Open MarythaV2 opened 1 year ago
i would try reinstalling with followin method (windows) use powershell first check wich python is available py -0 my system example
py -0
Installed Pythons found by C:\Windows\py.exe Launcher for Windows
-3.10-64 *
-3.9-64
-3.7-64
-3.6-64
i would suggest python 3.10.x create a fresh directory
mkdir bark-withclone
cd bark-withclone
py -3.10 -m venv venv
.\venv\Scripts\Activate.ps1
it should look like
(venv) PS G:\git\ai-audio\bark-withclone>
then first instal torch with
pip3 install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu117
pip3 install notebook
pip install soundfile
pip install git+https://github.com/serp-ai/bark-with-voice-clone.git
run the following tests with ipython:
import torch
torch.cuda.is_available()
output should be True
import bark
from bark import SAMPLE_RATE, generate_audio
should return without error mkdir notebooks copy your notebook to this directory cd notebooks jupyter notebook open your notebook and try again
this is how i got it up and running hope this helps
thank you, that fixed it!
I've been trying to run the clone_voice.ipynb file but I have been getting stuck at the importing of codec, but there have been some other errors after that. I have codec, but I'm getting the error message " cannot import name 'codec_encode' from 'bark.generation' ", " name 'torchaudio' is not defined ", and " name 'torch' is not defined ".