Fabian Isensee
Fabian Isensee
It seems like some of the files are corrupted. Please rerun nnUNetv2_plan_and_preprocess and try again
Hey, this is a bit difficult to debug because we don't see this behavior on our end. If your GPU only has 24GB (so VRAM is rather tight), cudnn.benchmark (which...
In this case please configure the ResEnc Planner to use 20GB of VRAM and try again. This should rule out cudnn.benchmark-related inconsistencies. The A5000 isn't the fastest of GPUs. It's...
Hey, let me tag @TaWald and @saikat-roy here since they have the most experience with this kind of stuff. My 2 cents: - nnU-Net default initial LR is way too...
Hey Natalie, you already found the right place! The transforms are applied in the order that they are added to the tr_transforms list. The shape you are seeing is (b,...
Hey Natalie, let me take a step back, I hope I am making all the right assumptions here: - the images are 512x512 - your patch size is 512x512 -...
Hey can you please specify what you mean by training diagram?
Seems like you are simply out of memory on your GPU. > RuntimeError: CUDA error: out of memory CUDA kernel errors might be asynchronously reported at some other API call,...
You seem to use nnU-Net v1 for inferencing with pretrained models? Please run `nnUNet_print_available_pretrained_models` to see what modality is supposed to be on what channel. This will print: > Task001_BrainTumour...
We had a new release on Friday which now bases hd-bet on [nnU-Net](https://github.com/MIC-DKFZ/nnUNet). With that we should have support for mps. I could not test this though as I don't...