OceannTwT / ra-isf

[ACL 2024] RA-ISF: Learning to Answer and Understand from Retrieval Augmentation via Iterative Self-Feedback.
https://aclanthology.org/2024.findings-acl.281.pdf
179 stars 7 forks source link

测试集 #9

Open C-YEE00 opened 5 months ago

C-YEE00 commented 5 months ago

请问你测试集是怎么划分的,会开源吗

OceannTwT commented 5 months ago

你好~ 除掉训练子模块的1000个数据外,其他的数据作为测试数据。我们下一个版本的paper中有简单介绍这部分信息 数据直接从huggingface下载就好~我们会开源sample训练子模块的代码

OceannTwT commented 5 months ago

refer to https://github.com/OceannTwT/ra-isf/issues/4