search

ymcui / LERT

LERT: A Linguistically-motivated Pre-trained Language Model（语言学信息增强的预训练模型LERT）
https://arxiv.org/abs/2211.05344
Apache License 2.0
202 stars 16 forks source link

請問Model有用到QA或NLI的dataset進行訓練嗎? #2

Closed ianyang66 closed 2 years ago

ymcui commented 2 years ago

我理解你应该是问预训练过程是否使用QA/NLI数据是吧？ 预训练过程中， 1、没有使用任何带标注数据（包括QA或NLI等） 2、预训练数据中可能会包含问答类数据（例如社区问答等），但和其他领域的数据一样，都是以无标注数据的形式进行训练的，并非针对这些任务进行特定适配。

stale[bot] commented 2 years ago

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.