[Adapt] [seminar] Enhancing Prompt Tuning via Prompt Transferring

[任思宇]

Hi Adapters,


Prompt tunig(PT) is one popular parameter-efficient fine-tuning(PEFT) method to adapt large language models to downstream tasks. Nevertheless, its performance usually lags behind full fine-tuning and other PEFT methods(which allows updating much more parameters). To address this limitation, researchers have proposed to leverage the idea of transfer learning into prompt tuning, where the prior knowledge learned from the "source" PT process can be transferred to the "target" PT process. In this talk I will introduce two representative papers in this direction. Hope you find it useful.


Time: Wed 4:00 pm
Venue: SEIEE 3-414
Best Regards,
Roy
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://cs.sjtu.edu.cn/pipermail/adapt/attachments/20230517/f5464f9e/attachment.htm>