search

erew123 / alltalk_tts

AllTalk is based on the Coqui TTS engine, similar to the Coqui_tts extension for Text generation webUI, however supports a variety of advanced features, such as a settings page, low VRAM support, DeepSpeed, narrator, model finetuning, custom models, wav file maintenance. It can also be used with 3rd Party software via JSON calls.
GNU Affero General Public License v3.0
816 stars 91 forks source link

Impossible to install CUDA Toolkit - Docker #148

Closed Fgabz closed 5 months ago

Fgabz commented 5 months ago

Describe the bug

Hello, I'm having issue installing CUDA 11.8 Toolkit. I'm using the docker image provided on the repo and runing it on runpod (I've tried other provider same problem). Each time I'm trying to install the the toolkit, I'm not able to install the nvidia driver 550. Does anyone got also this issue?

erew123 commented 5 months ago

Hi @Fgabz

I'm assuming you are installing it for Finetuning?

I don't know Docker too well (someone else built that docker), however you may be able to run this pip install nvidia-cublas-cu11 and then maybe restart the docker environment and it may work. If it does, there will be no need to install the CUDA 11.8 Toolkit.

I'm going to be releasing a new update sometime soon and with that, attempt to get a new Docker built.

Thanks

Fgabz commented 5 months ago

Yes the goal is for finetuning. I'm going to try this, I'll let you know.

erew123 commented 5 months ago

Hi @Fgabz

Correction, these are needed:

pip install nvidia-cublas-cu11 nvidia-cudnn-cu11

export LD_LIBRARY_PATH=`python3 -c 'import os; import nvidia.cublas.lib; import nvidia.cudnn.lib; print(os.path.dirname(nvidia.cublas.lib.__file__) + ":" + os.path.dirname(nvidia.cudnn.lib.__file__))'
Fgabz commented 5 months ago

All good @erew123 now it works ! thanks !