quickq官方软件ios

deepseek-r1: incentivizing reasoning capability in llms viareinforcement learning