Closed tomas122727 closed 6 months ago
The hubert model is loaded in half precision mode to save memory. I suspect it may not work on CPU inference, but I am not in a position to test right now. Do you know if pytorch can see your GPU?
I do not know exactly what python version is required, but my conda environment for so-vits-svc 5.0 uses python 3.10.13.
yes it can
even my previous so vits 4.0 works fine and so vits svc 3.0 works fine
it also keeps saying this at launch INFO: Could not find files for the given pattern(s).
nevermind i just fixed it by reinstalling torchcrepe module
it wasnt compiled by cuda keep this issue archieved in case someone runs into this
Error code C:\Users\tomas_000\miniconda3\envs\so_svc5\lib\site-packages\torch\nn\utils\weight_norm.py:28: UserWarning: torch.nn.utils.weight_norm is deprecated in favor of torch.nn.utils.parametrizations.weight_norm. warnings.warn("torch.nn.utils.weight_norm is deprecated in favor of torch.nn.utils.parametrizations.weight_norm.") Loading Fluttershy models\Fluttershy\singer\Fluttershy.spk.npy Traceback (most recent call last): File "M:\so-vits-svc-5.0-main\inference_gui.py", line 1149, in convert out_audio = self.infer_tool.infer( File "M:\so-vits-svc-5.0-main\inference_util.py", line 302, in infer vec = np.array(self.pred_vec(audio_data)) File "M:\so-vits-svc-5.0-main\inference_util.py", line 129, in pred_vec vec = self.hubert.units( File "C:\Users\tomas_000\miniconda3\envs\so_svc5\lib\site-packages\torch\utils_contextlib.py", line 115, in decorate_context return func(*args, kwargs) File "M:\so-vits-svc-5.0-main\hubert\hubertmodel.py", line 71, in units x, = self.encode(wav) File "M:\so-vits-svc-5.0-main\hubert\hubert_model.py", line 42, in encode x = self.feature_projection(x.transpose(1, 2)) File "C:\Users\tomas_000\miniconda3\envs\so_svc5\lib\site-packages\torch\nn\modules\module.py", line 1511, in _wrapped_call_impl return self._call_impl(*args, *kwargs) File "C:\Users\tomas_000\miniconda3\envs\so_svc5\lib\site-packages\torch\nn\modules\module.py", line 1520, in _call_impl return forward_call(args, kwargs) File "M:\so-vits-svc-5.0-main\hubert\hubert_model.py", line 106, in forward x = self.norm(x) File "C:\Users\tomas_000\miniconda3\envs\so_svc5\lib\site-packages\torch\nn\modules\module.py", line 1511, in _wrapped_call_impl return self._call_impl(*args, *kwargs) File "C:\Users\tomas_000\miniconda3\envs\so_svc5\lib\site-packages\torch\nn\modules\module.py", line 1520, in _call_impl return forward_call(args, **kwargs) File "C:\Users\tomas_000\miniconda3\envs\so_svc5\lib\site-packages\torch\nn\modules\normalization.py", line 201, in forward return F.layer_norm( File "C:\Users\tomas_000\miniconda3\envs\so_svc5\lib\site-packages\torch\nn\functional.py", line 2546, in layer_norm return torch.layer_norm(input, normalized_shape, weight, bias, eps, torch.backends.cudnn.enabled) RuntimeError: "LayerNormKernelImpl" not implemented for 'Half'
how to fix this also what python version am i suppose to install on my pc ?