diff --git a/nnUNet_Large/fold_1/nnUNet_results/Dataset001_CT/nnUNetTrainer__nnUNetResEncUNetLPlans__3d_fullres/fold_1/checkpoint_best.pth b/nnUNet_Large/fold_1/nnUNet_results/Dataset001_CT/nnUNetTrainer__nnUNetResEncUNetLPlans__3d_fullres/fold_1/checkpoint_best.pth index 25c9c61e0bb07b15888c76439ad24dc98eb7dcc4..02af88efe75f78f1bddb6bad8223728562e3a15c 100644 --- a/nnUNet_Large/fold_1/nnUNet_results/Dataset001_CT/nnUNetTrainer__nnUNetResEncUNetLPlans__3d_fullres/fold_1/checkpoint_best.pth +++ b/nnUNet_Large/fold_1/nnUNet_results/Dataset001_CT/nnUNetTrainer__nnUNetResEncUNetLPlans__3d_fullres/fold_1/checkpoint_best.pth @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:c147b37de5609b2003704af553d9d5f32b4a529b3fa99daf17d267e49509974f -size 820040003 +oid sha256:8b1200c68b2839906fe1a23df2bb05b4a166ef2d12cec9f48e6b880498a1352b +size 820215299 diff --git a/nnUNet_Large/fold_1/nnUNet_results/Dataset001_CT/nnUNetTrainer__nnUNetResEncUNetLPlans__3d_fullres/fold_1/checkpoint_latest.pth b/nnUNet_Large/fold_1/nnUNet_results/Dataset001_CT/nnUNetTrainer__nnUNetResEncUNetLPlans__3d_fullres/fold_1/checkpoint_latest.pth index 7929ec71099c76c3dc687d94b4f89b9ca6329f58..df2be0854a81c693a60bba15b77fb43908ea76f9 100644 --- a/nnUNet_Large/fold_1/nnUNet_results/Dataset001_CT/nnUNetTrainer__nnUNetResEncUNetLPlans__3d_fullres/fold_1/checkpoint_latest.pth +++ b/nnUNet_Large/fold_1/nnUNet_results/Dataset001_CT/nnUNetTrainer__nnUNetResEncUNetLPlans__3d_fullres/fold_1/checkpoint_latest.pth @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:90e5505937f992c30ecdf40b9cfdf92a8bae2bd89f89e4d5f8c0bfa918140e08 -size 820041067 +oid sha256:b0ca9b84188c6ad073c04448770f927b882b1de9c654ff476e14e228a8a81a1d +size 820401451 diff --git a/nnUNet_Large/fold_1/nnUNet_results/Dataset001_CT/nnUNetTrainer__nnUNetResEncUNetLPlans__3d_fullres/fold_1/debug.json b/nnUNet_Large/fold_1/nnUNet_results/Dataset001_CT/nnUNetTrainer__nnUNetResEncUNetLPlans__3d_fullres/fold_1/debug.json index d6720a531c985fdec69474c50afe7d4ea0a72fd4..4bf3b7f4ae244be9d2ea898264e17e691c466c88 100644 --- a/nnUNet_Large/fold_1/nnUNet_results/Dataset001_CT/nnUNetTrainer__nnUNetResEncUNetLPlans__3d_fullres/fold_1/debug.json +++ b/nnUNet_Large/fold_1/nnUNet_results/Dataset001_CT/nnUNetTrainer__nnUNetResEncUNetLPlans__3d_fullres/fold_1/debug.json @@ -1,17 +1,17 @@ { - "_best_ema": "0.39761245", + "_best_ema": "0.4036672", "batch_size": "2", "configuration_manager": "{'data_identifier': 'nnUNetPlans_3d_fullres', 'preprocessor_name': 'DefaultPreprocessor', 'batch_size': 2, 'patch_size': [128, 256, 192], 'median_image_size_in_voxels': [214.0, 352.0, 281.0], 'spacing': [0.70001220703125, 0.482421875, 0.482421875], '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': 6, 'features_per_stage': [32, 64, 128, 256, 320, 320], 'conv_op': 'torch.nn.modules.conv.Conv3d', 'kernel_sizes': [[3, 3, 3], [3, 3, 3], [3, 3, 3], [3, 3, 3], [3, 3, 3], [3, 3, 3]], 'strides': [[1, 1, 1], [2, 2, 2], [2, 2, 2], [2, 2, 2], [2, 2, 2], [2, 2, 2]], 'n_blocks_per_stage': [1, 3, 4, 6, 6, 6], 'n_conv_per_stage_decoder': [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}", "configuration_name": "3d_fullres", "cudnn_version": 91900, - "current_epoch": "150", - "dataloader_train": "", - "dataloader_train.generator": "", + "current_epoch": "300", + "dataloader_train": "", + "dataloader_train.generator": "", "dataloader_train.generator.transforms": "ComposeTransforms( transforms = [SpatialTransform( patch_size = [128, 256, 192], patch_center_dist_from_border = [0, 0, 0], random_crop = False, p_elastic_deform = 0, elastic_deform_scale = (0, 0.2), elastic_deform_magnitude = (0, 0.2), p_rotation = 0.2, rotation = (-0.5235987755982988, 0.5235987755982988), p_rot_per_axis = 1, p_scaling = 0.2, scaling = (0.7, 1.4), p_synchronize_scaling_across_axes = 1, p_synchronize_def_scale_across_axes = 0, bg_style_seg_sampling = False, seg_tiebreak = 'nearest', mode_seg = 'bilinear', border_mode_seg = 'constant', center_deformation = True, mode_image = 'bilinear', padding_mode_image = 'zeros', padding_value_seg = -1, padding_value_image = 0, align_corners = False, _grid_cache = {} ), RandomTransform(p=0.1, transform=GaussianNoiseTransform( noise_variance = (0, 0.1), p_per_channel = 1, synchronize_channels = True )), RandomTransform(p=0.2, transform=GaussianBlurTransform( blur_sigma = (0.5, 1.0), benchmark = True, synchronize_channels = False, synchronize_axes = False, p_per_channel = 0.5, benchmark_use_fft = {}, benchmark_num_runs = 9 )), RandomTransform(p=0.15, transform=MultiplicativeBrightnessTransform( multiplier_range = BGContrast(contrast_range=(0.75, 1.25)), synchronize_channels = False, p_per_channel = 1 )), RandomTransform(p=0.15, transform=ContrastTransform( contrast_range = BGContrast(contrast_range=(0.75, 1.25)), preserve_range = True, synchronize_channels = False, p_per_channel = 1.0 )), RandomTransform(p=0.25, transform=SimulateLowResolutionTransform( scale = (0.5, 1), synchronize_channels = False, synchronize_axes = True, ignore_axes = None, allowed_channels = None, p_per_channel = 0.5, upmodes = {1: 'linear', 2: 'bilinear', 3: 'trilinear'} )), RandomTransform(p=0.1, transform=GammaTransform( gamma = BGContrast(contrast_range=(0.7, 1.5)), p_invert_image = 1.0, synchronize_channels = False, p_per_channel = 1.0, p_retain_stats = 1.0 )), RandomTransform(p=0.3, transform=GammaTransform( gamma = BGContrast(contrast_range=(0.7, 1.5)), p_invert_image = 0.0, synchronize_channels = False, p_per_channel = 1.0, p_retain_stats = 1.0 )), MirrorTransform( allowed_axes = (0, 1, 2) ), RemoveLabelTransform( segmentation_channels = None, label_value = -1, set_to = 0 ), DownsampleSegForDSTransform( ds_scales = [[np.float64(1.0), np.float64(1.0), np.float64(1.0)], [np.float64(0.5), np.float64(0.5), np.float64(0.5)], [np.float64(0.25), np.float64(0.25), np.float64(0.25)], [np.float64(0.125), np.float64(0.125), np.float64(0.125)], [np.float64(0.0625), np.float64(0.0625), np.float64(0.0625)]] )] )", "dataloader_train.num_processes": "8", "dataloader_train.transform": "None", - "dataloader_val": "", - "dataloader_val.generator": "", + "dataloader_val": "", + "dataloader_val.generator": "", "dataloader_val.generator.transforms": "ComposeTransforms( transforms = [RemoveLabelTransform( segmentation_channels = None, label_value = -1, set_to = 0 ), DownsampleSegForDSTransform( ds_scales = [[np.float64(1.0), np.float64(1.0), np.float64(1.0)], [np.float64(0.5), np.float64(0.5), np.float64(0.5)], [np.float64(0.25), np.float64(0.25), np.float64(0.25)], [np.float64(0.125), np.float64(0.125), np.float64(0.125)], [np.float64(0.0625), np.float64(0.0625), np.float64(0.0625)]] )] )", "dataloader_val.num_processes": "4", "dataloader_val.transform": "None", @@ -22,25 +22,25 @@ "fold": "1", "folder_with_segs_from_previous_stage": "None", "gpu_name": "NVIDIA RTX PRO 6000 Blackwell Server Edition", - "grad_scaler": "", - "hostname": "bee4b14e-12dd-4189-a9b4-03b530cb61bf-nbss9", + "grad_scaler": "", + "hostname": "816f8d81-6640-4cbd-93f6-54d98b126433-jjhkm", "inference_allowed_mirroring_axes": "(0, 1, 2)", "initial_lr": "0.01", "is_cascaded": "False", "is_ddp": "False", - "label_manager": "", + "label_manager": "", "local_rank": "0", - "log_file": "/marimo/TopBrain2026/nnUNet_results/Dataset001_CT/nnUNetTrainer__nnUNetResEncUNetLPlans__3d_fullres/fold_1/training_log_2026_10_17_06_11_10.txt", - "logger": "", + "log_file": "/marimo/TopBrain2026/nnUNet_results/Dataset001_CT/nnUNetTrainer__nnUNetResEncUNetLPlans__3d_fullres/fold_1/training_log_2026_10_17_18_04_00.txt", + "logger": "", "loss": "DeepSupervisionWrapper(\n (loss): DC_and_CE_loss(\n (ce): RobustCrossEntropyLoss()\n (dc): MemoryEfficientSoftDiceLoss()\n )\n)", - "lr_scheduler": "", + "lr_scheduler": "", "my_init_kwargs": "{'plans': {'dataset_name': 'Dataset001_CT', 'plans_name': 'nnUNetResEncUNetLPlans', 'original_median_spacing_after_transp': [0.70001220703125, 0.482421875, 0.482421875], 'original_median_shape_after_transp': [206, 365, 286], 'image_reader_writer': 'SimpleITKIO', 'transpose_forward': [0, 1, 2], 'transpose_backward': [0, 1, 2], 'configurations': {'2d': {'data_identifier': 'nnUNetPlans_2d', 'preprocessor_name': 'DefaultPreprocessor', 'batch_size': 70, 'patch_size': [384, 320], 'median_image_size_in_voxels': [352.0, 281.0], 'spacing': [0.482421875, 0.482421875], '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': 7, 'features_per_stage': [32, 64, 128, 256, 512, 512, 512], 'conv_op': 'torch.nn.modules.conv.Conv2d', 'kernel_sizes': [[3, 3], [3, 3], [3, 3], [3, 3], [3, 3], [3, 3], [3, 3]], 'strides': [[1, 1], [2, 2], [2, 2], [2, 2], [2, 2], [2, 2], [2, 2]], 'n_blocks_per_stage': [1, 3, 4, 6, 6, 6, 6], 'n_conv_per_stage_decoder': [1, 1, 1, 1, 1, 1], 'conv_bias': True, 'norm_op': 'torch.nn.modules.instancenorm.InstanceNorm2d', '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': True}, '3d_fullres': {'data_identifier': 'nnUNetPlans_3d_fullres', 'preprocessor_name': 'DefaultPreprocessor', 'batch_size': 2, 'patch_size': [128, 256, 192], 'median_image_size_in_voxels': [214.0, 352.0, 281.0], 'spacing': [0.70001220703125, 0.482421875, 0.482421875], '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': 6, 'features_per_stage': [32, 64, 128, 256, 320, 320], 'conv_op': 'torch.nn.modules.conv.Conv3d', 'kernel_sizes': [[3, 3, 3], [3, 3, 3], [3, 3, 3], [3, 3, 3], [3, 3, 3], [3, 3, 3]], 'strides': [[1, 1, 1], [2, 2, 2], [2, 2, 2], [2, 2, 2], [2, 2, 2], [2, 2, 2]], 'n_blocks_per_stage': [1, 3, 4, 6, 6, 6], 'n_conv_per_stage_decoder': [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}}, 'experiment_planner_used': 'nnUNetPlannerResEncL', 'label_manager': 'LabelManager', 'foreground_intensity_properties_per_channel': {'0': {'max': 1813.9854736328125, 'mean': 211.27015686035156, 'median': 192.95703125, 'min': -91.01630401611328, 'percentile_00_5': 66.00397491455078, 'percentile_99_5': 618.0081787109375, 'std': 107.36483764648438}}}, 'configuration': '3d_fullres', 'fold': 1, 'dataset_json': {'channel_names': {'0': 'CT'}, 'labels': {'background': 0, 'BA': 1, 'R-P1P2': 2, 'L-P1P2': 3, 'R-ICA-C6-C7': 4, 'R-M1': 5, 'L-ICA-C6-C7': 6, 'L-M1': 7, 'R-Pcom': 8, 'L-Pcom': 9, 'Acom': 10, 'R-A1A2': 11, 'L-A1A2': 12, 'R-A3': 13, 'L-A3': 14, '3rd-A2': 15, '3rd-A3': 16, 'R-M2': 17, 'R-M3': 18, 'L-M2': 19, 'L-M3': 20, 'R-P3P4': 21, 'L-P3P4': 22, 'R-VA': 23, 'L-VA': 24, 'R-SCA': 25, 'L-SCA': 26, 'R-AICA': 27, 'L-AICA': 28, 'R-PICA': 29, 'L-PICA': 30, 'R-AChA': 31, 'L-AChA': 32, 'R-OA': 33, 'L-OA': 34, 'R-ICA-C1-C5': 35, 'L-ICA-C1-C5': 36}, 'numTraining': 25, 'file_ending': '.nii.gz'}, 'device': device(type='cuda')}", "network": "ResidualEncoderUNet", "num_epochs": "1000", "num_input_channels": "1", "num_iterations_per_epoch": "250", "num_val_iterations_per_epoch": "50", - "optimizer": "SGD (\nParameter Group 0\n dampening: 0\n differentiable: False\n foreach: None\n fused: None\n initial_lr: 0.01\n lr: 0.008648416647233837\n maximize: False\n momentum: 0.99\n nesterov: True\n weight_decay: 3e-05\n)", + "optimizer": "SGD (\nParameter Group 0\n dampening: 0\n differentiable: False\n foreach: None\n fused: None\n initial_lr: 0.01\n lr: 0.00726350460441003\n maximize: False\n momentum: 0.99\n nesterov: True\n weight_decay: 3e-05\n)", "output_folder": "/marimo/TopBrain2026/nnUNet_results/Dataset001_CT/nnUNetTrainer__nnUNetResEncUNetLPlans__3d_fullres/fold_1", "output_folder_base": "/marimo/TopBrain2026/nnUNet_results/Dataset001_CT/nnUNetTrainer__nnUNetResEncUNetLPlans__3d_fullres", "oversample_foreground_percent": "0.33", diff --git