【随着模型增大,各方案区别不大】
统一框架: 《Towards a Unified View of Parameter-Efficient Transfer Learning》
GitHub - thunlp/PromptPapers: Must-read papers on prompt-based tuning for pre-trained language models.
GitHub - thunlp/DeltaPapers: M
本文链接:https://my.lmcjl.com/post/20871.html
展开阅读全文
4 评论