Update Dockerfile

Dockerfile

@@ -1,18 +1,42 @@
-FROM python:3.11-slim
-
-WORKDIR /app
+FROM ubuntu:22.04
 
 RUN apt-get update && \
-    apt-get install -y --no-install-recommends gcc g++ cmake && \
-    pip install --no-cache-dir llama-cpp-python gradio huggingface-hub && \
-    apt-get purge -y gcc g++ cmake && \
-    apt-get autoremove -y && \
+    apt-get install -y \
+    build-essential \
+    libssl-dev \
+    zlib1g-dev \
+    libopenblas-dev \
+    libomp-dev \
+    cmake \
+    pkg-config \
+    git \
+    python3-pip \
+    curl \
+    libcurl4-openssl-dev \
+    wget && \
     rm -rf /var/lib/apt/lists/*
 
-COPY app.py .
+RUN pip3 install --upgrade pip && \
+    pip3 install openai fastapi uvicorn pydantic orjson httptools
+
+RUN pip install httpx[http2]
+
+RUN git clone https://github.com/ggerganov/llama.cpp && \
+    cd llama.cpp && \
+    cmake -B build -S . \
+        -DLLAMA_BUILD_SERVER=ON \
+        -DGGML_BLAS=ON \
+        -DGGML_BLAS_VENDOR=OpenBLAS \
+        -DCMAKE_BUILD_TYPE=Release && \
+    cmake --build build --config Release --target llama-server -j $(nproc)
 
-RUN useradd -m -u 1000 user && chown -R user:user /app
-USER user
+RUN mkdir -p /models && \
+    wget -O /models/model.gguf https://huggingface.co/unsloth/Qwen3.5-0.8B-GGUF/resolve/main/Qwen3.5-0.8B-Q8_0.gguf
+
+COPY app.py /app.py
+COPY start.sh /start.sh
+RUN chmod +x /start.sh
 
 EXPOSE 7860
-CMD ["python", "app.py"]
+
+CMD ["/start.sh"]