a/nnUNet_Large/fold_1/nnUNet_results/Dataset001_CT/nnUNetTrainer__nnUNetResEncUNetLPlans__3d_fullres/fold_1/progress.png b/nnUNet_Large/fold_1/nnUNet_results/Dataset001_CT/nnUNetTrainer__nnUNetResEncUNetLPlans__3d_fullres/fold_1/progress.png index 579489c41fb186e021f2026251078a85f3904ae9..111df885686067f44dc3e3d4ff51281c2805a608 100644 --- a/nnUNet_Large/fold_1/nnUNet_results/Dataset001_CT/nnUNetTrainer__nnUNetResEncUNetLPlans__3d_fullres/fold_1/progress.png +++ b/nnUNet_Large/fold_1/nnUNet_results/Dataset001_CT/nnUNetTrainer__nnUNetResEncUNetLPlans__3d_fullres/fold_1/progress.png @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:17a9a8861b3704d7d25ab70dbe084707f3df75dbcdfa7e3ea3f8a4c2d12899ea -size 1267151 +oid sha256:1dd472e7c08b4de4969c06dfa173a74dcb8d79d5f760e8d19f4d694e3f2af262 +size 1503404 diff --git a/nnUNet_Large/fold_1/nnUNet_results/Dataset001_CT/nnUNetTrainer__nnUNetResEncUNetLPlans__3d_fullres/fold_1/training_log_2026_10_17_18_04_00.txt b/nnUNet_Large/fold_1/nnUNet_results/Dataset001_CT/nnUNetTrainer__nnUNetResEncUNetLPlans__3d_fullres/fold_1/training_log_2026_10_17_18_04_00.txt new file mode 100644 index 0000000000000000000000000000000000000000..81dc11efd1ff3f1a52c85a3a3472c3140001c062 --- /dev/null +++ b/nnUNet_Large/fold_1/nnUNet_results/Dataset001_CT/nnUNetTrainer__nnUNetResEncUNetLPlans__3d_fullres/fold_1/training_log_2026_10_17_18_04_00.txt @@ -0,0 +1,1082 @@ + +####################################################################### +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. +####################################################################### + +2026-10-17 18:04:04.758965: do_dummy_2d_data_aug: False +2026-10-17 18:04:04.760270: Creating new 5-fold cross-validation split... +2026-10-17 18:04:04.796447: Desired fold for training: 1 +2026-10-17 18:04:04.796610: This split has 20 training and 5 validation cases. + +This is the configuration used by this training: +Configuration name: 3d_fullres + {'data_identifier': 'nnUNetPlans_3d_fullres', 'preprocessor_name': 'DefaultPreprocessor', 'batch_size': 2, 'patch_size': [128, 256, 192], 'median_image_size_in_voxels': [214.0, 352.0, 281.0], 'spacing': [0.70001220703125, 0.482421875, 0.482421875], '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': 6, 'features_per_stage': [32, 64, 128, 256, 320, 320], 'conv_op': 'torch.nn.modules.conv.Conv3d', 'kernel_sizes': [[3, 3, 3], [3, 3, 3], [3, 3, 3], [3, 3, 3], [3, 3, 3], [3, 3, 3]], 'strides': [[1, 1, 1], [2, 2, 2], [2, 2, 2], [2, 2, 2], [2, 2, 2], [2, 2, 2]], 'n_blocks_per_stage': [1, 3, 4, 6, 6, 6], 'n_conv_per_stage_decoder': [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} + +These are the global plan.json settings: + {'dataset_name': 'Dataset001_CT', 'plans_name': 'nnUNetResEncUNetLPlans', 'original_median_spacing_after_transp': [0.70001220703125, 0.482421875, 0.482421875], 'original_median_shape_after_transp': [206, 365, 286], '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': 1813.9854736328125, 'mean': 211.27015686035156, 'median': 192.95703125, 'min': -91.01630401611328, 'percentile_00_5': 66.00397491455078, 'percentile_99_5': 618.0081787109375, 'std': 107.36483764648438}}} + +2026-10-17 18:04:10.091159: Unable to plot network architecture: +2026-10-17 18:04:10.091688: No module named 'hiddenlayer' +2026-10-17 18:04:10.137426: +2026-10-17 18:04:10.138452: Epoch 300 +2026-10-17 18:04:10.139274: Current learning rate: 0.00725 +2026-10-17 18:09:01.326229: train_loss -0.6901 +2026-10-17 18:09:01.326790: val_loss -0.5391 +2026-10-17 18:09:01.327741: Pseudo dice [np.float32(0.8704), np.float32(0.6303), np.float32(0.5907), np.float32(0.7354), np.float32(0.5252), np.float32(0.7129), np.float32(0.5732), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.6513), np.float32(0.6106), np.float32(0.6329), np.float32(0.6276), np.float32(0.0), np.float32(0.0), np.float32(0.5915), np.float32(0.4698), np.float32(0.5347), np.float32(0.4274), np.float32(0.5057), np.float32(0.4624), np.float32(0.6886), np.float32(0.6783), np.float32(0.4694), np.float32(0.4666), np.float32(0.0), np.float32(0.0), np.float32(0.4216), np.float32(0.467), np.float32(0.0), np.float32(0.0), np.float32(0.6047), np.float32(0.6165), np.float32(0.6773), np.float32(0.6519)] +2026-10-17 18:09:01.328124: Epoch time: 291.19 s +2026-10-17 18:09:01.328306: Yayy! New best EMA pseudo Dice: 0.4088999927043915 +2026-10-17 18:09:02.932089: +2026-10-17 18:09:02.932832: Epoch 301 +2026-10-17 18:09:02.933175: Current learning rate: 0.00724 +2026-10-17 18:13:30.311464: train_loss -0.6974 +2026-10-17 18:13:30.312112: val_loss -0.55 +2026-10-17 18:13:30.312655: Pseudo dice [np.float32(0.8576), np.float32(0.5995), np.float32(0.6041), np.float32(0.748), np.float32(0.5893), np.float32(0.7659), np.float32(0.6467), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.658), np.float32(0.6298), np.float32(0.5689), np.float32(0.5835), np.float32(0.0), np.float32(0.0), np.float32(0.647), np.float32(0.5126), np.float32(0.6241), np.float32(0.5079), np.float32(0.6193), np.float32(0.5339), np.float32(0.6208), np.float32(0.6045), np.float32(0.4325), np.float32(0.4605), np.float32(0.0), np.float32(0.0), np.float32(0.4837), np.float32(0.4741), np.float32(0.0), np.float32(0.0), np.float32(0.6328), np.float32(0.6249), np.float32(0.7042), np.float32(0.7117)] +2026-10-17 18:13:30.313469: Epoch time: 267.38 s +2026-10-17 18:13:30.313954: Yayy! New best EMA pseudo Dice: 0.41370001435279846 +2026-10-17 18:13:32.099941: +2026-10-17 18:13:32.100680: Epoch 302 +2026-10-17 18:13:32.101169: Current learning rate: 0.00724 +2026-10-17 18:17:30.435527: train_loss -0.6991 +2026-10-17 18:17:30.436032: val_loss -0.4123 +2026-10-17 18:17:30.436638: Pseudo dice [np.float32(0.8359), np.float32(0.3727), np.float32(0.3748), np.float32(0.4699), np.float32(0.3233), np.float32(0.4629), np.float32(0.3754), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.4203), np.float32(0.414), np.float32(0.3974), np.float32(0.4101), np.float32(0.0), np.float32(0.0), np.float32(0.3265), np.float32(0.2075), np.float32(0.3935), np.float32(0.3457), np.float32(0.1733), np.float32(0.3116), np.float32(0.4862), np.float32(0.4354), np.float32(0.3147), np.float32(0.3937), np.float32(0.0), np.float32(0.0), np.float32(0.1509), np.float32(0.3271), np.float32(0.0), np.float32(0.0), np.float32(0.3531), np.float32(0.3546), np.float32(0.4284), np.float32(0.4601)] +2026-10-17 18:17:30.437436: Epoch time: 238.34 s +2026-10-17 18:17:31.326963: +2026-10-17 18:17:31.327681: Epoch 303 +2026-10-17 18:17:31.328073: Current learning rate: 0.00723 +2026-10-17 18:21:22.198743: train_loss -0.715 +2026-10-17 18:21:22.199424: val_loss -0.5983 +2026-10-17 18:21:22.199872: Pseudo dice [np.float32(0.8866), np.float32(0.6574), np.float32(0.657), np.float32(0.8136), np.float32(0.6592), np.float32(0.8009), np.float32(0.7036), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.7091), np.float32(0.7133), np.float32(0.6247), np.float32(0.6451), np.float32(0.0), np.float32(0.0), np.float32(0.6841), np.float32(0.5309), np.float32(0.6677), np.float32(0.5425), np.float32(0.651), np.float32(0.5694), np.float32(0.7256), np.float32(0.6542), np.float32(0.5014), np.float32(0.5507), np.float32(0.0), np.float32(0.0), np.float32(0.4217), np.float32(0.4834), np.float32(0.0), np.float32(0.0), np.float32(0.7335), np.float32(0.6961), np.float32(0.7721), np.float32(0.7569)] +2026-10-17 18:21:22.200284: Epoch time: 230.87 s +2026-10-17 18:21:23.112840: +2026-10-17 18:21:23.113240: Epoch 304 +2026-10-17 18:21:23.113773: Current learning rate: 0.00722 +2026-10-17 18:25:05.122167: train_loss -0.7217 +2026-10-17 18:25:05.122886: val_loss -0.5018 +2026-10-17 18:25:05.123496: Pseudo dice [np.float32(0.8608), np.float32(0.5459), np.float32(0.5559), np.float32(0.677), np.float32(0.5), np.float32(0.6595), np.float32(0.5301), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.5957), np.float32(0.5754), np.float32(0.6002), np.float32(0.5908), np.float32(0.0), np.float32(0.0), np.float32(0.5612), np.float32(0.4659), np.float32(0.5539), np.float32(0.4275), np.float32(0.5236), np.float32(0.4528), np.float32(0.303), np.float32(0.392), np.float32(0.4297), np.float32(0.4607), np.float32(0.0), np.float32(0.0), np.float32(0.4126), np.float32(0.4819), np.float32(0.0), np.float32(0.0), np.float32(0.5735), np.float32(0.5608), np.float32(0.5829), np.float32(0.5484)] +2026-10-17 18:25:05.124043: Epoch time: 222.01 s +2026-10-17 18:25:06.048341: +2026-10-17 18:25:06.049133: Epoch 305 +2026-10-17 18:25:06.049547: Current learning rate: 0.00721 +2026-10-17 18:29:25.542953: train_loss -0.7221 +2026-10-17 18:29:25.543489: val_loss -0.5377 +2026-10-17 18:29:25.543970: Pseudo dice [np.float32(0.8502), np.float32(0.5757), np.float32(0.5824), np.float32(0.7022), np.float32(0.5571), np.float32(0.7058), np.float32(0.6092), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.6192), np.float32(0.6126), np.float32(0.5922), np.float32(0.5928), np.float32(0.0), np.float32(0.0), np.float32(0.6154), np.float32(0.5026), np.float32(0.5815), np.float32(0.5216), np.float32(0.6062), np.float32(0.4902), np.float32(0.6068), np.float32(0.6676), np.float32(0.4896), np.float32(0.4811), np.float32(0.0), np.float32(0.0), np.float32(0.4052), np.float32(0.5223), np.float32(0.0), np.float32(0.0), np.float32(0.5384), np.float32(0.6143), np.float32(0.6588), np.float32(0.6512)] +2026-10-17 18:29:25.544694: Epoch time: 259.5 s +2026-10-17 18:29:26.688088: +2026-10-17 18:29:26.688787: Epoch 306 +2026-10-17 18:29:26.689093: Current learning rate: 0.0072 +2026-10-17 18:33:14.328257: train_loss -0.7257 +2026-10-17 18:33:14.329096: val_loss -0.5258 +2026-10-17 18:33:14.329604: Pseudo dice [np.float32(0.8877), np.float32(0.5749), np.float32(0.5707), np.float32(0.6932), np.float32(0.5242), np.float32(0.6681), np.float32(0.5782), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.5773), np.float32(0.5766), np.float32(0.514), np.float32(0.5044), np.float32(0.0), np.float32(0.0), np.float32(0.5835), np.float32(0.4484), np.float32(0.5605), np.float32(0.5046), np.float32(0.589), np.float32(0.5074), np.float32(0.5987), np.float32(0.5711), np.float32(0.5285), np.float32(0.5242), np.float32(0.0), np.float32(0.0), np.float32(0.2547), np.float32(0.5126), np.float32(0.0), np.float32(0.0), np.float32(0.5978), np.float32(0.6108), np.float32(0.657), np.float32(0.6614)] +2026-10-17 18:33:14.329930: Epoch time: 227.64 s +2026-10-17 18:33:14.330338: Yayy! New best EMA pseudo Dice: 0.4142000079154968 +2026-10-17 18:33:17.894834: +2026-10-17 18:33:17.895413: Epoch 307 +2026-10-17 18:33:17.895800: Current learning rate: 0.00719 +2026-10-17 18:37:29.231273: train_loss -0.7228 +2026-10-17 18:37:29.232683: val_loss -0.5699 +2026-10-17 18:37:29.233464: Pseudo dice [np.float32(0.8847), np.float32(0.6773), np.float32(0.6395), np.float32(0.7625), np.float32(0.6307), np.float32(0.7638), np.float32(0.6592), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.6664), np.float32(0.6637), np.float32(0.6025), np.float32(0.6248), np.float32(0.0), np.float32(0.0), np.float32(0.6792), np.float32(0.5303), np.float32(0.6553), np.float32(0.5543), np.float32(0.6456), np.float32(0.5056), np.float32(0.6976), np.float32(0.6711), np.float32(0.5005), np.float32(0.503), np.float32(0.0), np.float32(0.0), np.float32(0.4372), np.float32(0.4533), np.float32(0.0), np.float32(0.0), np.float32(0.6945), np.float32(0.6751), np.float32(0.7538), np.float32(0.7488)] +2026-10-17 18:37:29.233893: Epoch time: 251.34 s +2026-10-17 18:37:29.234101: Yayy! New best EMA pseudo Dice: 0.42080000042915344 +2026-10-17 