Support RTX 50XX series (sm_120)
Hi,
I get the following error when running the gpu docker container:
/usr/local/lib/python3.10/site-packages/torch/cuda/__init__.py:287: UserWarning:
NVIDIA GeForce RTX 5060 Ti with CUDA capability sm_120 is not compatible with the current PyTorch installation.
The current PyTorch install supports CUDA capabilities sm_50 sm_60 sm_70 sm_75 sm_80 sm_86 sm_90.
If you want to use the NVIDIA GeForce RTX 5060 Ti GPU with PyTorch, please check the instructions at https://pytorch.org/get-started/locally/
Can you provide a new docker container working with the new RTX 5000 Series?
@EPNW-Eric did you ever figure it out?
TBO I'm not sure, since I tested a lot of different whisper variants back then and I can't remember exactly which worked and which not.
But it seems I wrote a custom Dockerfile with a higher CUDA version. You can try it, but no guarantee that it actually works:
FROM nvidia/cuda:13.0.0-cudnn-devel-ubuntu24.04
RUN apt-get update && \
apt-get install -y --no-install-recommends \
portaudio19-dev \
python3.12 \
python3.12-dev \
python3-pip \
python3-venv \
ffmpeg \
libhdf5-dev \
git \
git-lfs \
build-essential \
nano \
wget \
&& rm -rf /var/lib/apt/lists/*
RUN python3 -m venv /opt/venv
ENV PATH="/opt/venv/bin:$PATH"
RUN pip install --upgrade pip
RUN pip install -U "huggingface_hub[cli]" && hf download TheTobyB/whisper-large-v3-turbo-german-ct2
RUN pip install whisper-live==0.7.1
WORKDIR /app
RUN wget https://raw.githubusercontent.com/collabora/WhisperLive/4b46371dac00e51f0cd9962c77b0a5e0daf037ad/run_server.py
CMD ["python", "run_server.py"]
Also note that I seem to pre-download a specific german model (TheTobyB/whisper-large-v3-turbo-german-ct2), I think you can omit that line if you don't need it.
If you decide to try the Dockerfile, can you give me feedback if it works 😅 ?
@EPNW-Eric thanks for the heads up. thats helpful. i was curious before i bought but guess i will jump in! appreciate the feedback