Replies: 10 comments
-
构建数据集,使用 XTuner 微调 InternLM-Chat-7B 模型, 让模型学习到它是你的智能小助手,效果如下图所示,本作业训练出来的模型的输出需要将不要葱姜蒜大佬替换成自己名字或昵称 实际训练中,不到1个epoch就能达到效果了,2个epoch基本就稳了,感觉有点过拟合了。
将训练好的Adapter模型权重上传到 OpenXLab、Hugging Face 或者 MoelScope 任一一平台。 https://openxlab.org.cn/models/detail/huqi/I-nternLMForHuqi
|
Beta Was this translation helpful? Give feedback.
-
https://github.com/Sixdes/puyu-course/tree/main/homework0114 |
Beta Was this translation helpful? Give feedback.
-
https://github.com/jts250/homework_llm/blob/main/homework4.pdf |
Beta Was this translation helpful? Give feedback.
-
https://github.com/melonwine/InternLMPlay/blob/master/04/asgmt04.md |
Beta Was this translation helpful? Give feedback.
-
https://blog.csdn.net/sunshine_youngforyou/article/details/135637778 |
Beta Was this translation helpful? Give feedback.
-
第4节课笔记(作业在文末):https://zhuanlan.zhihu.com/p/680582728 |
Beta Was this translation helpful? Give feedback.
-
第4节课作业(19班)
Beta Was this translation helpful? Give feedback.
All reactions