18:37:31.230077: +2026-10-17 18:37:31.230718: Epoch 308 +2026-10-17 18:37:31.231056: Current learning rate: 0.00718 +2026-10-17 18:41:35.234753: train_loss -0.7295 +2026-10-17 18:41:35.235504: val_loss -0.3731 +2026-10-17 18:41:35.236043: Pseudo dice [np.float32(0.8867), np.float32(0.3006), np.float32(0.3371), np.float32(0.3922), np.float32(0.3584), np.float32(0.3943), np.float32(0.3558), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.2884), np.float32(0.2979), np.float32(0.4126), np.float32(0.3617), np.float32(0.0), np.float32(0.0), np.float32(0.3352), np.float32(0.2752), np.float32(0.2726), np.float32(0.2537), np.float32(0.2706), np.float32(0.2295), np.float32(0.5278), np.float32(0.4792), np.float32(0.1711), np.float32(0.1718), np.float32(0.0), np.float32(0.0), np.float32(0.0988), np.float32(0.209), np.float32(0.0), np.float32(0.0), np.float32(0.3632), np.float32(0.4097), np.float32(0.4111), np.float32(0.3858)] +2026-10-17 18:41:35.236620: Epoch time: 244.01 s +2026-10-17 18:41:36.185451: +2026-10-17 18:41:36.186316: Epoch 309 +2026-10-17 18:41:36.186709: Current learning rate: 0.00717 +2026-10-17 18:45:46.810089: train_loss -0.7243 +2026-10-17 18:45:46.810974: val_loss -0.6352 +2026-10-17 18:45:46.811472: Pseudo dice [np.float32(0.833), np.float32(0.7694), np.float32(0.7268), np.float32(0.8572), np.float32(0.7284), np.float32(0.8683), np.float32(0.7676), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.7431), np.float32(0.7598), np.float32(0.6358), np.float32(0.6645), np.float32(0.0), np.float32(0.0), np.float32(0.7342), np.float32(0.6117), np.float32(0.734), np.float32(0.6301), np.float32(0.682), np.float32(0.537), np.float32(0.7977), np.float32(0.8475), np.float32(0.4457), np.float32(0.5734), np.float32(0.0), np.float32(0.0), np.float32(0.4627), np.float32(0.4522), np.float32(0.0), np.float32(0.0), np.float32(0.828), np.float32(0.7582), np.float32(0.8873), np.float32(0.8932)] +2026-10-17 18:45:46.811748: Epoch time: 250.63 s +2026-10-17 18:45:47.671006: +2026-10-17 18:45:47.671469: Epoch 310 +2026-10-17 18:45:47.671833: Current learning rate: 0.00716 +2026-10-17 18:49:53.122742: train_loss -0.7074 +2026-10-17 18:49:53.123823: val_loss -0.4622 +2026-10-17 18:49:53.124362: Pseudo dice [np.float32(0.8495), np.float32(0.496), np.float32(0.5331), np.float32(0.6737), np.float32(0.4634), np.float32(0.6448), np.float32(0.5091), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.5845), np.float32(0.539), np.float32(0.5593), np.float32(0.5286), np.float32(0.0), np.float32(0.0), np.float32(0.5481), np.float32(0.4606), np.float32(0.5189), np.float32(0.4311), np.float32(0.585), np.float32(0.4865), np.float32(0.3218), np.float32(0.3684), np.float32(0.4004), np.float32(0.3258), np.float32(0.0), np.float32(0.0), np.float32(0.524), np.float32(0.4437), np.float32(0.0), np.float32(0.0), np.float32(0.5896), np.float32(0.5592), np.float32(0.6023), np.float32(0.5536)] +2026-10-17 18:49:53.124887: Epoch time: 245.45 s +2026-10-17 18:49:53.951943: +2026-10-17 18:49:53.952527: Epoch 311 +2026-10-17 18:49:53.952793: Current learning rate: 0.00715 +2026-10-17 18:54:01.585487: train_loss -0.7077 +2026-10-17 18:54:01.586245: val_loss -0.5242 +2026-10-17 18:54:01.587265: Pseudo dice [np.float32(0.8864), np.float32(0.6215), np.float32(0.5979), np.float32(0.7222), np.float32(0.5405), np.float32(0.6959), np.float32(0.5626), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.6073), np.float32(0.619), np.float32(0.5869), np.float32(0.5933), np.float32(0.0), np.float32(0.0), np.float32(0.6134), np.float32(0.5024), np.float32(0.5456), np.float32(0.4278), np.float32(0.5454), np.float32(0.4594), np.float32(0.6445), np.float32(0.6064), np.float32(0.511), np.float32(0.6011), np.float32(0.0), np.float32(0.0), np.float32(0.3863), np.float32(0.529), np.float32(0.0), np.float32(0.0), np.float32(0.6254), np.float32(0.6268), np.float32(0.7213), np.float32(0.7367)] +2026-10-17 18:54:01.587778: Epoch time: 247.64 s +2026-10-17 18:54:02.466906: +2026-10-17 18:54:02.467591: Epoch 312 +2026-10-17 18:54:02.467884: Current learning rate: 0.00714 +2026-10-17 18:58:10.108229: train_loss -0.7209 +2026-10-17 18:58:10.108783: val_loss -0.4845 +2026-10-17 18:58:10.109771: Pseudo dice [np.float32(0.8676), np.float32(0.5531), np.float32(0.5421), np.float32(0.6127), np.float32(0.4916), np.float32(0.563), np.float32(0.4745), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.5356), np.float32(0.5384), np.float32(0.5429), np.float32(0.5661), np.float32(0.0), np.float32(0.0), np.float32(0.5376), np.float32(0.4728), np.float32(0.4383), np.float32(0.375), np.float32(0.4154), np.float32(0.3633), np.float32(0.6675), np.float32(0.6406), np.float32(0.3335), np.float32(0.4137), np.float32(0.0), np.float32(0.0), np.float32(0.2489), np.float32(0.2955), np.float32(0.0), np.float32(0.0), np.float32(0.4803), np.float32(0.3844), np.float32(0.5804), np.float32(0.5612)] +2026-10-17 18:58:10.110490: Epoch time: 247.64 s +2026-10-17 18:58:10.996875: +2026-10-17 18:58:10.997612: Epoch 313 +2026-10-17 18:58:10.997999: Current learning rate: 0.00713 +2026-10-17 19:02:39.011042: train_loss -0.7173 +2026-10-17 19:02:39.011682: val_loss -0.5834 +2026-10-17 19:02:39.012167: Pseudo dice [np.float32(0.8878), np.float32(0.6523), np.float32(0.6382), np.float32(0.7685), np.float32(0.5957), np.float32(0.7566), np.float32(0.6419), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.6426), np.float32(0.6662), np.float32(0.6381), np.float32(0.6111), np.float32(0.0), np.float32(0.0), np.float32(0.6063), np.float32(0.5297), np.float32(0.5607), np.float32(0.4962), np.float32(0.5414), np.float32(0.3713), np.float32(0.7926), np.float32(0.7566), np.float32(0.4178), np.float32(0.5318), np.float32(0.0), np.float32(0.0), np.float32(0.3424), np.float32(0.4932), np.float32(0.0), np.float32(0.0), np.float32(0.6745), np.float32(0.6332), np.float32(0.7727), np.float32(0.7457)] +2026-10-17 