Fix: use CUDA 12.6 image for libcudart.so.13 compatibility

Dockerfile

@@ -1,8 +1,7 @@
 # ClarityGuard - HuggingFace Spaces L4 GPU
 # llama-server precompiled locally and uploaded to the repo as a binary.
-# No compile step in HF Spaces, avoiding CUDA build OOM/timeouts.
 
-FROM nvidia/cuda:12.4.1-runtime-ubuntu22.04
+FROM nvidia/cuda:12.6.3-runtime-ubuntu22.04
 
 ENV DEBIAN_FRONTEND=noninteractive
 ENV PYTHONUNBUFFERED=1
@@ -22,10 +21,8 @@ RUN apt-get update && apt-get install -y \
     python3 python3-pip \
     git git-lfs curl \
     libgomp1 \
-    libcublas-12-4 \
     && rm -rf /var/lib/apt/lists/*
 
-# Copy precompiled llama-server and its shared libraries from the repo.
 COPY bin/llama-server /opt/llama-cpp/llama-server
 COPY bin/*.so* /usr/local/lib/
 RUN chmod +x /opt/llama-cpp/llama-server && ldconfig