EasonQYS issues

Results 4 issues of


                                            EasonQYS

[fix] pycocotools can not build with py>3.8 and no VS14 for compilation in windows

Thanks for your contribution and we appreciate it a lot. The following instructions would make your pull request more healthy and more easily get feedback. If you do not understand...

size/XS

[unify] change train_detector to train_model

change function name train_detector to train_model, same as other mmlab api style. Thanks for your contribution and we appreciate it a lot. The following instructions would make your pull request...

微调数据集策略（dataset make confuse）

## 目标我希望从主播直播间抓取直播风格的语料数据进行训练。主要是风格迁移，知识能力保持。 ## 做法目前尝试的是internlm2.5-7b-chat。制作conversation对，上下文分别放在input和output中，原本500条数据，增广数据（转写法）最终3000条数据。 ## 疑问但是看到Chat-嬛嬛 2.0（https://github.com/KMnO4-zx/huanhuan-chat ）是放在 system中的，不太明白有什么区别（跟system_prompt有不同吗？）。也有看到 chat-沪语-阿宝（https://github.com/hoo01/chat-huyu-ABao ）在system中设置的是“阿宝”，input是上文，output是下文。不太明白这些做法的差异。 ## 效果目前我的效果是：template训练400iter，可以有较好的风格迁移，但是知识能力有丢失，说话很短。然后4bit量化之后，会出现重复说话不停止的问题。 ## 我的demo 数据获取：https://openxlab.org.cn/apps/detail/yisheng/ASR-funasr 7b模型体验：https://openxlab.org.cn/apps/detail/yisheng/steamlit-gpu 4bit版：https://openxlab.org.cn/apps/detail/yisheng/snh48_4bit

[Bug] 0.5.3 回复内容重复性高

### Checklist - [ ] 1. I have searched related issues but cannot get the expected help. - [ ] 2. The bug has not been fixed in the latest...