Update Dockerfile

Dockerfile

@@ -1,16 +1,32 @@
-# Use the official lightweight CPU-only image from Docling
-FROM quay.io/docling-project/docling-serve-cpu:latest
+# Use a Python base with high compatibility
+FROM python:3.10-slim
 
-# Set environment variables for Hugging Face Spaces
-ENV HOST=0.0.0.0
-ENV PORT=7860
-ENV DOCLING_SERVE_ENABLE_UI=1
+# Install system dependencies for OCR and Image processing
+RUN apt-get update && apt-get install -y \
+    libgl1-mesa-glx \
+    libglib2.0-0 \
+    && rm -rf /var/lib/apt/lists/*
 
-# Pre-download the SmolDocling model to speed up startup (optional but recommended)
-# This ensures the 256M model is ready in the container cache
-RUN python3 -c "from docling.document_converter import DocumentConverter; DocumentConverter()"
+WORKDIR /app
+
+# Install GLM-OCR and required inference engines
+RUN pip install --no-cache-dir \
+    torch --index-url https://download.pytorch.org/whl/cpu \
+    transformers \
+    accelerate \
+    einops \
+    pillow \
+    fastapi \
+    uvicorn \
+    python-multipart
+
+# Script to pre-load the model (0.9B is ~1.8GB on disk)
+RUN python3 -c "from transformers import AutoModelForCausalLM, AutoTokenizer; \
+    AutoModelForCausalLM.from_pretrained('THUDM/glm-4v-9b', trust_remote_code=True)"
+
+# Create a simple FastAPI app wrapper
+COPY app.py .
 
 EXPOSE 7860
 
-# Start the server
-CMD ["docling-serve", "run", "--host", "0.0.0.0", "--port", "7860"]
+CMD ["uvicorn", "app:app", "--host", "0.0.0.0", "--port", "7860"]