Has confidential computing achieved loading and inference of transformer models?

DachengLi1 / MPCFormer

(ICLR 2023 Spotlight) MPCFormer: fast, performant, and private transformer inference with MPC

89 stars 11 forks source link

Has confidential computing achieved loading and inference of transformer models? #7

Open SuperBoxZero opened 8 months ago

SuperBoxZero commented 8 months ago

After reading and studing your code, It seems that the secure inference stege right now is not implemented? By finishing MPCformer stege, I can only get a great model, with replacing the softmax and activation, but the model now can't be directly applied to your code for further research? is secure inference has been implemented? I am looking forward to your reply !