-
Notifications
You must be signed in to change notification settings - Fork 41
Description
I downloaded this project from GitHub and tried to run the test program. My Python version is 3.10, PyTorch version is 2.4.0, CUDA version is 11.8, and I'm using a GTX 1660 Super GPU. However, I encountered many issues.
First, three models failed to download from the server, so I manually added the files "Herbert_base.pt", "rmvpe.pt", and "rmvpe.onnx" by these 3 urls **_(1:https://huggingface.co/Daswer123/RVC_Base/resolve/main/hubert_base.pt ; 2:https://huggingface.co/lj1995/VoiceConversionWebUI/blob/main/rmvpe.pt ; 3:https://huggingface.co/lj1995/VoiceConversionWebUI/blob/main/rmvpe.onnx)_**to the .\rvc-python\base_model folder. Then, when running the program, I encountered the following error: ModuleNotFoundError: No module named 'utils',after that,I tried to solve the problem again through pip installer ntils, but it didn't seem to work, and in the end, I got stuck in the following error:
2025-04-04 18:21:04 | INFO | rvc_python.configs.config | Found GPU NVIDIA GeForce GTX 1660 SUPER, force to fp32 2025-04-04 18:21:04 | INFO | rvc_python.modules.vc.modules | Loading: klee.pth overwrite preprocess and configs.json is_half:False, device:cuda:0 Доступные модели: [] Traceback (most recent call last): File "C:\Users\Administrator\Downloads\rvc-python-main\test.py", line 8, in <module> rvc.load_model("klee.pth") File "C:\Users\Administrator\Downloads\rvc-python-main\rvc_python\infer.py", line 92, in load_model self.vc.get_vc(model_path, version) File "C:\Users\Administrator\Downloads\rvc-python-main\rvc_python\modules\vc\modules.py", line 103, in get_vc self.cpt = torch.load(sid, map_location="cpu") File "C:\Users\Administrator\PycharmProjects\pythonProject5\venv\lib\site-packages\torch\serialization.py", line 1014, in load return _load(opened_zipfile, File "C:\Users\Administrator\PycharmProjects\pythonProject5\venv\lib\site-packages\torch\serialization.py", line 1422, in _load result = unpickler.load() File "C:\Users\Administrator\PycharmProjects\pythonProject5\venv\lib\site-packages\torch\serialization.py", line 1415, in find_class return super().find_class(mod_name, name) AttributeError: Can't get attribute 'HParams' on <module 'utils' from 'C:\\Users\\Administrator\\PycharmProjects\\pythonProject5\\venv\\lib\\site-packages\\utils\\__init__.py'>
I'm not sure if this is a bug or if my environment is incorrect, or if my graphics card selection is incorrect, or if the library hasn't been downloaded properly. Can anyone help me!
lastly,This is my latest code, and the errors shown above are caused by this code.
`from rvc_python.infer import RVCInference
rvc = RVCInference(device="cuda:0")
rvc.load_model("klee.pth")
rvc.infer_file("2.wav", "output.wav")`
I hope this can provide some help in solving the problem.
I am just a junior high school student, and there may be some areas that I am not particularly clear about. There may be some simple questions that you think are easy, please understand me,thanks a lot!!!