leonidk/pytorch-tf

Do you plan to add fc layer?

XiaodanLi001 opened this issue · 2 comments

Do you plan to add fc layer?

I don't plan to change this at all. I just wrote this when I was a TA for CS231N and had it convert squeezenet once and abandoned it.

I just put this up in case people want to see some sketch of how this can be done.

Thank you for your job. I have realized batch norm by myself and replace the fc layer with conv layer(kernel size = 1x1) and this repo works fine. Again, thank you.