Update Dockerfile

Dockerfile

@@ -2,23 +2,24 @@ FROM python:3.10-slim
 
 WORKDIR /app
 
-# Install system dependencies
+# Install system dependencies (minimal for pre-compiled wheels)
 RUN apt-get update && apt-get install -y \
-    build-essential \
-    cmake \
     wget \
     curl \
     && rm -rf /var/lib/apt/lists/*
 
-# Copy requirements and install Python dependencies
+# Copy requirements first for better caching
 COPY requirements.txt .
-RUN pip install --no-cache-dir -r requirements.txt
 
-# Download the model file
-RUN echo "📥 Downloading CapybaraHermes model (5GB)..." && \
-    wget -O capybarahermes-2.5-mistral-7b.Q5_K_M.gguf \
+# Install Python dependencies with pre-compiled llama-cpp-python
+RUN pip install --no-cache-dir --upgrade pip && \
+    pip install --no-cache-dir -r requirements.txt
+
+# Download model file (this is the time-consuming part, but should work)
+RUN echo "📥 Downloading CapybaraHermes model..." && \
+    wget --progress=bar:force:noscroll -O capybarahermes-2.5-mistral-7b.Q5_K_M.gguf \
     https://huggingface.co/TheBloke/CapybaraHermes-2.5-Mistral-7B-GGUF/resolve/main/capybarahermes-2.5-mistral-7b.Q5_K_M.gguf && \
-    echo "✅ Model downloaded successfully ($(du -h capybarahermes-2.5-mistral-7b.Q5_K_M.gguf | cut -f1))"
+    echo "✅ Model downloaded: $(du -h capybarahermes-2.5-mistral-7b.Q5_K_M.gguf | cut -f1)"
 
 # Copy application files
 COPY api.py .
@@ -31,5 +32,5 @@ EXPOSE 7860
 HEALTHCHECK --interval=30s --timeout=30s --start-period=300s --retries=3 \
   CMD curl --fail http://localhost:7860/_stcore/health || exit 1
 
-# Run Streamlit on port 7860 (HF Spaces requirement)
+# Run Streamlit on port 7860
 CMD ["streamlit", "run", "app.py", "--server.port=7860", "--server.address=0.0.0.0", "--server.headless=true"]