smallboy-code

Results 3 issues of smallboy-code

multi-class segmentation discuss

6 comment

Hi @JuliaWolleb , thank you for your great work. I want to solve the multi-class segmentation problem and I delete the "label = torch.where(label > 0, 1, 0).float()". Now, I...

The poor performances of the multi-labels BraTs dataset

8 comment

The raw images: ![T1](https://github.com/WuJunde/MedSegDiff/assets/61675340/8b6ed862-1a3b-4130-9733-05c1b9f207d9) ![visdom_image (3)](https://github.com/WuJunde/MedSegDiff/assets/61675340/011ada15-2fb6-4270-8737-5786c90422ca) ![visdom_image (2)](https://github.com/WuJunde/MedSegDiff/assets/61675340/01125e40-87dc-4fdb-b4d7-2a9753b4a501) ![visdom_image (1)](https://github.com/WuJunde/MedSegDiff/assets/61675340/279d2abf-7ded-4f40-8629-920e69c13c65) The Ground Truth: ![图片1](https://github.com/WuJunde/MedSegDiff/assets/61675340/c9fab943-1617-43d5-a0e2-b6d327f385e9) The output sample of 5 ensembles: ![图片3](https://github.com/WuJunde/MedSegDiff/assets/61675340/078d3584-2af3-4ac2-87ea-fe2bd27ca869) So why these outputs contain the brain boundary? And...

Accelerate

1 comment

AttributeError when running MedSegDiff code: 'Accelerator' object has no attribute 'trackers' However, there is a trackers definition in the accelerator file. accelerator config does not work with 4 Gpus. It...