llama.cpp-server / Dockerfile
gingdev's picture
Update Dockerfile
e7cb0ac verified
raw
history blame contribute delete
676 Bytes
FROM ghcr.io/ggerganov/llama.cpp:server
RUN apt update && \
apt install -y wget
ENV HOME=/home/user
# https://huggingface.co/docs/hub/spaces-sdks-docker-first-demo
# Create user then move server to user's home directory. Finally, we create a models folder.
RUN useradd -m -u 1000 user && \
mv /server $HOME && \
mkdir $HOME/models && \
chown -R user:user $HOME
USER user
WORKDIR $HOME
RUN wget https://huggingface.co/gingdev/llama7b-ictu/resolve/main/llama7b_q4_k_m.gguf?download=true -O ictu.gguf
ENTRYPOINT [ "./server" ]
CMD [ "--host", "0.0.0.0", "--port", "7860", "--model", "ictu.gguf", "-c", "4096", "--chat-template", "chatml", "--embedding" ]