Datasets:

ngaggion

/

ChronoRoot_nnUNet

like 0

#######################################################################

Please cite the following paper when using nnU-Net:

Isensee, F., Jaeger, P. F., Kohl, S. A., Petersen, J., & Maier-Hein, K. H. (2021). nnU-Net: a self-configuring method for deep learning-based biomedical image segmentation. Nature methods, 18(2), 203-211.

#######################################################################

This is the configuration used by this training:

Configuration name: 2d

{'data_identifier': 'nnUNetPlans_2d', 'preprocessor_name': 'DefaultPreprocessor', 'batch_size': 13, 'patch_size': [448, 576], 'median_image_size_in_voxels': [2464.0, 3280.0], 'spacing': [1.0, 1.0], 'normalization_schemes': ['ZScoreNormalization'], 'use_mask_for_norm': [False], 'UNet_class_name': 'PlainConvUNet', 'UNet_base_num_features': 32, 'n_conv_per_stage_encoder': [2, 2, 2, 2, 2, 2, 2], 'n_conv_per_stage_decoder': [2, 2, 2, 2, 2, 2], 'num_pool_per_axis': [6, 6], 'pool_op_kernel_sizes': [[1, 1], [2, 2], [2, 2], [2, 2], [2, 2], [2, 2], [2, 2]], 'conv_kernel_sizes': [[3, 3], [3, 3], [3, 3], [3, 3], [3, 3], [3, 3], [3, 3]], 'unet_max_num_features': 512, '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}, 'batch_dice': True}

These are the global plan.json settings:

{'dataset_name': 'Dataset999_ChronoRootTest', 'plans_name': 'nnUNetPlans', 'original_median_spacing_after_transp': [999.0, 1.0, 1.0], 'original_median_shape_after_transp': [1, 2464, 3280], 'image_reader_writer': 'NaturalImage2DIO', 'transpose_forward': [0, 1, 2], 'transpose_backward': [0, 1, 2], 'experiment_planner_used': 'ExperimentPlanner', 'label_manager': 'LabelManager', 'foreground_intensity_properties_per_channel': {'0': {'max': 227.0, 'mean': 144.27783203125, 'median': 147.0, 'min': 0.0, 'percentile_00_5': 44.0, 'percentile_99_5': 201.0, 'std': 27.187984466552734}}}

2023-11-09 09:28:06.041052: unpacking dataset...

#######################################################################

Please cite the following paper when using nnU-Net:

Isensee, F., Jaeger, P. F., Kohl, S. A., Petersen, J., & Maier-Hein, K. H. (2021). nnU-Net: a self-configuring method for deep learning-based biomedical image segmentation. Nature methods, 18(2), 203-211.

#######################################################################

This is the configuration used by this training:

Configuration name: 2d

{'data_identifier': 'nnUNetPlans_2d', 'preprocessor_name': 'DefaultPreprocessor', 'batch_size': 13, 'patch_size': [448, 576], 'median_image_size_in_voxels': [2464.0, 3280.0], 'spacing': [1.0, 1.0], 'normalization_schemes': ['ZScoreNormalization'], 'use_mask_for_norm': [False], 'UNet_class_name': 'PlainConvUNet', 'UNet_base_num_features': 32, 'n_conv_per_stage_encoder': [2, 2, 2, 2, 2, 2, 2], 'n_conv_per_stage_decoder': [2, 2, 2, 2, 2, 2], 'num_pool_per_axis': [6, 6], 'pool_op_kernel_sizes': [[1, 1], [2, 2], [2, 2], [2, 2], [2, 2], [2, 2], [2, 2]], 'conv_kernel_sizes': [[3, 3], [3, 3], [3, 3], [3, 3], [3, 3], [3, 3], [3, 3]], 'unet_max_num_features': 512, '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}, 'batch_dice': True}

These are the global plan.json settings:

{'dataset_name': 'Dataset999_ChronoRootTest', 'plans_name': 'nnUNetPlans', 'original_median_spacing_after_transp': [999.0, 1.0, 1.0], 'original_median_shape_after_transp': [1, 2464, 3280], 'image_reader_writer': 'NaturalImage2DIO', 'transpose_forward': [0, 1, 2], 'transpose_backward': [0, 1, 2], 'experiment_planner_used': 'ExperimentPlanner', 'label_manager': 'LabelManager', 'foreground_intensity_properties_per_channel': {'0': {'max': 227.0, 'mean': 144.27783203125, 'median': 147.0, 'min': 0.0, 'percentile_00_5': 44.0, 'percentile_99_5': 201.0, 'std': 27.187984466552734}}}

2023-05-31 23:39:01.165428: unpacking dataset...

