Submission for the Huawei Developer Competition - Open-Source LLM AI Challenge and secured the sixth position. This includes the source code for the backend.
The actual model inference is performed on the Huawei Cloud ModelArts service. This backend is primarily responsible for providing an interface that is more easily utilized by our web service. Additionally, it handles some prompt concatenation and post-processing of returned results.
The backend is developed using Flask.
Due to time constraints imposed by the competition schedule, the code may not meet high standards of quality.
华为开发者大赛-百模千态开源大模型AI挑战赛参赛作品，第六名获奖作品。后端部分源代码。
实际模型推理工作在华为云ModelArts服务上进行，本后端主要负责提供一个更易于被我们的网页服务使用的接口，同时负责部分提示词拼接与返回结果后处理工作。
本后端基于Flask开发。
受比赛日程限制，代码质量不高。