Copy llama.cpp shared libraries into runtime image

Dockerfile

@@ -34,7 +34,8 @@ ENV DEBIAN_FRONTEND=noninteractive \
     HF_HUB_DISABLE_TELEMETRY=1 \
     HF_HOME=/data/hf-cache \
     MODEL_DIR=/data/models \
-    MODEL_PATH=/data/models/model.gguf
+  MODEL_PATH=/data/models/model.gguf \
+  LD_LIBRARY_PATH=/usr/local/lib:/usr/local/lib/llama.cpp
 ENV PATH=/opt/venv/bin:$PATH
 
 RUN apt-get update && apt-get install -y --no-install-recommends \
@@ -51,6 +52,8 @@ RUN groupadd -g 65532 appuser || true \
   && useradd -u 65532 -g 65532 -M -s /usr/sbin/nologin appuser || true
 
 COPY --from=builder /opt/llama.cpp/build/bin/llama-server /usr/local/bin/llama-server
+COPY --from=builder /opt/llama.cpp/build/bin/lib*.so* /usr/local/lib/
+RUN ldconfig
 
 WORKDIR /app
 COPY requirements.txt /app/requirements.txt