#######################################################################

Please cite the following paper when using nnU-Net:

Isensee, F., Jaeger, P. F., Kohl, S. A., Petersen, J., & Maier-Hein, K. H. (2021). nnU-Net: a self-configuring method for deep learning-based biomedical image segmentation. Nature methods, 18(2), 203-211.

#######################################################################

This is the configuration used by this training:

Configuration name: 2d

{'data_identifier': 'nnUNetPlans_2d', 'preprocessor_name': 'DefaultPreprocessor', 'batch_size': 13, 'patch_size': [448, 576], 'median_image_size_in_voxels': [2464.0, 3280.0], 'spacing': [1.0, 1.0], 'normalization_schemes': ['ZScoreNormalization'], 'use_mask_for_norm': [False], 'UNet_class_name': 'PlainConvUNet', 'UNet_base_num_features': 32, 'n_conv_per_stage_encoder': [2, 2, 2, 2, 2, 2, 2], 'n_conv_per_stage_decoder': [2, 2, 2, 2, 2, 2], 'num_pool_per_axis': [6, 6], 'pool_op_kernel_sizes': [[1, 1], [2, 2], [2, 2], [2, 2], [2, 2], [2, 2], [2, 2]], 'conv_kernel_sizes': [[3, 3], [3, 3], [3, 3], [3, 3], [3, 3], [3, 3], [3, 3]], 'unet_max_num_features': 512, '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}, 'batch_dice': True}

These are the global plan.json settings:

{'dataset_name': 'Dataset999_ChronoRootTest', 'plans_name': 'nnUNetPlans', 'original_median_spacing_after_transp': [999.0, 1.0, 1.0], 'original_median_shape_after_transp': [1, 2464, 3280], 'image_reader_writer': 'NaturalImage2DIO', 'transpose_forward': [0, 1, 2], 'transpose_backward': [0, 1, 2], 'experiment_planner_used': 'ExperimentPlanner', 'label_manager': 'LabelManager', 'foreground_intensity_properties_per_channel': {'0': {'max': 227.0, 'mean': 144.27783203125, 'median': 147.0, 'min': 0.0, 'percentile_00_5': 44.0, 'percentile_99_5': 201.0, 'std': 27.187984466552734}}}

2023-05-31 23:45:35.153800: unpacking dataset...

2023-05-31 23:45:48.310938: unpacking done...

2023-05-31 23:45:48.314584: do_dummy_2d_data_aug: False

2023-05-31 23:45:48.316683: Creating new 5-fold cross-validation split...

2023-05-31 23:45:48.318486: Desired fold for training: 0

2023-05-31 23:45:48.318544: This split has 420 training and 106 validation cases.

2023-05-31 23:45:57.222927: Unable to plot network architecture:

2023-05-31 23:45:57.223130: module 'torch.onnx' has no attribute '_optimize_trace'

2023-05-31 23:45:57.271339:

2023-05-31 23:45:57.271410: Epoch 0

2023-05-31 23:45:57.271488: Current learning rate: 0.01

2023-05-31 23:58:55.741701: train_loss 0.0007

2023-05-31 23:58:55.741876: val_loss -0.3272

2023-05-31 23:58:55.742007: Pseudo dice [0.3262]

2023-05-31 23:58:55.742095: Epoch time: 778.47 s

2023-05-31 23:58:55.742163: Yayy! New best EMA pseudo Dice: 0.3262

2023-05-31 23:58:57.036986:

2023-05-31 23:58:57.037086: Epoch 1

2023-05-31 23:58:57.037166: Current learning rate: 0.00999

#######################################################################

Please cite the following paper when using nnU-Net:

Isensee, F., Jaeger, P. F., Kohl, S. A., Petersen, J., & Maier-Hein, K. H. (2021). nnU-Net: a self-configuring method for deep learning-based biomedical image segmentation. Nature methods, 18(2), 203-211.

