- Qwen2.5:1.5B on Ollama
- Auto tool execution via <|tool|> blocks
- Tools: lookup_order, cancel_order, check_refund, update_address (India), track_shipment
- Memory (data/memory.json) and logs (logs/chatml.log)
- Demos: test_bot.py, test_tools.py
- Windows/PowerShell launchers + Dockerfile
pip install -r requirements.txt ollama serve ollama pull qwen2.5:1.5b uvicorn app:app --reload --port 8080 python test_bot.py
docker build -t support-bot:3.4 . docker run --rm -p 8080:8080 -e LLM_API=http://host.docker.internal:11434/api/chat -e LLM_MODEL=qwen2.5:1.5b support-bot:3.4
Generated: 2025-11-10T04:08:17.855928Z