Upload Dockerfile

Dockerfile

@@ -17,16 +17,4 @@ COPY . .
 # Expose the port for Gradio (Spaces expects the app on port 7860)
 EXPOSE 7860
 
-# Start vLLM in the background and then the Gradio app
-CMD bash -c "wget -O /tmp/tool_chat_template_llama3.1_json.jinja https://github.com/vllm-project/vllm/raw/refs/heads/main/examples/tool_chat_template_llama3.1_json.jinja && \
-    vllm.entrypoints.openai.api_server \
-    --model unsloth/llama-3-8b-Instruct-bnb-4bit \
-    --enable-auto-tool-choice \
-    --tool-call-parser llama3_json \
-    --chat-template /tmp/tool_chat_template_llama3.1_json.jinja \
-    --quantization bitsandbytes \
-    --load-format bitsandbytes \
-    --dtype half \
-    --max-model-len 8192 \
-    --download-dir models/vllm > vllm.log 2>&1 & \
-    python app.py"
+CMD ["python", "app.py"]