19:02:39.012917: Epoch time: 268.02 s +2026-10-17 19:02:39.925553: +2026-10-17 19:02:39.926160: Epoch 314 +2026-10-17 19:02:39.926471: Current learning rate: 0.00712 +2026-10-17 19:06:32.615257: train_loss -0.7169 +2026-10-17 19:06:32.616016: val_loss -0.434 +2026-10-17 19:06:32.617105: Pseudo dice [np.float32(0.8783), np.float32(0.4778), np.float32(0.4591), np.float32(0.5794), np.float32(0.3989), np.float32(0.5228), np.float32(0.4067), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.5117), np.float32(0.4837), np.float32(0.4944), np.float32(0.4256), np.float32(0.0), np.float32(0.0), np.float32(0.4656), np.float32(0.3439), np.float32(0.3699), np.float32(0.2728), np.float32(0.3189), np.float32(0.3213), np.float32(0.533), np.float32(0.4948), np.float32(0.3189), np.float32(0.3453), np.float32(0.0), np.float32(0.0), np.float32(0.0898), np.float32(0.3448), np.float32(0.0), np.float32(0.0), np.float32(0.4324), np.float32(0.4254), np.float32(0.4932), np.float32(0.4697)] +2026-10-17 19:06:32.617637: Epoch time: 232.69 s +2026-10-17 19:06:33.542013: +2026-10-17 19:06:33.542757: Epoch 315 +2026-10-17 19:06:33.543120: Current learning rate: 0.00711 +2026-10-17 19:10:49.199330: train_loss -0.6937 +2026-10-17 19:10:49.199850: val_loss -0.1923 +2026-10-17 19:10:49.200730: Pseudo dice [np.float32(0.879), np.float32(0.0771), np.float32(0.0544), np.float32(0.1108), np.float32(0.0206), np.float32(0.0444), np.float32(0.0535), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.0692), np.float32(0.0522), np.float32(0.0824), np.float32(0.0918), np.float32(0.0), np.float32(0.0), np.float32(0.033), np.float32(0.0252), np.float32(0.0461), np.float32(0.0703), np.float32(0.0008), np.float32(0.0869), np.float32(0.102), np.float32(0.1216), np.float32(0.0449), np.float32(0.065), np.float32(0.0), np.float32(0.0), np.float32(0.0325), np.float32(0.0111), np.float32(0.0), np.float32(0.0), np.float32(0.014), np.float32(0.0464), np.float32(0.0625), np.float32(0.0398)] +2026-10-17 19:10:49.201692: Epoch time: 255.66 s +2026-10-17 19:10:50.142119: +2026-10-17 19:10:50.142943: Epoch 316 +2026-10-17 19:10:50.143219: Current learning rate: 0.0071 +2026-10-17 19:15:15.962033: train_loss -0.6439 +2026-10-17 19:15:15.962931: val_loss -0.4281 +2026-10-17 19:15:15.963622: Pseudo dice [np.float32(0.8615), np.float32(0.4365), np.float32(0.4251), np.float32(0.4935), np.float32(0.4275), np.float32(0.505), np.float32(0.4566), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.3948), np.float32(0.4592), np.float32(0.354), np.float32(0.4158), np.float32(0.0), np.float32(0.0), np.float32(0.4747), np.float32(0.4038), np.float32(0.3838), np.float32(0.4204), np.float32(0.3729), np.float32(0.4167), np.float32(0.6284), np.float32(0.565), np.float32(0.4004), np.float32(0.4016), np.float32(0.0), np.float32(0.0), np.float32(0.0738), np.float32(0.2354), np.float32(0.0), np.float32(0.0), np.float32(0.416), np.float32(0.37), np.float32(0.4767), np.float32(0.4849)] +2026-10-17 19:15:15.964143: Epoch time: 265.82 s +2026-10-17 19:15:16.886501: +2026-10-17 19:15:16.887291: Epoch 317 +2026-10-17 19:15:16.887817: Current learning rate: 0.0071 +2026-10-17 19:19:20.900991: train_loss -0.6701 +2026-10-17 19:19:20.902556: val_loss -0.4959 +2026-10-17 19:19:20.903182: Pseudo dice [np.float32(0.8587), np.float32(0.5605), np.float32(0.5068), np.float32(0.6726), np.float32(0.5127), np.float32(0.6435), np.float32(0.5766), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.5701), np.float32(0.5712), np.float32(0.5225), np.float32(0.6017), np.float32(0.0), np.float32(0.0), np.float32(0.596), np.float32(0.5017), np.float32(0.5386), np.float32(0.4907), np.float32(0.5951), np.float32(0.4669), np.float32(0.5279), np.float32(0.4936), np.float32(0.4455), np.float32(0.4013), np.float32(0.0), np.float32(0.0), np.float32(0.4003), np.float32(0.5882), np.float32(0.0), np.float32(0.0), np.float32(0.6595), np.float32(0.6125), np.float32(0.6355), np.float32(0.6465)] +2026-10-17 19:19:20.903639: Epoch time: 244.02 s +2026-10-17 19:19:21.827725: +2026-10-17 19:19:21.828540: Epoch 318 +2026-10-17 19:19:21.829005: Current learning rate: 0.00709 +2026-10-17 19:23:16.111266: train_loss -0.6947 +2026-10-17 19:23:16.112441: val_loss -0.3493 +2026-10-17 19:23:16.112992: Pseudo dice [np.float32(0.8507), np.float32(0.2858), np.float32(0.317), np.float32(0.403), np.float32(0.2696), np.float32(0.3094), np.float32(0.319), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.3767), np.float32(0.3517), np.float32(0.4071), np.float32(0.4398), np.float32(0.0), np.float32(0.0), np.float32(0.3235), np.float32(0.2936), np.float32(0.3159), np.float32(0.2673), np.float32(0.2761), np.float32(0.3072), np.float32(0.2155), np.float32(0.3422), np.float32(0.2574), np.float32(0.2182), np.float32(0.0), np.float32(0.0), np.float32(0.4059), np.float32(0.0902), np.float32(0.0), np.float32(0.0), np.float32(0.2543), np.float32(0.3229), np.float32(0.2951), np.float32(0.2727)] +2026-10-17 19:23:16.113248: Epoch time: 234.29 s +2026-10-17 19:23:17.019730: +2026-10-17 19:23:17.020515: Epoch 319 +2026-10-17 19:23:17.020880: Current learning rate: 0.00708 +2026-10-17 19:27:13.554448: train_loss -0.6666 +2026-10-17 19:27:13.555364: val_loss -0.4536 +2026-10-17 19:27:13.555693: Pseudo dice [np.float32(0.8589), np.float32(0.4866), np.float32(0.508), np.float32(0.5888), np.float32(0.4245), np.float32(0.5527), np.float32(0.4729), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.5385), np.float32(0.4677), np.float32(0.5355), np.float32(0.4613), np.float32(0.0), np.float32(0.0), np.float32(0.516), np.float32(0.3994), np.float32(0.4614), np.float32(0.3843), np.float32(0.458), np.float32(0.4695), np.float32(0.2702), np.float32(0.4453), np.float32(0.4678), np.float32(0.3639), np.float32(0.0), np.float32(0.0), np.float32(0.4633), np.float32(0.5129), np.float32(0.0), np.float32(0.0), np.float32(0.4975), np.float32(0.4689), np.float32(0.4896), np.float32(0.4823)] +2026-10-17 