Currently, I am using GPT-4o on Azure, whether I can deployment open source Qwen in local side? whether need re-train for this scenario ? Thanks.