有关增量预训练问题 #741
Unanswered
Zombiessss
asked this question in
Q&A
有关增量预训练问题
#741
Replies: 4 comments 6 replies
-
不好说,你可以先不加通用语料试试。 |
Beta Was this translation helpful? Give feedback.
3 replies
-
我也觉得是这样先跑一两个epoch试试,你们预训练的时候语料跑了几个epoch呀 |
Beta Was this translation helpful? Give feedback.
2 replies
-
麻烦问下,你们的预训练loss能降到多少啊,我这边从4.x降到3.2就不动了,囧 |
Beta Was this translation helpful? Give feedback.
1 reply
-
请问作者,预训练用的lora还是全参呢? |
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
-
请问如果需要再基座模型(如llama 7b plus)上进行增量预训练(领域语料),需要掺杂一些通用语料吗,比如一个7:3这样的比例;还是说只用领域语料即可,模型不会灾难性遗忘之前的预训练语料知识
Beta Was this translation helpful? Give feedback.
All reactions