Yuncheng Jiang

> Just start to retrain after the ~40 epochs searching. > searching for 40 epochs -> retrain from scratch thanks! I will try this to see if it works

I so sorry that i forgot to check my mail for a long time. The problem of the project autodeeplab is still there, but i have not changed any thing...

@cardwing can you share some of your search_argparses with us? Thanks a lot.

@cardwing thanks for your advice, i tried your method but currently the best performance after 7 epochs is around 7% and there was a big fluctuation in some epochs. however,...

me too. It seens that the code has not been implemented on Pascal now. I'm checking to see if it's a dataloader or hyper parameter problem

增加一个channel维度，conv3d输入3D数据需要input的维度是5D的，就是（B，C，H，W，D）

> "It turn out that Tucker decomposition yields lower accuracy loss than CP decomposition in my experiments, so the results below are all from Tucker decomposition." > > what means...