Update Dockerfile

Dockerfile

@@ -1,24 +1,35 @@
-# Start with the official full image
 FROM ghcr.io/ggml-org/llama.cpp:full
-
-# Set up the working directory
 WORKDIR /app
 
-# 1. Install Python and Pip (The base image is often minimal Ubuntu)
-RUN apt-get update && \
-    apt-get install -y python3 python3-pip git && \
-    rm -rf /var/lib/apt/lists/*
-
-# 2. Copy your requirements and install Python libraries
-COPY requirements.txt .
-RUN pip3 install --no-cache-dir -r requirements.txt
+# Install Python dependencies for downloading models
+RUN apt update && apt install -y python3-pip
+RUN pip install -U huggingface_hub
 
-# 3. Copy your application code
-COPY . .
+# --- MODEL DOWNLOAD SECTION ---
+# 1. Download Qwen 2.5 Coder 14B (Q4_K_M)
+# We rename it to 'model.gguf' to make the CMD section easier to manage.
+RUN python3 -c 'from huggingface_hub import hf_hub_download; \
+    repo="Qwen/Qwen2.5-Coder-14B-Instruct-GGUF"; \
+    filename="qwen2.5-coder-14b-instruct-q4_k_m.gguf"; \
+    hf_hub_download(repo_id=repo, filename=filename, local_dir="/app")' && \
+    mv /app/qwen2.5-coder-14b-instruct-q4_k_m.gguf /app/model.gguf
 
-# 4. Create a folder for models to live in
-RUN mkdir -p /app/models
+# --- OPTIONAL VISION MODEL (Comment out the section above and uncomment this to switch) ---
+# RUN python3 -c 'from huggingface_hub import hf_hub_download; \
+#    repo="Qwen/Qwen2.5-VL-7B-Instruct-GGUF"; \
+#    filename="qwen2.5-vl-7b-instruct-q4_k_m.gguf"; \
+#    hf_hub_download(repo_id=repo, filename=filename, local_dir="/app")' && \
+#    mv /app/qwen2.5-vl-7b-instruct-q4_k_m.gguf /app/model.gguf
 
-# 5. IMPORTANT: Override the base image's entrypoint. 
-# The base image defaults to running 'llama-server', but we want to run 'app.py'
-ENTRYPOINT ["python3", "app.py"]
+# --- SERVER STARTUP ---
+# We use "--n-gpu-layers 0" for CPU Spaces. 
+# We use "-c 8192" because 14B takes a lot of RAM; 32k context might crash the space.
+CMD ["--server", \
+     "-m", "/app/model.gguf", \
+     "--host", "0.0.0.0", \
+     "--port", "7860", \
+     "-t", "4", \
+     "--cache-type-k", "q8_0", \
+     "--cache-type-v", "iq4_nl", \
+     "-c", "8192", \
+     "--n-gpu-layers", "0"]