ivrit-ai-streaming / Dockerfile
AshDavid12
take out keep alive
544d523
raw
history blame contribute delete
869 Bytes
FROM nvidia/cuda:12.2.2-cudnn8-runtime-ubuntu22.04
ENV PYTHON_VERSION=3.11
RUN export DEBIAN_FRONTEND=noninteractive \
&& apt-get -qq update \
&& apt-get -qq install --no-install-recommends \
python${PYTHON_VERSION} \
python${PYTHON_VERSION}-venv \
python3-pip \
libcublas11 \
&& rm -rf /var/lib/apt/lists/*
# Set the working directory
WORKDIR /
# Create a writable cache directory for Hugging Face
RUN mkdir -p /hf_cache && chmod -R 777 /hf_cache
# Set the environment variable for the Hugging Face cache
ENV HF_HOME=/hf_cache
# Copy the requirements.txt file and install the dependencies
COPY requirements.txt .
RUN pip install --no-cache-dir -r requirements.txt
# Copy the current directory contents into the container at /app
COPY . .
# Run FastAPI with Uvicorn
CMD ["uvicorn", "infer:app", "--host", "0.0.0.0", "--port","7860"]