Final Deployment Fix: Metadata & Docker Speedup

Dockerfile

@@ -27,9 +27,10 @@ RUN mkdir -p /app/models && chmod 777 /app/models
 COPY requirements.txt .
 
 # Install dependencies
-# Note: We enforce a specific llama-cpp-python install for CPU (OpenBLAS is good, but simple build works too)
+# We use a PRE-BUILT wheel for llama-cpp-python to avoid compiling from source (Saves ~10 minutes)
 RUN pip install --no-cache-dir --upgrade pip && \
-    pip install --no-cache-dir -r requirements.txt
+    pip install --no-cache-dir -r requirements.txt && \
+    pip install llama-cpp-python --extra-index-url https://abetlen.github.io/llama-cpp-python/whl/cpu
 
 # Copy the rest of the application
 COPY . .

requirements.txt

@@ -5,5 +5,4 @@ cython
 
 numpy
 setuptools
-llama-cpp-python
 huggingface-hub