FileNotFoundError: libtorchaudio.pyd' (or one of its dependencies)

[magrega]

Hi! I am having a trouble launching the GUI. I've tried it on two different Win10 PCs with NVIDIA cards. One is RTX 3060 and the other is GTX 1050 but both come to the same error.

I changed line 83 in bark_infinity_installer_webui.py from "pip install torchaudio==2.0.2+cu118 torch==2.0.1+cu118 torchvision --extra-index-url https://download.pytorch.org/whl/cu118", to "pip install torchaudio==2.1.2+cu118 torch==2.1.2+cu118 torchvision --extra-index-url https://download.pytorch.org/whl/cu118" as was mentioned in one of the issues but it didn't help.

Running command: "C:\Users\proshanovae\Downloads\bark-infinity-v0.23\installer_files\conda\condabin\conda.bat" activate "C:\Users\proshanovae\Downloads\bark-infinity-v0.23\installer_files\env" >nul && conda
Running command: "C:\Users\proshanovae\Downloads\bark-infinity-v0.23\installer_files\conda\condabin\conda.bat" activate "C:\Users\proshanovae\Downloads\bark-infinity-v0.23\installer_files\env" >nul && python bark_webui.py
Traceback (most recent call last):
  File "C:\Users\proshanovae\Downloads\bark-infinity-v0.23\bark\bark_webui.py", line 14, in <module>
    from bark_infinity import config
  File "C:\Users\proshanovae\Downloads\bark-infinity-v0.23\bark\bark_infinity\__init__.py", line 1, in <module>
    from .api import generate_audio, text_to_semantic, semantic_to_waveform, save_as_prompt
  File "C:\Users\proshanovae\Downloads\bark-infinity-v0.23\bark\bark_infinity\api.py", line 12, in <module>
    from .generation import get_SUNO_USE_DIRECTML
  File "C:\Users\proshanovae\Downloads\bark-infinity-v0.23\bark\bark_infinity\generation.py", line 7, in <module>
    from encodec import EncodecModel
  File "C:\Users\proshanovae\Downloads\bark-infinity-v0.23\installer_files\env\lib\site-packages\encodec\__init__.py", line 12, in <module>
    from .model import EncodecModel
  File "C:\Users\proshanovae\Downloads\bark-infinity-v0.23\installer_files\env\lib\site-packages\encodec\model.py", line 19, in <module>
    from .utils import _check_checksum, _linear_overlap_add, _get_checkpoint_url
  File "C:\Users\proshanovae\Downloads\bark-infinity-v0.23\installer_files\env\lib\site-packages\encodec\utils.py", line 14, in <module>
    import torchaudio
  File "C:\Users\proshanovae\Downloads\bark-infinity-v0.23\installer_files\env\lib\site-packages\torchaudio\__init__.py", line 1, in <module>
    from . import (  # noqa: F401
  File "C:\Users\proshanovae\Downloads\bark-infinity-v0.23\installer_files\env\lib\site-packages\torchaudio\_extension\__init__.py", line 45, in <module>
    _load_lib("libtorchaudio")
  File "C:\Users\proshanovae\Downloads\bark-infinity-v0.23\installer_files\env\lib\site-packages\torchaudio\_extension\utils.py", line 64, in _load_lib
    torch.ops.load_library(path)
  File "C:\Users\proshanovae\Downloads\bark-infinity-v0.23\installer_files\env\lib\site-packages\torch\_ops.py", line 1478, in load_library
    ctypes.CDLL(path)
  File "C:\Users\proshanovae\Downloads\bark-infinity-v0.23\installer_files\env\lib\ctypes\__init__.py", line 374, in __init__
    self._handle = _dlopen(self._name, mode)
FileNotFoundError: Could not find module 'C:\Users\proshanovae\Downloads\bark-infinity-v0.23\installer_files\env\Lib\site-packages\torchaudio\lib\libtorchaudio.pyd' (or one of its dependencies). Try using the full path with constructor syntax.
If there is an error, take note of it here.

[magrega]

On reinstall I get this

ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts.
torchaudio 2.0.2+cu118 requires torch==2.0.1, but you have torch 2.8.0 which is incompatible.
torchvision 0.15.2+cu118 requires torch==2.0.1, but you have torch 2.8.0 which is incompatible.

[magrega]

Fixed it by activating env manually, uninstalling torch related packages pip uninstall torch torchvision torchaudio and then installing the downgraded version of torch pip install torch==2.0.1 torchvision==0.15.2 torchaudio==2.0.2 --index-url https://download.pytorch.org/whl/cu118

[madwurmz]

Fixed it by activating env manually, uninstalling torch related packages pip uninstall torch torchvision torchaudio and then installing the downgraded version of torch pip install torch==2.0.1 torchvision==0.15.2 torchaudio==2.0.2 --index-url https://download.pytorch.org/whl/cu118

yes this is a pretty old repo, fun to see you use it, but do we really expect these experimental projects to have support for years after release :O

[magrega]

Well, I've been looking for a local ai to clone voices in the russian language. Bark is the only thing I could find

[madwurmz]

Well, I've been looking for a local ai to clone voices in the russian language. Bark is the only thing I could find

You should check out VibeVoice, local tts , does training and supports foreign languages . Previously I used RVC that's also local and able to do foreign voice training. The training in Bark is actually just a gimmick, that will hallucinate a lot.

[magrega]

You should check out VibeVoice, local tts , does training and supports foreign languages . Previously I used RVC that's also local and able to do foreign voice training.

Could you give some pointers where should I start with VibeVoice? I found an instance in LM Studio. Will I be able to train it to speak in Russian?

[madwurmz]

You should check out VibeVoice, local tts , does training and supports foreign languages . Previously I used RVC that's also local and able to do foreign voice training.

Could you give some pointers where should I start with VibeVoice? I found an instance in LM Studio. Will I be able to train it to speak in Russian?

I just got this info, but haven’t tested it myself yet. Check out this video: https://youtu.be/XtaPZmlyMMw?si=N-MWcl2vIOHmg1GD The training result are perfect, and he mentions it supports multiple languages