update dockerfiles & requirements - HF

Dockerfile

@@ -6,12 +6,9 @@ WORKDIR /app
 
 # 3. Install minimal system dependencies
 #    - libgl1/libglib2.0-0: OpenCV needs these
-#    - build-essential/python3-dev: llama-cpp-python builds from source
 RUN apt-get update && apt-get install -y --no-install-recommends \
     libgl1 \
     libglib2.0-0 \
-    build-essential \
-    python3-dev \
     && rm -rf /var/lib/apt/lists/*
 
 # 4. Install PyTorch CPU-only FIRST (saves ~1.5GB vs full CUDA torch)
@@ -19,20 +16,25 @@ RUN apt-get update && apt-get install -y --no-install-recommends \
 RUN pip install --no-cache-dir \
     torch --index-url https://download.pytorch.org/whl/cpu
 
-# 5. Install remaining Python dependencies
+# 5. Install llama-cpp-python from pre-built CPU wheels (avoids 30+ min C++ compile)
+RUN pip install --no-cache-dir \
+    llama-cpp-python \
+    --extra-index-url https://abetlen.github.io/llama-cpp-python/whl/cpu
+
+# 6. Install remaining Python dependencies
 COPY requirements.txt .
 RUN pip install --no-cache-dir -r requirements.txt
 
-# 6. Copy your code
+# 7. Copy your code
 COPY . .
 
-# 7. Download all ML models (Face, Text, LLaMA 3 GGUF) during build
+# 8. Download all ML models (Face, Text, LLaMA 3 GGUF) during build
 RUN python download_models.py
 
-# 8. Environment & Port settings (7860 is HF Spaces standard)
+# 9. Environment & Port settings (7860 is HF Spaces standard)
 ENV PYTHONPATH=/app
 ENV USE_EMBEDDED_LLM=True
 EXPOSE 7860
 
-# 9. Run the app with Uvicorn
+# 10. Run the app with Uvicorn
 CMD ["uvicorn", "app.main:app", "--host", "0.0.0.0", "--port", "7860"]

requirements.txt

@@ -21,9 +21,9 @@ pillow
 gdown
 
 # --- NLP (PyTorch CPU-only via --index-url, see Dockerfile) ---
+# llama-cpp-python is installed separately in Dockerfile from pre-built CPU wheels
 transformers>=4.40.0
 sentencepiece==0.1.99
-llama-cpp-python>=0.2.77
 huggingface-hub>=0.23.0
 
 # --- Utilities ---