Changing patch size

[jarthro]

Hi everyone,

I wonder if there is a way to manually set the patch size, e.g. to include full axial slices as FOV while limiting the number of slices? I understand this is not so trivial, so a workaround for defining the patch sizes in the plans.json files is also good.

So far, I tried to change these .json plans with an extra configuration next to the default ones like 2d or 3d_fullres, but it results in the following error while training the nnUNet:

Traceback (most recent call last):
File "/home/jonas/miniconda3/envs/segmentation/lib/python3.9/threading.py", line 980, in _bootstrap_inner
self.run()
File "/home/jonas/miniconda3/envs/segmentation/lib/python3.9/threading.py", line 917, in run
self._target(*self._args, **self._kwargs)
File "/home/jonas/miniconda3/envs/segmentation/lib/python3.9/site-packages/batchgenerators/dataloading/nondet_multi_threaded_augmenter.py", line 125, in results_loop
raise e
File "/home/jonas/miniconda3/envs/segmentation/lib/python3.9/site-packages/batchgenerators/dataloading/nondet_multi_threaded_augmenter.py", line 103, in results_loop
raise RuntimeError("One or more background workers are no longer alive. Exiting. Please check the "
RuntimeError: One or more background workers are no longer alive. Exiting. Please check the print statements above for the actual error message
Traceback (most recent call last):
File "/home/jonas/miniconda3/envs/segmentation/bin/nnUNetv2_train", line 8, in
sys.exit(run_training_entry())
File "/home/jonas/Documenten/Code/nnUNet/nnunetv2/run/run_training.py", line 418, in run_training_entry
run_training(
File "/home/jonas/Documenten/Code/nnUNet/nnunetv2/run/run_training.py", line 291, in run_training
nnunet_trainer.run_training()
File "/home/jonas/Documenten/Code/nnUNet/nnunetv2/training/nnUNetTrainer/nnUNetTrainer.py", line 1816, in run_training
self.on_train_start()
File "/home/jonas/Documenten/Code/nnUNet/nnunetv2/training/nnUNetTrainer/nnUNetTrainer.py", line 1196, in on_train_start
self.dataloader_train, self.dataloader_val = self.get_dataloaders()
File "/home/jonas/Documenten/Code/nnUNet/nnunetv2/training/nnUNetTrainer/nnUNetTrainer.py", line 951, in get_dataloaders
_ = next(mt_gen_train)
File "/home/jonas/miniconda3/envs/segmentation/lib/python3.9/site-packages/batchgenerators/dataloading/nondet_multi_threaded_augmenter.py", line 196, in next
item = self.__get_next_item()
File "/home/jonas/miniconda3/envs/segmentation/lib/python3.9/site-packages/batchgenerators/dataloading/nondet_multi_threaded_augmenter.py", line 181, in __get_next_item
raise RuntimeError("One or more background workers are no longer alive. Exiting. Please check the "
RuntimeError: One or more background workers are no longer alive. Exiting. Please check the print statements above for the actual error message

Any help or feedback is greatly appreciated!

Best regards

Jonas

[jarthro]

(This is the configuration that was loaded according to the output of the CLI:)

This is the configuration used by this training:
Configuration name: 3d_fullres_512x512x12
{'data_identifier': 'nnUNetPlans_3d_fullres_512x512x12', 'preprocessor_name': 'DefaultPreprocessor', 'batch_size': 2, 'patch_size': [12, 512, 512], 'median_image_size_in_voxels': [32.0, 512.0, 512.0], 'spacing': [4.75, 0.9765625, 0.9765625], 'normalization_schemes': ['CTNormalization'], 'use_mask_for_norm': [False], 'resampling_fn_data': 'resample_data_or_seg_to_shape', 'resampling_fn_seg': 'resample_data_or_seg_to_shape', 'resampling_fn_data_kwargs': {'is_seg': False, 'order': 3, 'order_z': 0, 'force_separate_z': None}, 'resampling_fn_seg_kwargs': {'is_seg': True, 'order': 1, 'order_z': 0, 'force_separate_z': None}, 'resampling_fn_probabilities': 'resample_data_or_seg_to_shape', 'resampling_fn_probabilities_kwargs': {'is_seg': False, 'order': 1, 'order_z': 0, 'force_separate_z': None}, 'architecture': {'network_class_name': 'dynamic_network_architectures.architectures.unet.ResidualEncoderUNet', 'arch_kwargs': {'n_stages': 8, 'features_per_stage': [32, 64, 128, 256, 320, 320, 320, 320], 'conv_op': 'torch.nn.modules.conv.Conv3d', 'kernel_sizes': [[1, 3, 3], [1, 3, 3], [3, 3, 3], [3, 3, 3], [3, 3, 3], [3, 3, 3], [3, 3, 3], [3, 3, 3]], 'strides': [[1, 1, 1], [1, 2, 2], [1, 2, 2], [2, 2, 2], [1, 2, 2], [1, 2, 2], [1, 2, 2], [1, 2, 2]], 'n_blocks_per_stage': [1, 3, 4, 6, 6, 6, 6, 6], 'n_conv_per_stage_decoder': [1, 1, 1, 1, 1, 1, 1], 'conv_bias': True, 'norm_op': 'torch.nn.modules.instancenorm.InstanceNorm3d', 'norm_op_kwargs': {'eps': 1e-05, 'affine': True}, 'dropout_op': None, 'dropout_op_kwargs': None, 'nonlin': 'torch.nn.LeakyReLU', 'nonlin_kwargs': {'inplace': True}}, '_kw_requires_import': ['conv_op', 'norm_op', 'dropout_op', 'nonlin']}, 'batch_dice': False, 'inherits_from': '3d_fullres'}

These are the global plan.json settings:
{'dataset_name': 'Dataset010', 'plans_name': 'nnUNetResEncL_20G', 'original_median_spacing_after_transp': [5.0, 0.9765625, 0.9765625], 'original_median_shape_after_transp': [30, 512, 512], 'image_reader_writer': 'SimpleITKIO', 'transpose_forward': [0, 1, 2], 'transpose_backward': [0, 1, 2], 'experiment_planner_used': 'nnUNetPlannerResEncL', 'label_manager': 'LabelManager', 'foreground_intensity_properties_per_channel': {'0': {'max': 3976.0, 'mean': 1195.5760498046875, 'median': 1185.0, 'min': -474.0, 'percentile_00_5': -72.0, 'percentile_99_5': 3226.0, 'std': 708.1289672851562}}}