-
Notifications
You must be signed in to change notification settings - Fork 2.5k
Description
I am facing error:- /opt/nvidia/nvidia_entrypoint.sh: line 67: /usr/bin/python3: No such file or directory
Can you guys help me out please below is my dockerimage
FROM nvidia/cuda:12.4.1-devel-ubuntu22.04
ENV DEBIAN_FRONTEND=noninteractive
ENV VENV_PATH=/opt/venv
ENV PATH="$VENV_PATH/bin:$PATH"
Install system packages & create the symlink for the NVIDIA entrypoint
RUN apt-get update &&
apt-get install -y --no-install-recommends
git
python3
python3-venv
python3-dev
wget
build-essential
ca-certificates &&
ln -sf /usr/bin/python3.10 /usr/bin/python3 &&
rm -rf /var/lib/apt/lists/*
Setup Virtual Env
RUN python3 -m venv $VENV_PATH && pip install --upgrade pip
WORKDIR /workspace
Clone repo
RUN git clone https://github.com/**/Almora-v3
WORKDIR /workspace/Almora-v3
1. Install specific CUDA-optimized Torch first
RUN pip install torch==2.5.1 torchvision==0.20.1 torchaudio==2.5.1 --index-url https://download.pytorch.org/whl/cu124
2. Install Runpod and other requirements
RUN pip install runpod packaging ninja &&
pip install -r requirements.txt -i https://mirrors.aliyun.com/pypi/simple/
3. Install Flash Attention (must be after torch/ninja)
RUN pip install flash-attn --no-build-isolation
RUN mkdir -p /workspace/Almora-v3/input_cache
COPY handler.py /workspace/Almora-v3/handler.py
CMD ["/usr/bin/python3", "-u", "handler.py"]