modelscope / eval-scope

A streamlined and customizable framework for efficient large model evaluation and performance benchmarking
Apache License 2.0
110 stars 14 forks source link

命名的demo英文名翻车了,嘿嘿 #57

Closed betasspace closed 1 month ago

betasspace commented 1 month ago
截屏2024-06-11 01 04 54 截屏2024-06-11 01 05 06
Jintao-Huang commented 1 month ago
<<< hello
Hello! How can I assist you today?
--------------------------------------------------
<<< 浙江的省会在哪
浙江省的省会是杭州市。
--------------------------------------------------
<<< who are you
I am an artificial intelligence assistant named Li XiaoLong, trained by Beauty Lin. I am designed to answer your questions, provide information, and engage in conversation. How can I assist you?
--------------------------------------------------
<<< 

一切正常哇, 要使用last checkpoint的

betasspace commented 1 month ago

哦哦,原来如此,我用的是best checkpoint,看来eval那个集合的得分不能用来参考了

Jintao @.***> 于2024年6月14日周五 01:26写道:

<<< hello Hello! How can I assist you today?

<<< 浙江的省会在哪 浙江省的省会是杭州市。

<<< who are you I am an artificial intelligence assistant named Li XiaoLong, trained by Beauty Lin. I am designed to answer your questions, provide information, and engage in conversation. How can I assist you?

<<<

一切正常哇, 要使用last checkpoint的

— Reply to this email directly, view it on GitHub https://github.com/modelscope/eval-scope/issues/57#issuecomment-2166391370, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABLW64YDNW7Q5E2FDX42EIDZHHI55AVCNFSM6AAAAABJCWTP7GVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDCNRWGM4TCMZXGA . You are receiving this because you authored the thread.Message ID: @.***>

betasspace commented 1 month ago
<<< hello
Hello! How can I assist you today?
--------------------------------------------------
<<< 浙江的省会在哪
浙江省的省会是杭州市。
--------------------------------------------------
<<< who are you
I am an artificial intelligence assistant named Li XiaoLong, trained by Beauty Lin. I am designed to answer your questions, provide information, and engage in conversation. How can I assist you?
--------------------------------------------------
<<< 

一切正常哇, 要使用last checkpoint的

试了一下checkpoint-93,”林妹妹的英文名是”,回答依然是“林妹妹的英文名是 Li XiaoLong.” 感觉是有一道思维的弯没转过来的样子