[ISSUE for v2]:

[ParanoiaSam]

Voice Changer Version

git clone https://github.com/w-okada/voice-changer.git

Operational System

Linux Mint

GPU

AMD Radeon RX 6650 XT

CUDA Version

ROCm

Read carefully and check the options

• If you use win_cuda_torch_cuda edition, setup cuda? see here
• If you use win_cuda edition, setup cuda and cudnn? see here
• If you use mac edition, client is not launched automatically. Use chrome to open application.?
• I've tried to change the Chunk Size
• I've tried to set the Index to zero
• I've read the tutorial
• I've tried to extract to another folder (or re-extract) the .zip file

Does pre-installed model work?

No

Model Type

The 4 default voices

Issue Description

I'm trying to install the conda version of w-okada's voice-changer. These are the steps I am taking to install it on my Linux Mint os with an AMD GPU with ROCm versions of torch, tourchaudio, and torchversion.

The instuctions for the current version on the AMD website ask for python 3.12 but the instuctions for voice-changer ask for python 3.10. I've tried both and I'm still stuck.

I've had some success by asking chatbots what to do and it sometimes worked. Definitely not a stable install and it hasn't always recognized my GPU. I'm including a step-by-step log of what I've tried.

Application Screenshot

No response

Logs on console

I'm trying to install the conda version of w-okada's voice-changer. These are the steps I am taking to install it on my Linux Mint os with an AMD GPU with ROCm versions of torch, tourchaudio, and torchversion.

The instuctions for the current version on the AMD website ask for python 3.12 but the instuctions for voice-changer ask for python 3.10. I've tried both and I'm still stuck. the following is with the python 3.12 and rocm6.4.2 versions of those files.

[(https://rocm.docs.amd.com/projects/radeon/en/latest/docs/install/wsl/install-pytorch.html)]

$ mkdir voice-changer-holder
$ cd voice-changer-holder

inside of folder now

$ wget https://repo.radeon.com/rocm/manylinux/rocm-rel-6.4.2/torch-2.6.0%2Brocm6.4.2.git76481f7c-cp312-cp312-linux_x86_64.whl
$ wget https://repo.radeon.com/rocm/manylinux/rocm-rel-6.4.2/torchvision-0.21.0%2Brocm6.4.2.git4040d51f-cp312-cp312-linux_x86_64.whl
$ wget https://repo.radeon.com/rocm/manylinux/rocm-rel-6.4.2/pytorch_triton_rocm-3.2.0%2Brocm6.4.2.git7e948ebf-cp312-cp312-linux_x86_64.whl
wget https://repo.radeon.com/rocm/manylinux/rocm-rel-6.4.2/torchaudio-2.6.0%2Brocm6.4.2.gitd8831425-cp312-cp312-linux_x86_64.whl
$ git clone https://github.com/w-okada/voice-changer.git

$ ls

pytorch_triton_rocm-3.2.0+rocm6.4.2.git7e948ebf-cp312-cp312-linux_x86_64.whl
torch-2.6.0+rocm6.4.2.git76481f7c-cp312-cp312-linux_x86_64.whl
torchaudio-2.6.0+rocm6.4.2.gitd8831425-cp312-cp312-linux_x86_64.whl
torchvision-0.21.0+rocm6.4.2.git4040d51f-cp312-cp312-linux_x86_64.whl
voice-changer/

$ conda create --name voicechanger python=3.12
$ conda activate voicechanger

in conda

$ pip3 uninstall torch torchvision pytorch-triton-rocm
$ pip3 install torch-2.6.0+rocm6.4.2.git76481f7c-cp312-cp312-linux_x86_64.whl torchvision-0.21.0+rocm6.4.2.git4040d51f-cp312-cp312-linux_x86_64.whl torchaudio-2.6.0+rocm6.4.2.gitd8831425-cp312-cp312-linux_x86_64.whl pytorch_triton_rocm-3.2.0+rocm6.4.2.git7e948ebf-cp312-cp312-linux_x86_64.whl

$ cd voice-changer/server/
$ pip install -r requirements.txt

AttributeError: module 'pkgutil' has no attribute 'ImpImporter'. Did you mean: 'zipimporter'?

