Enable GPU support with T4 hardware and CUDA base image

Dockerfile

@@ -1,7 +1,10 @@
-FROM python:3.11-slim
+FROM nvidia/cuda:11.8.0-cudnn8-runtime-ubuntu22.04
 
-# Install system dependencies
+# Install Python and system dependencies
 RUN apt-get update && apt-get install -y \
+    python3.11 \
+    python3-pip \
+    python3.11-venv \
     nodejs \
     npm \
     git \
@@ -26,7 +29,7 @@ COPY --chown=user requirements_embedding.txt .
 COPY --chown=user hono-proxy ./hono-proxy
 COPY --chown=user vespa-certs ./vespa-certs
 
-# Install Python dependencies
+# Install Python dependencies with GPU support
 RUN pip install --user --no-cache-dir -r requirements_embedding.txt
 
 # Install Node dependencies
@@ -40,17 +43,19 @@ RUN mkdir -p $HOME/.vespa/il-infra.colpali-server.default && \
 # Create startup script
 WORKDIR $HOME/app
 RUN echo '#!/bin/bash' > start.sh && \
-    echo 'echo "Starting services..."' >> start.sh && \
-    echo 'echo "Python version: $(python --version)"' >> start.sh && \
+    echo 'echo "Starting services with GPU support..."' >> start.sh && \
+    echo 'echo "Python version: $(python3 --version)"' >> start.sh && \
     echo 'echo "Node version: $(node --version)"' >> start.sh && \
+    echo 'echo "CUDA available: $(python3 -c "import torch; print(torch.cuda.is_available())")"' >> start.sh && \
+    echo 'echo "GPU device: $(python3 -c "import torch; print(torch.cuda.get_device_name(0) if torch.cuda.is_available() else \"No GPU\")")"' >> start.sh && \
     echo '' >> start.sh && \
-    echo '# Start embedding API' >> start.sh && \
-    echo 'python embedding_api.py &' >> start.sh && \
+    echo '# Start embedding API with GPU' >> start.sh && \
+    echo 'CUDA_VISIBLE_DEVICES=0 python3 embedding_api.py &' >> start.sh && \
     echo 'EMBED_PID=$!' >> start.sh && \
     echo 'echo "Started embedding API with PID: $EMBED_PID"' >> start.sh && \
     echo '' >> start.sh && \
     echo '# Wait for embedding API' >> start.sh && \
-    echo 'sleep 5' >> start.sh && \
+    echo 'sleep 15' >> start.sh && \
     echo '' >> start.sh && \
     echo '# Check if embedding API is running' >> start.sh && \
     echo 'if curl -f http://localhost:8001/health; then' >> start.sh && \

README.md

@@ -6,6 +6,7 @@ colorTo: purple
 sdk: docker
 pinned: false
 license: apache-2.0
+suggested_hardware: t4-small
 ---
 
 # ColPali Backend API