Update Dockerfile

Dockerfile

@@ -1,52 +1,33 @@
 FROM nvidia/cuda:12.0.0-cudnn8-devel-ubuntu22.04
 
-ENV MODEL_NAME="Phind-CodeLlama-34B-v2-GGUF"
-ENV DEFAULT_MODEL_FILE="phind-codellama-34b-v2.Q2_K.gguf"
-ENV MODEL_USER="TheBloke"
-ENV DEFAULT_MODEL_BRANCH="main"
-ENV MODEL_URL="https://huggingface.co/${MODEL_USER}/${MODEL_NAME}/resolve/${DEFAULT_MODEL_BRANCH}/${DEFAULT_MODEL_FILE}"
-ENV PATH="/usr/local/cuda/bin:$PATH"
-
+# Install dependencies
 RUN apt update && \
-    apt install --no-install-recommends -y build-essential python3 python3-pip wget curl git && \
+    apt install --no-install-recommends -y build-essential python3 python3-pip wget curl git cmake && \
     apt clean && rm -rf /var/lib/apt/lists/*
 
 WORKDIR /app
 
-RUN apt-get install -y wget && \
-    wget -qO- "https://cmake.org/files/v3.18/cmake-3.18.0-Linux-x86_64.tar.gz" | tar --strip-components=1 -xz -C /usr/local
-
-RUN CMAKE_ARGS="-DLLAMA_CUBLAS=on" FORCE_CMAKE=1 pip install git+https://github.com/abetlen/llama-cpp-python --no-cache-dir
-
-COPY requirements.txt ./
-
-RUN pip3 install --upgrade pip && \
-    pip3 install -r requirements.txt
-
-WORKDIR /app
+# Download ggml and mmproj models from HuggingFace
+RUN wget https://huggingface.co/mys/ggml_llava-v1.5-13b/raw/main/ggml-model-q4_k.gguf -O ggml-model-q4_k.gguf && \
+    wget https://huggingface.co/mys/ggml_llava-v1.5-13b/raw/main/mmproj-model-f16.gguf -O mmproj-model-f16.gguf
 
-RUN echo ${MODEL_URL} && \
-    wget -O /app/${DEFAULT_MODEL_FILE} ${MODEL_URL}
+# Clone and build llava-server
+RUN git clone https://github.com/trzy/llava-cpp-server.git llava && \
+    cd llava && \
+    git submodule init && \
+    git submodule update && \
+    make
 
+# Create a non-root user for security reasons
 RUN useradd -m -u 1000 user
 
-RUN mkdir -p /home/user/app && \
-    mv /app/${DEFAULT_MODEL_FILE} /home/user/app
-
-RUN chown -R user:user /home/user/app
-
 USER user
-ENV HOME=/home/user \
-	PATH=/home/user/.local/bin:$PATH \
-    MODEL_NAME=${MODEL_NAME} \
-    MODEL_FILE=/home/user/app/${DEFAULT_MODEL_FILE} 
+ENV HOME=/home/user
 
 WORKDIR $HOME/app
 
-COPY --chown=user . .
-
-RUN ls -al
-
-EXPOSE 8000
+# Expose the port
+EXPOSE 8080
 
-CMD ["uvicorn", "main:app", "--host", "0.0.0.0", "--port", "8000"]
+# Start the llava-server with models
+CMD ["./llava/bin/llava-server", "-m", "./ggml-model-q4_k.gguf", "--mmproj", "./mmproj-model-f16.gguf", "--host", "0.0.0.0", "--port", "8080"]