I can't fix this. moving on to python 3.10

$ cd ~/voice-changer-holder/

$ conda deactivate

$ conda env remove --name voicechanger

$ git clone https://github.com/w-okada/voice-changer.git

$ wget https://repo.radeon.com/rocm/manylinux/rocm-rel-5.7/torch-2.0.1%2Brocm5.7-cp310-cp310-linux_x86_64.whl
$ wget https://repo.radeon.com/rocm/manylinux/rocm-rel-5.7/torchvision-0.15.2%2Brocm5.7-cp310-cp310-linux_x86_64.whl

$ pip3 install --force-reinstall torch-2.0.1+rocm5.7-cp310-cp310-linux_x86_64.whl torchvision-0.15.2+rocm5.7-cp310-cp310-linux_x86_64.whl

$ cd voice-changer/server/
$ pip install -r requirements.txt

$ python MMVCServerSIO.py

Traceback (most recent call last):
File "/home/sam/voice-changer-holder/voice-changer/server/MMVCServerSIO.py", line 18, in
from voice_changer.VoiceChangerManager import VoiceChangerManager
File "/home/sam/voice-changer-holder/voice-changer/server/voice_changer/VoiceChangerManager.py", line 9, in
from voice_changer.Local.ServerDevice import ServerDevice, ServerDeviceCallbacks
File "/home/sam/voice-changer-holder/voice-changer/server/voice_changer/Local/ServerDevice.py", line 9, in
from voice_changer.Local.AudioDeviceList import checkSamplingRate, list_audio_device
File "/home/sam/voice-changer-holder/voice-changer/server/voice_changer/Local/AudioDeviceList.py", line 1, in
import sounddevice as sd
File "/home/sam/anaconda3/envs/voicechanger/lib/python3.10/site-packages/sounddevice.py", line 72, in
_lib = _ffi.dlopen(_libname)
OSError: cannot load library 'libportaudio.so.2': /home/sam/anaconda3/envs/voicechanger/bin/../lib/libstdc++.so.6: version `GLIBCXX_3.4.32' not found (required by /lib/x86_64-linux-gnu/libjack.so.0)

$ conda install -c conda-forge libstdcxx-ng

$ python MMVCServerSIO.py

[Voice Changer] post_update_settings ex: No module named 'fairseq'
Traceback (most recent call last):
File "/home/sam/voice-changer-holder/voice-changer/server/restapi/MMVC_Rest_Fileuploader.py", line 71, in post_update_settings
info = self.voiceChangerManager.update_settings(key, val)
File "/home/sam/voice-changer-holder/voice-changer/server/voice_changer/VoiceChangerManager.py", line 348, in update_settings
self.generateVoiceChanger(newVal)
File "/home/sam/voice-changer-holder/voice-changer/server/voice_changer/VoiceChangerManager.py", line 256, in generateVoiceChanger
from voice_changer.RVC.RVCr2 import RVCr2
File "/home/sam/voice-changer-holder/voice-changer/server/voice_changer/RVC/RVCr2.py", line 12, in
from voice_changer.RVC.embedder.EmbedderManager import EmbedderManager
File "/home/sam/voice-changer-holder/voice-changer/server/voice_changer/RVC/embedder/EmbedderManager.py", line 5, in
from voice_changer.RVC.embedder.FairseqContentvec import FairseqContentvec
File "/home/sam/voice-changer-holder/voice-changer/server/voice_changer/RVC/embedder/FairseqContentvec.py", line 3, in
from voice_changer.RVC.embedder.FairseqHubert import FairseqHubert
File "/home/sam/voice-changer-holder/voice-changer/server/voice_changer/RVC/embedder/FairseqHubert.py", line 4, in
from fairseq import checkpoint_utils
ModuleNotFoundError: No module named 'fairseq'

$ pip install fairseq

failed to install

$ conda install conda-forge::fairseq
$ conda install conda-forge::pyworld

installed this based on the recommendation in the Closed Issues tab.
$ python MMVCServerSIO.py

