-
Notifications
You must be signed in to change notification settings - Fork 41
Open
Description
I have been trying to fix this error for a while, with no effectiveness. I have seen the other posts in here, none of them worked for me. This error occurs when I use the API mode of the library (I don't get any errors when I'm using the Python Module), I get the error:
2025-11-30 21:16:02.796 | INFO | rvc_python.api:rvc_convert:40 - Received request to convert audio
2025-11-30 21:16:02 | WARNING | rvc_python.modules.vc.modules | Traceback (most recent call last):
File "E:\AI\AI bot\RVCvenv\lib\site-packages\rvc_python\modules\vc\modules.py", line 184, in vc_single
audio_opt = self.pipeline.pipeline(
File "E:\AI\AI bot\RVCvenv\lib\site-packages\rvc_python\modules\vc\pipeline.py", line 359, in pipeline
pitch, pitchf = self.get_f0(
File "E:\AI\AI bot\RVCvenv\lib\site-packages\rvc_python\modules\vc\pipeline.py", line 157, in get_f0
f0 = self.model_rmvpe.infer_from_audio(x, thred=0.03)
File "E:\AI\AI bot\RVCvenv\lib\site-packages\rvc_python\lib\rmvpe.py", line 605, in infer_from_audio
hidden = self.mel2hidden(mel)
File "E:\AI\AI bot\RVCvenv\lib\site-packages\rvc_python\lib\rmvpe.py", line 584, in mel2hidden
hidden = self.model(mel)
File "E:\AI\AI bot\RVCvenv\lib\site-packages\torch\nn\modules\module.py", line 1518, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
File "E:\AI\AI bot\RVCvenv\lib\site-packages\torch\nn\modules\module.py", line 1527, in _call_impl
return forward_call(*args, **kwargs)
File "E:\AI\AI bot\RVCvenv\lib\site-packages\rvc_python\lib\rmvpe.py", line 409, in forward
x = self.cnn(self.unet(mel)).transpose(1, 2).flatten(-2)
File "E:\AI\AI bot\RVCvenv\lib\site-packages\torch\nn\modules\module.py", line 1518, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
File "E:\AI\AI bot\RVCvenv\lib\site-packages\torch\nn\modules\module.py", line 1527, in _call_impl
return forward_call(*args, **kwargs)
File "E:\AI\AI bot\RVCvenv\lib\site-packages\rvc_python\lib\rmvpe.py", line 367, in forward
x, concat_tensors = self.encoder(x)
File "E:\AI\AI bot\RVCvenv\lib\site-packages\torch\nn\modules\module.py", line 1518, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
File "E:\AI\AI bot\RVCvenv\lib\site-packages\torch\nn\modules\module.py", line 1527, in _call_impl
return forward_call(*args, **kwargs)
File "E:\AI\AI bot\RVCvenv\lib\site-packages\rvc_python\lib\rmvpe.py", line 244, in forward
x = self.bn(x)
File "E:\AI\AI bot\RVCvenv\lib\site-packages\torch\nn\modules\module.py", line 1518, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
File "E:\AI\AI bot\RVCvenv\lib\site-packages\torch\nn\modules\module.py", line 1527, in _call_impl
return forward_call(*args, **kwargs)
File "E:\AI\AI bot\RVCvenv\lib\site-packages\torch\nn\modules\batchnorm.py", line 171, in forward
return F.batch_norm(
File "E:\AI\AI bot\RVCvenv\lib\site-packages\torch\nn\functional.py", line 2478, in batch_norm
return torch.batch_norm(
RuntimeError: "batch_norm" not implemented for 'Half'
2025-11-30 21:16:02.984 | ERROR | rvc_python.api:rvc_convert:57 - 'tuple' object has no attribute 'dtype'
INFO: 127.0.0.1:60576 - "POST /convert HTTP/1.1" 500 Internal Server Error
I'm a beginner programmer, so I don't know much about this stuff, so I would be thankful if someone could help me on this please. By the way, I don't if this is important, but I'm using Python 3.10.7 and torch 2.1.1+cu118, on Windows 10.
Metadata
Metadata
Assignees
Labels
No labels