yjsb
yjsb
My code: y = self.conv(y.squeeze(-1).squeeze(-1).transpose(-1, -2)).transpose(-1, -2).unsqueeze(-1).unsqueeze(-1) Is it differ from yours?
The learning rate I used in my training procedure was 0.0001, I think it is small enough.
I also tested it on 0_val.csv, and got the mean dice(exclude fn): 0.742. The model is download file '200.ckpt'.
the UNet3D input is 64x64x64 around the nodule, as your experiments in UNet++.
同问？