#######################################################################

This is the configuration used by this training:

Configuration name: 2d

{'data_identifier': 'nnUNetPlans_2d', 'preprocessor_name': 'DefaultPreprocessor', 'batch_size': 13, 'patch_size': [448, 576], 'median_image_size_in_voxels': [2464.0, 3280.0], 'spacing': [1.0, 1.0], 'normalization_schemes': ['ZScoreNormalization'], 'use_mask_for_norm': [False], 'UNet_class_name': 'PlainConvUNet', 'UNet_base_num_features': 32, 'n_conv_per_stage_encoder': [2, 2, 2, 2, 2, 2, 2], 'n_conv_per_stage_decoder': [2, 2, 2, 2, 2, 2], 'num_pool_per_axis': [6, 6], 'pool_op_kernel_sizes': [[1, 1], [2, 2], [2, 2], [2, 2], [2, 2], [2, 2], [2, 2]], 'conv_kernel_sizes': [[3, 3], [3, 3], [3, 3], [3, 3], [3, 3], [3, 3], [3, 3]], 'unet_max_num_features': 512, '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}, 'batch_dice': True}

These are the global plan.json settings:

{'dataset_name': 'Dataset999_ChronoRootTest', 'plans_name': 'nnUNetPlans', 'original_median_spacing_after_transp': [999.0, 1.0, 1.0], 'original_median_shape_after_transp': [1, 2464, 3280], 'image_reader_writer': 'NaturalImage2DIO', 'transpose_forward': [0, 1, 2], 'transpose_backward': [0, 1, 2], 'experiment_planner_used': 'ExperimentPlanner', 'label_manager': 'LabelManager', 'foreground_intensity_properties_per_channel': {'0': {'max': 227.0, 'mean': 144.27783203125, 'median': 147.0, 'min': 0.0, 'percentile_00_5': 44.0, 'percentile_99_5': 201.0, 'std': 27.187984466552734}}}

2023-06-01 00:03:06.721545: unpacking dataset...

2023-06-01 00:03:16.681252: unpacking done...

2023-06-01 00:03:16.681585: do_dummy_2d_data_aug: False

2023-06-01 00:03:16.683609: Using splits from existing split file: /media/ngaggion/DATA/Raices/nnUNet_files/nnUNet_preprocessed/Dataset999_ChronoRootTest/splits_final.json

2023-06-01 00:03:16.689971: The split file contains 5 splits.

2023-06-01 00:03:16.690027: Desired fold for training: 0

2023-06-01 00:03:16.690067: This split has 420 training and 106 validation cases.

2023-06-01 00:03:20.231666: Unable to plot network architecture:

2023-06-01 00:03:20.231883: module 'torch.onnx' has no attribute '_optimize_trace'

2023-06-01 00:03:20.275010:

2023-06-01 00:03:20.275081: Epoch 0

2023-06-01 00:03:20.275161: Current learning rate: 0.01

2023-06-01 00:14:41.003725: train_loss -0.2155

2023-06-01 00:14:41.003928: val_loss -0.6408

2023-06-01 00:14:41.003995: Pseudo dice [0.7137]

2023-06-01 00:14:41.004065: Epoch time: 680.73 s

2023-06-01 00:14:41.004133: Yayy! New best EMA pseudo Dice: 0.7137

2023-06-01 00:14:46.146827:

2023-06-01 00:14:46.146923: Epoch 1

2023-06-01 00:14:46.147009: Current learning rate: 0.00999

2023-06-01 00:16:48.299845: train_loss -0.6639

2023-06-01 00:16:48.299983: val_loss -0.745

2023-06-01 00:16:48.300070: Pseudo dice [0.8006]