Update Dockerfile

Dockerfile

@@ -1,24 +1,13 @@
 FROM ollama/ollama:latest
 
-# HF Spaces expect the server on port 7860 (free CPU tier)
+# HF Spaces (free CPU) expects port 7860
 ENV OLLAMA_HOST=0.0.0.0:7860
 ENV OLLAMA_MAX_LOADED_MODELS=1
 
 EXPOSE 7860
 
-# Start Ollama, pull Qwen3 8B quantized, then keep running
-CMD ["bash", "-c", "\
-    echo '🚀 Starting Ollama API server (Qwen3 8B - free CPU)...' && \
-    ollama serve & \
-    sleep 10 && \
-    echo '📥 Pulling Qwen3 8B (q4_K_M quantized)...' && \
-    ollama pull qwen3:8b-q4_K_M && \
-    echo '✅ Qwen3 8B is ready! API available at https://your-space.hf.space' && \
-    echo '   Model name: qwen3:8b-q4_K_M' && \
-    tail -f /dev/null \
-"]
+# Override default entrypoint so we can run normal commands
+ENTRYPOINT []
 
-# Optional: change the model here if you want another quantization
-# Examples:
-#   qwen3:8b-q5_K_M     (better quality, slightly more RAM)
-#   qwen3:8b-q4_0       (lighter)
+# Start Ollama server + pull Qwen3 8B (quantized for CPU) + keep alive
+CMD ollama serve & sleep 15 && ollama pull qwen3:8b-q4_K_M && echo "✅ Qwen3 8B ready!" && tail -f /dev/null