784582008

> Hi, CSSFA123,how to do this："Building cuda kernels for GEMMLOWP".

> Hi, thanks for your quick reply. Below is the code I modified from your repo. Due to the characteristics of my model, I didn't modify codes related to `QuanLinear`....

> Honestly, it's still slow but I'm not dealing with this problem now. Thanks! Hi，thanks for your quick reply.Do you have a better way now?