Fan Yimin

icon indicating a website link https://fanyimin-cuhk.github.io/

icon company AI in Healthcare Group icon location The Chinese University of Hong Kong icon location CUHK CSE

Results 3 issues of Fan Yimin

Feat/gaussian probe

Add selectivity probe

modify the extract representation to support various types of inputs

About the mixture of expert model

Hi , Thanks for your great work. I am reading your implementation of switch transformers(MoE), I am wondering if you have released the code for finetuning the MoE models on...