/home/sam/anaconda3/envs/voicechanger/lib/python3.10/site-packages/scipy/init.py:143: UserWarning: A NumPy version >=1.19.5 and <1.27.0 is required for this version of SciPy (detected version 2.2.6)
warnings.warn(f"A NumPy version >={np_minversion} and <{np_maxversion}"
Traceback (most recent call last):
File "/home/sam/voice-changer-holder/voice-changer/server/MMVCServerSIO.py", line 18, in
from voice_changer.VoiceChangerManager import VoiceChangerManager
File "/home/sam/voice-changer-holder/voice-changer/server/voice_changer/VoiceChangerManager.py", line 9, in
from voice_changer.Local.ServerDevice import ServerDevice, ServerDeviceCallbacks
File "/home/sam/voice-changer-holder/voice-changer/server/voice_changer/Local/ServerDevice.py", line 13, in
import librosa
File "/home/sam/anaconda3/envs/voicechanger/lib/python3.10/site-packages/librosa/init.py", line 209, in
from . import core
File "/home/sam/anaconda3/envs/voicechanger/lib/python3.10/site-packages/librosa/core/init.py", line 5, in
from .convert import * # pylint: disable=wildcard-import
File "/home/sam/anaconda3/envs/voicechanger/lib/python3.10/site-packages/librosa/core/convert.py", line 7, in
from . import notation
File "/home/sam/anaconda3/envs/voicechanger/lib/python3.10/site-packages/librosa/core/notation.py", line 8, in
from ..util.exceptions import ParameterError
File "/home/sam/anaconda3/envs/voicechanger/lib/python3.10/site-packages/librosa/util/init.py", line 77, in
from .utils import * # pylint: disable=wildcard-import
File "/home/sam/anaconda3/envs/voicechanger/lib/python3.10/site-packages/librosa/util/utils.py", line 5, in
import scipy.ndimage
File "/home/sam/anaconda3/envs/voicechanger/lib/python3.10/site-packages/scipy/ndimage/init.py", line 154, in
from ._interpolation import * # noqa: F401 F403
File "/home/sam/anaconda3/envs/voicechanger/lib/python3.10/site-packages/scipy/ndimage/_interpolation.py", line 37, in
from scipy import special
File "", line 1075, in _handle_fromlist
File "/home/sam/anaconda3/envs/voicechanger/lib/python3.10/site-packages/scipy/init.py", line 200, in getattr
return _importlib.import_module(f'scipy.{name}')
File "/home/sam/anaconda3/envs/voicechanger/lib/python3.10/importlib/init.py", line 126, in import_module
return _bootstrap._gcd_import(name[level:], package, level)
File "/home/sam/anaconda3/envs/voicechanger/lib/python3.10/site-packages/scipy/special/init.py", line 663, in
from . import _ufuncs
File "scipy/special/_ufuncs.pyx", line 1, in init scipy.special._ufuncs
ValueError: numpy.dtype size changed, may indicate binary incompatibility. Expected 96 from C header, got 88 from PyObject

And I'm stuck. The Ai chatbots recommend changing pip to different versions of pip24.1 and pip24.0 but that causes a rabbit hole of issues. It makes me install 5 or 6 pip packages and it still gets stuck. Am I missing a step?

[Valenthyne]

Unfortunately, I am in the same boat with being unable to run things here, particularly with hitting a wall as far as the "fairseq" package goes. I had to follow a hodgepodge of tutorials and links to try and even get to the stage of launching the server and accessing it through the browser, only to ultimately fail at that point.

#1541 also appears to have run into a similar roadblock. Likewise, there are a good few issues in general created by Linux users who don't seem to be able to get the software working.

Put very simply, the documentation for getting this project set up and running on Linux, particularly with an AMD GPU, is in dire need of an update. I hope for and kindly request the developers to become more attentive on these matters and not neglect Linux as a target.

I have to wonder if there are specific issues getting it working on Ubuntu-based distributions.