DeepSeek-R1 / DeepSeek Reasoning Model (deepseek-reasoner) with the AI Automation Suggester integration for Home Assistant?

[Hedda]

Could it be possible and a good idea to try to get DeepSeek-R1 / DeepSeek Reasoning Model (deepseek-reasoner) with the AI Automation Suggester integration for Home Assistant?

The benefit would be that more people could be able to run smaller variants of DeepSeek R1 distilled models on-premis with their own hardware at home instead of having to use LLM cloud services and still get acceptable results for this purpose.

Spotted several people running distilled model of Deepseek R-1 LLM on the Nvidia Jetson Orin Nano SBC with only 16GB unified memory:

https://www.youtube.com/watch?v=e-EG3B5Uj78

https://www.youtube.com/watch?v=YnoeWXh3-q0

https://www.nextbigfuture.com/2025/01/tiny-deepseek-1-5b-models-run-on-99-nvidia-jetson-nano.html

https://blogs.nvidia.com/blog/deepseek-r1-nim-microservice/

If setup as a local local Ollama server then can use this from Home Assistant's Assist and other applications via the Ollama integration:

https://www.home-assistant.io/integrations/ollama/

https://www.home-assistant.io/voice_control/assist_create_open_ai_personality/