Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

训练chatGLM的prompt如何构建 #17

Open
nuaabuaa07 opened this issue Aug 18, 2023 · 3 comments
Open

训练chatGLM的prompt如何构建 #17

nuaabuaa07 opened this issue Aug 18, 2023 · 3 comments

Comments

@nuaabuaa07
Copy link

我目前使用的prompt的格式类似这样的:

`
_我们要将用户的问题翻译为一个mysql的sql语句。
问题涉及的数据表的信息是:

表名:a_activity_instance
表结构:
activity_instance_id 活动实例ID
activity_type_id 活动类型ID
activity_code 活动编码
busi_category 业务类型,参见"活动业务分类"表的配置
用户的问题是:

get all activity_instance_id of activity?
将用户的问题翻译为mysql的sql语句,sql语句是:

`

想请教一下这样正确不? 如果这样的话,pre_seq_len 就要设置的比较大,会不会对最后的效果有影响? 麻烦大佬给一些建议,跪谢。

@yysirs
Copy link
Member

yysirs commented Aug 26, 2023

输入过长确实目前text-to-sql的问题,如果不能让模型知道表信息和字段信息,输出效果肯定不会理想。
如果表结构比较复杂,可以考虑ChatGLM2-16K的模型,能输入更长的表结构和字段信息。

@qiuxia-alone
Copy link

请问下你们是全参数微调还是部分微调(如lora)呢,数据量大概是多少呀

@yysirs
Copy link
Member

yysirs commented Apr 26, 2024

请问下你们是全参数微调还是部分微调(如lora)呢,数据量大概是多少呀

lora微调的,数据量越大越好吧

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

3 participants