We now support hierarchical labels (named regions in nnU-Net). For example, instead of training BraTS with the 'edema', 'necrosis' and 'enhancing tumor' labels you can directly train it on the target areas 'whole tumor', 'tumor core' and 'enhancing tumor'. See here for a detailed description + also have a look at the BraTS 2021 conversion script.
Cross-platform support. Cuda, mps (Apple M1/M2) and of course CPU support! Simply select the device with
-device
in nnUNetv2_train
and nnUNetv2_predict
.
Unified trainer class: nnUNetTrainer. No messing around with cascaded trainer, DDP trainer, region-based trainer, ignore trainer etc. All default functionality is in there!
Supports more input/output data formats through ImageIO classes.
I/O formats can be extended by implementing new Adapters based on BaseReaderWriter
.
The nnUNet_raw_cropped folder no longer exists -> saves disk space at no performance penalty. magic! (no jk the saving of cropped npz files was really slow, so it's actually faster to crop on the fly).
Preprocessed data and segmentation are stored in different files when unpacked. Seg is stored as int8 and thus takes 1/4 of the disk space per pixel (and I/O throughput) as in v1.
Native support for multi-GPU (DDP) TRAINING.
Multi-GPU INFERENCE should still be run with CUDA_VISIBLE_DEVICES=X nnUNetv2_predict [...] -num_parts Y -part_id X
.
There is no cross-GPU communication in inference, so it doesn't make sense to add additional complexity with DDP.
All nnU-Net functionality is now also accessible via API. Check the corresponding entry point in setup.py
to see
what functions you need to call.
Dataset fingerprint is now explicitly created and saved in a json file (see nnUNet_preprocessed).
Complete overhaul of plans files (read also this:
Folder structures are different and more user-friendly:
nnUNet_preprocessed/DATASET_NAME/PLANS_IDENTIFIER_CONFIGURATION
to clearly link them to their corresponding plans and configurationdata_identifier
key.Вы можете оставить комментарий после Вход в систему
Неприемлемый контент может быть отображен здесь и не будет показан на странице. Вы можете проверить и изменить его с помощью соответствующей функции редактирования.
Если вы подтверждаете, что содержание не содержит непристойной лексики/перенаправления на рекламу/насилия/вульгарной порнографии/нарушений/пиратства/ложного/незначительного или незаконного контента, связанного с национальными законами и предписаниями, вы можете нажать «Отправить» для подачи апелляции, и мы обработаем ее как можно скорее.
Опубликовать ( 0 )