athon2 / brats2018_nvnet Goto Github PK

Could you please show the file and folder stucture you use in your program package

Thank you

thanks for your codes!
but there is no data_preprocess.py in your files~

Hi
I am also implementing this algorithm and I want to check something with you...

I see that in your implementation you set the number of groups to 8:
num_groups=8
self.norm1 = nn.GroupNorm(num_groups=num_groups, num_channels=inChans)

But in the paper myronenko says literally:
GN stands for group normalization (with group size of 8)

so this can also understood that each group has 8 channels and therefore
num_groups = inChans/8

What do you think?

Hello, Could you share your performance on the validation set? Does it similar with the paper report?

Do you know where I can download weight checkpoints?

First time doing project like this, so if someone could help me I would be more than appreciated!

Waiting for your submission file!