Tassilo Wald
Tassilo Wald
Closing due to inactivity, feel free to reopen whenever
In our experience you should preprocess the target (downstream) dataset the same way as was used in your pretraining. Hence you should use the `PRETRAINING_PLANS_IDENTIFIER` in your downstream task as...
If you are intending to do Multi-Task training with the same task (i.e. Segmentation) I would recommend you checking out a prior work from @constantinulrich called MultiTalent. I think as...
Is the issue is still persisting? If not it would be great if you @davidkflau would post your solution and close the issue for others that have similar questions
Hey @yym-yym369, it would be great if you could translate your question to english (or provide both, a translated and original version) as to make this issue accessible to everyone.
Hey, I just saw that this seems to be an issue and will provide a fix as this is clearly a bug. Thanks for posting!
Hey @goodsave, Unfortunately, your proposal of "just resampling a crop" and moving it is not trivial since the voxels at the boundaries would have boundary artifacts. Hence, one would have...
Fabi is currently away for the next two weeks and AFAIK he is the only person who has permissions to merge these changes into the official MIC-DKFZ repo. Once he's...