields, support the feature of resuming training from the interrupted position? If it doesn't currently support it, will it be added in future versions?

6get-xiaofan

I messed around with the trainset_preprocess_pipeline_print.py using ChatGPT, and I managed to preprocess my audio with no errors. If anyone else has the "Failed to load audio" error, maybe try...

CeciliaXCIX

go-web.bat doesn't open the website

1

When I click go-web.bat it doesn't open the website and cmd doesn't list the command, How to fix it please?

Hugoyim0097

Because retrieval features are used during inference to replace input features in order to prevent speaker identity leakage, but how can we ensure that the generated speech still corresponds to...

suzhenghang