-
|
I just installed a K80 and I am trying to run VoxNovel but it is telling me my driver is too old and to update or to compile CUDA from source, which I did, the K80 runs on 11.4 but I see it is compatible with 11.8 I was looking at the linux requirements and it shows CUDA 12 as needed to run VoxNovel is this correct or is there any way I can get this to run with 11.8? Thanks |
Beta Was this translation helpful? Give feedback.
Replies: 8 comments 37 replies
-
It should run fine with CUDA 11.8I know this because:As in order to fix Headless VoxNovel Google Colab I HAD to downgrade the Google Colab env to CUDA 11.8 to get it to run in the first place Referenced Google Colab: https://colab.research.google.com/drive/15pp2hFBo2fD3legDQfWY5DMt-aI-HKKF?usp=sharing |
Beta Was this translation helpful? Give feedback.
-
|
I keep getting this error when starting up, not sure exactly how to proceed, I have been trying to get this to work and I am not sure what I am doing wrong |
Beta Was this translation helpful? Give feedback.
-
|
Not that I know of on the top of my head...
I guess just check if you can get XTTS running on its own, cause thats theist and default TTS backend that VoxNovel uses. line used for it after a Synthesizing speech by 🐸XTTS 🐍 Python API import torch
from TTS.api import TTS
# Get device
device = "cuda" if torch.cuda.is_available() else "cpu"
# List available 🐸TTS models
print(TTS().list_models())
# Init TTS
tts = TTS("tts_models/multilingual/multi-dataset/xtts_v2").to(device)
# Run TTS
# ❗ Since this model is multi-lingual voice cloning model, we must set the target speaker_wav and language
# Text to speech list of amplitude values as output
wav = tts.tts(text="Hello world!", speaker_wav="my/cloning/audio.wav", language="en")
# Text to speech to a file
tts.tts_to_file(text="Hello world!", speaker_wav="my/cloning/audio.wav", language="en", file_path="output.wav") |
Beta Was this translation helpful? Give feedback.
-
|
No sorry I mean it broke the epub2tts which is another program I was using for text to audiobook. I think it uses a different version of some dependency so when I install VoxNovel it replaces something. Anyway I'm pretty confident I will get it working now that I know a lot more about the way CUDA and Pytorch work. Ill keep you updated. Also I had a question, I noticed when I ran it before that most of the voices are British sounding, is there somewhere I can get samples of the voices and is there a way to port the voices from epub2tts into VoxNovel or are they already present? I haven't played around with VoxNovel enough to find the range of voices but a book with say 30 unique characters it would be nice to have enough voices to supply each one with their own. |
Beta Was this translation helpful? Give feedback.
-
voxnovel voice clone buttonYes there is a clone voice button in the VoxNovel gui actually, just give it a like 15-30 second audio sample of any person talking, give the name of the voice actor and select the gender for auto selection, and then that voice will be added to the list of available voices actors. It should walk you through the process anyway
About getting epub2tts voicesAlso epub2tts seems to also use XTTS as the TTS engine, so it might have a bunch of voice samples somewhere in its codebase so you can probs just easily grab voices from there. when selecting one of the voices from the dropdown of VoxNovel for a character
|
Beta Was this translation helpful? Give feedback.
-
|
Perfect, I will play around with it and see what I can do. Thanks for all the help! |
Beta Was this translation helpful? Give feedback.
-
|
Oh, also why not install miniconda and just set up two different conda python envs with different pips for both programs? |
Beta Was this translation helpful? Give feedback.
-
|
Testing on my end in a Google Colab right now... |
Beta Was this translation helpful? Give feedback.
-
|
I mean it work in Google Colab on my end with CUDA 11.8 
See full run details in this ipynb colab file |
Beta Was this translation helpful? Give feedback.
-
|
Update: may of given the wrong colab file here is if it was wrong |
Beta Was this translation helpful? Give feedback.
-
|
I didn’t even know that was an option, I’ll have to look into that.I’m kind of new to this whole thing but I’ve been running Linux exclusively for 17 years On Oct 1, 2024, at 16:23, Drew Thomasson ***@***.***> wrote:
Oh, also why not install miniconda and just set up two different conda python envs with different pips for both programs?
—Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you authored the thread.Message ID: ***@***.***>
|
Beta Was this translation helpful? Give feedback.
-
|
lol yeah, You might already have it installed actually from VoxNovelalso if you used the auto-installer command for VoxNovel then you probs already have conda installed, cause that should be doing that, try typing in
and it should show the the conda envs that exist to activate one "start using one" type to create one you can do something like this for instance
Example for creating a conda env named booknlp that uses python 3.7 and how to activate it:conda create --name booknlp python=3.7
conda activate booknlp
#to deactivate type:
conda deactivateonce your in a activate conda env then you could just be able to do normal pip install commands tbh and it'll make those pip installs only apply to the conda env your currently in lol Yeah, miniconda makes it a LOT easer handling python projects that use multiple python versions or have diffrent conflicting dependencies etc |
Beta Was this translation helpful? Give feedback.
