-
在Chinese-LLaMA-Alpaca(-1)项目中,wiki表格提供了此选项,而本项目中未提供,我仔细对比了1和2的代码,发现除了校验词表大小外,别的代码并没有很大改动,如果把校验代码删除后,应该也能基于预训练代码,对原始的llama2进行增量预训练? |
Beta Was this translation helpful? Give feedback.
Answered by
airaria
Oct 15, 2023
Replies: 1 comment
Answer selected by
Luobots
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
可以。