Dockerfile.v2

FROM ubuntu:22.04

# Install basic dependencies
RUN apt-get update && \
    apt-get install -y ca-certificates curl build-essential git && \
    apt-get clean && \
    rm -rf /var/lib/apt/lists/*

# Install CUDA toolkit
RUN curl -fsSL https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2204/x86_64/cuda-keyring_1.0-1_all.deb -o cuda-keyring.deb && \
    dpkg -i cuda-keyring.deb && \
    apt-get update && \
    apt-get install -y cuda-toolkit-12-4 && \
    apt-get clean && \
    rm -rf /var/lib/apt/lists/*

# Install Go
ARG GOVERSION=1.23.4
RUN curl -fsSL https://golang.org/dl/go${GOVERSION}.linux-amd64.tar.gz | tar xz -C /usr/local
ENV PATH=/usr/local/go/bin:/usr/local/cuda/bin:$PATH

# Set up build environment
ENV CGO_ENABLED=1
ENV GOFLAGS="'-ldflags=-w -s'"

# Build Ollama
WORKDIR /go/src/github.com/ollama/ollama
COPY . .
RUN go build -trimpath -buildmode=pie -o /bin/ollama .

# Configure the runtime environment
ENV LD_LIBRARY_PATH=/usr/local/nvidia/lib:/usr/local/nvidia/lib64
ENV NVIDIA_DRIVER_CAPABILITIES=compute,utility
ENV NVIDIA_VISIBLE_DEVICES=all
ENV OLLAMA_HOST=0.0.0.0:11434
EXPOSE 11434
ENTRYPOINT ["/bin/ollama"]
CMD ["serve"]