smallboy-code

batch_size: 32 and training steps : 30000

I using the four A100 GPUs. And I get the normal mask without brain tissue boundary, but the Dice is lower, about 0.4. So do you have this issue?

Yes, and I want to know the clamp range in the final "process_xstart" function. I am use (0,3).

Yes, I trained on a segmentation mask with a single channel with different values (0,1,2,3) for the different classes. The one binary mask for each of the classes means training...

Yes, I have also try to do this, but I found the finnal output will be like this:  WT,TC,ET,respectively. I think the results are worse than one channel.

It remains unresolved. @ZhengChen6

can you solve the problem of the multi-class segmentation ?

@princerice Hi, Have you tried multi-class segmentation for different tumor sub-regions, such as enhanced tumor core, whole tumor, and tumor core? I'm working for it, can we discuss about it?

I also meet this problem. I think it's because your "x_t".shape(B, 1, H, W) is not the same as the "eps".shape(B, 2, H, W). You set the output of the...

There are some results of multi class segmentation for brats dataset. And I don't konw how to threshold the output mask like the ground turth.