-
|
Perfect, I will do this tomorrow, if I can’t figure something out I may hit you up for more advice if you don’t mind.On Oct 1, 2024, at 20:14, Drew Thomasson ***@***.***> wrote:
lol yeah,
You might already have it installed actually from VoxNovel
also if you used the auto-installer command for VoxNovel then you probs already have conda installed, cause that should be doing that,
try typing in
conda env list
and it should show the the conda envs that exist
to activate one "start using one" type conda activate envName
to create one you can do something like this for instance
conda create --name {yourCustomeEnvName} python={pythonNUMBER}
Example for creating a conda env named booknlp that uses python 3.7 and how to activate it:
conda create --name booknlp python=3.7
conda activate booknlp
#to deactivate type:
conda deactivate
—Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you authored the thread.Message ID: ***@***.***>
|
Beta Was this translation helpful? Give feedback.
-
|
Sounds like a plan 👍 |
Beta Was this translation helpful? Give feedback.
-
|
so Im running into the same issue, I installed VoxNovel thinking it would keep my version of Pytorch "2.1.0 cu118" but it uninstalled that version and installed some other version and now it will not run the GPU, its saying again UserWarning: CUDA initialization: The NVIDIA driver on your system is too old (found version 11040). Please update your GPU driver by downloading and installing a new version from the URL: http://www.nvidia.com/Download/index.aspx Alternatively, go to: https://pytorch.org to install a PyTorch version that has been compiled with your version of the CUDA driver. (Triggered internally at ../c10/cuda/CUDAFunctions.cpp:108.) when I attempt to install the proper version it gives this error I tried both PIP and Conda to install ill add both below pip install torch==2.1.0 torchvision==0.16.0 torchaudio==2.1.0 --index-url https://download.pytorch.org/whl/cu118 conda install pytorch==2.1.0 torchvision==0.16.0 torchaudio==2.1.0 pytorch-cuda=11.8 -c pytorch -c nvidia
LibMambaUnsatisfiableError: Encountered problems while solving:
Could not solve for environment specs not sure what I'm doing wrong |
Beta Was this translation helpful? Give feedback.
-
|
installed pytorch 2.4.1cu118 with the same results, here is a copy of my environment :128: RuntimeWarning: 'torch.utils.collect_env' found in sys.modules after import of package 'torch.utils', but prior to execution of 'torch.utils.collect_env'; this may result in unpredictable behaviour OS: Ubuntu 22.04.5 LTS (x86_64) Python version: 3.12.4 | packaged by Anaconda, Inc. | (main, Jun 18 2024, 15:12:24) [GCC 11.2.0] (64-bit runtime) Nvidia driver version: 470.256.02 CPU: Versions of relevant libraries: |
Beta Was this translation helpful? Give feedback.
-
|
lol, make sure to use the multiple conda envs to make it easier for you to organize and such
also when in a conda env or in general whatnot you can run the command pip freeze > requirements.txtto put all of the pips and exact versions in a requirements.txt file to later install with a pip install -r requirements.txt --no-depscommand might make it easier if you ever want to share the pip env with anyone if they need to replicate what you did, |
Beta Was this translation helpful? Give feedback.
-
|
Tell me if you ever get it working tho! |
Beta Was this translation helpful? Give feedback.
-
|
I got it to work, its using CUDA and my GPU I did what you said and made a separate conda env. then I did a manual install and resolved dependencies manually using the ubuntu_requirements.txt and resolving dependencies as they were shown in the error messages. |
Beta Was this translation helpful? Give feedback.
-
|
nice nice! |
Beta Was this translation helpful? Give feedback.
-
|
I guess if its working now and all you needed was the pip changes If you do pip freeze > requirements.txtand paste that into this thread for anyone else that runs into your issue so then they can just run a pip install --no-deps -r requirements.txtto replicate all your pips in your conda env?🥺 |
Beta Was this translation helpful? Give feedback.
-
|
Where do i find the requirements.txt file so I can paste it? |
Beta Was this translation helpful? Give feedback.
-
|
running pip freeze > requirements.txtshould just create a requirements.txt file in the local dir you ran that command in you also could give it anyname lol, doesnt have to be named requirements.txt, for instance pip freeze > CUDA11.8_k80_Ubuntu_requirments.txt |
Beta Was this translation helpful? Give feedback.
I got it to work, its using CUDA and my GPU
I did what you said and made a separate conda env.
then I did a manual install and resolved dependencies manually using the ubuntu_requirements.txt and resolving dependencies as they were shown in the error messages.