Update Dockerfile

Dockerfile

@@ -1,7 +1,7 @@
-# Use the standard Python 3.11 image for better compatibility with pre-built wheels
+# Use the standard Python 3.11 image for better compatibility
 FROM python:3.11
 
-# Install system dependencies (curl is still needed for the model download)
+# Install system dependencies
 RUN apt-get update && apt-get install -y curl && rm -rf /var/lib/apt/lists/*
 
 WORKDIR /app
@@ -14,7 +14,21 @@ ENV TRANSFORMERS_CACHE=/app/.cache \
     HF_HOME=/app/.cache \
     CHROMADB_DISABLE_TELEMETRY=true
 
-# Copy and install Python requirements. Pip will now find a pre-built wheel automatically.
+# --- FIX: Auto-detect architecture and install the correct pre-built wheel ---
+# This block detects the CPU architecture and downloads the corresponding binary,
+# completely avoiding slow compilation.
+RUN ARCH=$(uname -m) && \
+    case "$ARCH" in \
+        x86_64) \
+            WHEEL_URL="https://github.com/abetlen/llama-cpp-python/releases/download/v0.2.78/llama_cpp_python-0.2.78-cp311-cp311-manylinux_2_28_x86_64.whl" ;; \
+        aarch64) \
+            WHEEL_URL="https://github.com/abetlen/llama-cpp-python/releases/download/v0.2.78/llama_cpp_python-0.2.78-cp311-cp311-manylinux_2_28_aarch64.whl" ;; \
+        *) \
+            echo "Unsupported architecture: $ARCH" && exit 1 ;; \
+    esac && \
+    pip install --no-cache-dir "$WHEEL_URL"
+
+# Now, install the rest of the dependencies from requirements.txt
 COPY requirements.txt .
 RUN pip install --no-cache-dir -r requirements.txt