19:27:13.556115: Epoch time: 236.54 s +2026-10-17 19:27:14.487790: +2026-10-17 19:27:14.488498: Epoch 320 +2026-10-17 19:27:14.488731: Current learning rate: 0.00707 +2026-10-17 19:31:21.398797: train_loss -0.6283 +2026-10-17 19:31:21.399830: val_loss -0.4721 +2026-10-17 19:31:21.400317: Pseudo dice [np.float32(0.8779), np.float32(0.5192), np.float32(0.518), np.float32(0.5833), np.float32(0.4776), np.float32(0.5864), np.float32(0.5226), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.5303), np.float32(0.5332), np.float32(0.4798), np.float32(0.548), np.float32(0.0), np.float32(0.0), np.float32(0.5369), np.float32(0.4482), np.float32(0.4402), np.float32(0.3891), np.float32(0.4282), np.float32(0.4389), np.float32(0.4206), np.float32(0.506), np.float32(0.3971), np.float32(0.427), np.float32(0.0), np.float32(0.0), np.float32(0.2894), np.float32(0.3991), np.float32(0.0), np.float32(0.0), np.float32(0.4583), np.float32(0.5281), np.float32(0.4972), np.float32(0.5003)] +2026-10-17 19:31:21.400705: Epoch time: 246.91 s +2026-10-17 19:31:22.252931: +2026-10-17 19:31:22.253657: Epoch 321 +2026-10-17 19:31:22.253877: Current learning rate: 0.00706 +2026-10-17 19:35:25.805760: train_loss -0.6271 +2026-10-17 19:35:25.806342: val_loss -0.5395 +2026-10-17 19:35:25.806871: Pseudo dice [np.float32(0.8864), np.float32(0.5947), np.float32(0.5512), np.float32(0.6616), np.float32(0.5667), np.float32(0.6361), np.float32(0.5839), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.5125), np.float32(0.5257), np.float32(0.4871), np.float32(0.5487), np.float32(0.0), np.float32(0.0), np.float32(0.5174), np.float32(0.425), np.float32(0.4992), np.float32(0.4437), np.float32(0.3668), np.float32(0.3956), np.float32(0.7412), np.float32(0.8027), np.float32(0.5003), np.float32(0.5542), np.float32(0.0), np.float32(0.0), np.float32(0.1562), np.float32(0.4405), np.float32(0.0), np.float32(0.0), np.float32(0.5283), np.float32(0.4729), np.float32(0.6447), np.float32(0.6575)] +2026-10-17 19:35:25.807910: Epoch time: 243.56 s +2026-10-17 19:35:26.695886: +2026-10-17 19:35:26.696599: Epoch 322 +2026-10-17 19:35:26.697037: Current learning rate: 0.00705 +2026-10-17 19:39:25.670122: train_loss -0.6718 +2026-10-17 19:39:25.671210: val_loss -0.5174 +2026-10-17 19:39:25.671849: Pseudo dice [np.float32(0.8508), np.float32(0.5799), np.float32(0.5998), np.float32(0.7308), np.float32(0.5612), np.float32(0.7023), np.float32(0.6076), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.6322), np.float32(0.6148), np.float32(0.587), np.float32(0.5794), np.float32(0.0), np.float32(0.0), np.float32(0.6188), np.float32(0.5074), np.float32(0.6336), np.float32(0.5177), np.float32(0.5462), np.float32(0.5154), np.float32(0.472), np.float32(0.4788), np.float32(0.4738), np.float32(0.3266), np.float32(0.0), np.float32(0.0), np.float32(0.5256), np.float32(0.4308), np.float32(0.0), np.float32(0.0), np.float32(0.6738), np.float32(0.6096), np.float32(0.6854), np.float32(0.6585)] +2026-10-17 19:39:25.672163: Epoch time: 238.98 s +2026-10-17 19:39:26.616650: +2026-10-17 19:39:26.617256: Epoch 323 +2026-10-17 19:39:26.617610: Current learning rate: 0.00704 +2026-10-17 19:43:22.846908: train_loss -0.6495 +2026-10-17 19:43:22.847704: val_loss -0.4728 +2026-10-17 19:43:22.848376: Pseudo dice [np.float32(0.8858), np.float32(0.414), np.float32(0.4097), np.float32(0.6423), np.float32(0.5251), np.float32(0.659), np.float32(0.551), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.5196), np.float32(0.4845), np.float32(0.5058), np.float32(0.4581), np.float32(0.0), np.float32(0.0), np.float32(0.5458), np.float32(0.4181), np.float32(0.5454), np.float32(0.4093), np.float32(0.4134), np.float32(0.2115), np.float32(0.4691), np.float32(0.5533), np.float32(0.1908), np.float32(0.124), np.float32(0.0), np.float32(0.0), np.float32(0.3212), np.float32(0.2962), np.float32(0.0), np.float32(0.0), np.float32(0.5662), np.float32(0.5559), np.float32(0.6251), np.float32(0.6061)] +2026-10-17 19:43:22.848863: Epoch time: 236.23 s +2026-10-17 19:43:24.061659: +2026-10-17 19:43:24.062344: Epoch 324 +2026-10-17 19:43:24.062818: Current learning rate: 0.00703 +2026-10-17 19:47:24.939366: train_loss -0.6692 +2026-10-17 19:47:24.940108: val_loss -0.4966 +2026-10-17 19:47:24.941239: Pseudo dice [np.float32(0.858), np.float32(0.5163), np.float32(0.4923), np.float32(0.5563), np.float32(0.461), np.float32(0.5434), np.float32(0.4824), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.507), np.float32(0.5018), np.float32(0.5372), np.float32(0.5623), np.float32(0.0), np.float32(0.0), np.float32(0.495), np.float32(0.4627), np.float32(0.4376), np.float32(0.4094), np.float32(0.3105), np.float32(0.3567), np.float32(0.5418), np.float32(0.6459), np.float32(0.341), np.float32(0.3613), np.float32(0.0), np.float32(0.0), np.float32(0.1138), np.float32(0.2966), np.float32(0.0), np.float32(0.0), np.float32(0.4189), np.float32(0.397), np.float32(0.4737), np.float32(0.4861)] +2026-10-17 19:47:24.941720: Epoch time: 240.88 s +2026-10-17 19:47:25.856969: +2026-10-17 19:47:25.857777: Epoch 325 +2026-10-17 19:47:25.858054: Current learning rate: 0.00702 +2026-10-17 19:51:39.298628: train_loss -0.7127 +2026-10-17 19:51:39.299240: val_loss -0.5196 +2026-10-17 19:51:39.300087: Pseudo dice [np.float32(0.8829), np.float32(0.6211), np.float32(0.5643), np.float32(0.6264), np.float32(0.534), np.float32(0.6258), np.float32(0.5314), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.5406), np.float32(0.5757), np.float32(0.5171), np.float32(0.5993), np.float32(0.0), np.float32(0.0), np.float32(0.5796), np.float32(0.5077), np.float32(0.4768), np.float32(0.4406), np.float32(0.458), np.float32(0.4118), np.float32(0.7354), np.float32(0.7208), np.float32(0.4065), np.float32(0.4811), np.float32(0.0), np.float32(0.0), np.float32(0.106), np.float32(0.4902), np.float32(0.0), np.float32(0.0), np.float32(0.5054), np.float32(0.4788), np.float32(0.6527), np.float32(0.6663)] +2026-10-17 19:51:39.300574: Epoch time: 253.44 s +2026-10-17 19:51:40.280726: +2026-10-17 19:51:40.281454: Epoch 326 +2026-10-17 19:51:40.281807: Current learning rate: 0.00701 +2026-10-17 19:55:31.317731: train_loss -0.7199 +2026-10-17 19:55:31.318713: val_loss -0.2946 +2026-10-17 19:55:31.319392: Pseudo dice [np.float32(0.8416), np.float32(0.2496), np.float32(0.2616), np.float32(0.3023), np.float32(0.2288), np.float32(0.2681), np.float32(0.2272), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.2395), np.float32(0.2415), np.float32(0.3246), np.float32(0.2826), np.float32(0.0), np.float32(0.0), np.float32(0.3102), np.float32(0.3012), np.float32(0.1619), np.float32(0.1198), np.float32(0.1069), np.float32(0.1847), np.float32(0.1877), np.float32(0.3628), np.float32(0.147), np.float32(0.1822), np.float32(0.0), np.float32(0.0), np.float32(0.0437), np.float32(0.207), np.float32(0.0), np.float32(0.0), np.float32(0.1799), np.float32(0.1888), np.float32(0.2157), np.float32(0.2008)] +2026-10-17 19:55:31.319937: Epoch time: 231.04 s +2026-10-17 19:55:32.174464: +2026-10-17 19:55:32.175233: Epoch 327 +2026-10-17 19:55:32.175541: Current learning rate: 0.007 +2026-10-17 19:59:37.466233: train_loss -0.7163 +2026-10-17 19:59:37.467188: val_loss -0.5494 +2026-10-17 19:59:37.467724: Pseudo dice [np.float32(0.8814), np.float32(0.5814), np.float32(0.5855), np.float32(0.724), np.float32(0.5775), np.float32(0.702), np.float32(0.6003), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.6223), np.float32(0.614), np.float32(0.5879), np.float32(0.624), np.float32(0.0), np.float32(0.0), np.float32(0.6209), np.float32(0.5125), np.float32(0.5223), np.float32(0.449), np.float32(0.4072), np.float32(0.4108), np.float32(0.7165), np.float32(0.7516), np.float32(0.4172), np.float32(0.4897), np.float32(0.0), np.float32(0.0), np.float32(0.377), np.float32(0.5153), np.float32(0.0), np.float32(0.0), np.float32(0.5836), np.float32(0.6054), np.float32(0.71), np.float32(0.696)] +2026-10-17 19:59:37.468202: Epoch time: 245.29 s +2026-10-17 19:59:38.363906: +2026-10-17 19:59:38.364557: Epoch 328 +2026-10-17 19:59:38.364904: Current learning rate: 0.00699 +2026-10-17 20:03:39.384348: train_loss -0.6972 +2026-10-17 20:03:39.385130: val_loss -0.5772 +2026-10-17 20:03:39.386055: Pseudo dice [np.float32(0.8345), np.float32(0.6262), np.float32(0.6163), np.float32(0.793), np.float32(0.5913), np.float32(0.7697), np.float32(0.6273), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.6414), np.float32(0.6395), np.float32(0.612), np.float32(0.602), np.float32(0.0), np.float32(0.0), np.float32(0.6246), np.float32(0.5435), np.float32(0.5643), np.float32(0.4569), np.float32(0.5076), np.float32(0.4866), np.float32(0.6908), np.float32(0.7335), np.float32(0.4841), np.float32(0.5102), np.float32(0.0), np.float32(0.0), np.float32(0.4616), np.float32(0.573), np.float32(0.0), np.float32(0.0), np.float32(0.6778), np.float32(0.6756), np.float32(0.7685), np.float32(0.7453)] +2026-10-17 20:03:39.386493: Epoch time: 241.02 s +2026-10-17 20:03:40.290714: +2026-10-17 20:03:40.291307: Epoch 329 +2026-10-17 20:03:40.291687: Current learning rate: 0.00698 +2026-10-17 20:07:52.345316: train_loss -0.6881 +2026-10-17 20:07:52.346193: val_loss -0.3431 +2026-10-17 20:07:52.347024: Pseudo dice [np.float32(0.8778), np.float32(0.3063), np.float32(0.3491), np.float32(0.3842), np.float32(0.2476), np.float32(0.3566), np.float32(0.279), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.2934), np.float32(0.2303), np.float32(0.4509), np.float32(0.4117), np.float32(0.0), np.float32(0.0), np.float32(0.3071), np.float32(0.2502), np.float32(0.3241), np.float32(0.2381), np.float32(0.2319), np.float32(0.2598), np.float32(0.3329), np.float32(0.393), np.float32(0.2457), np.float32(0.194), np.float32(0.0), np.float32(0.0), np.float32(0.251), np.float32(0.3945), np.float32(0.0), np.float32(0.0), np.float32(0.2753), np.float32(0.3543), np.float32(0.3813), np.float32(0.3362)] +2026-10-17 20:07:52.347416: Epoch time: 252.06 s +2026-10-17 20:07:53.281807: +2026-10-17 20:07:53.282604: Epoch 330 +2026-10-17 20:07:53.282960: Current learning rate: 0.00697 +2026-10-17 20:11:45.968194: train_loss -0.6971 +2026-10-17 20:11:45.968835: val_loss -0.5342 +2026-10-17 20:11:45.969296: Pseudo dice [np.float32(0.8331), np.float32(0.6126), np.float32(0.591), np.float32(0.7246), np.float32(0.585), np.float32(0.6628), np.float32(0.6089), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.5835), np.float32(0.5461), np.float32(0.5641), np.float32(0.5545), np.float32(0.0), np.float32(0.0), np.float32(0.6482), np.float32(0.5233), np.float32(0.6051), np.float32(0.4861), np.float32(0.52), np.float32(0.5166), np.float32(0.594), np.float32(0.6998), np.float32(0.4719), np.float32(0.4029), np.float32(0.0), np.float32(0.0), np.float32(0.4814), np.float32(0.5631), np.float32(0.0), np.float32(0.0), np.float32(0.6022), np.float32(0.5591), np.float32(0.7026), np.float32(0.6803)] +2026-10-17 20:11:45.969610: Epoch time: 232.69 s +2026-10-17 20:11:46.894142: +2026-10-17 20:11:46.894932: Epoch 331 +2026-10-17 20:11:46.895388: Current learning rate: 0.00696 +2026-10-17 20:16:08.430566: train_loss -0.7039 +2026-10-17 20:16:08.431584: val_loss -0.4582 +2026-10-17 20:16:08.432187: Pseudo dice [np.float32(0.8546), np.float32(0.4378), np.float32(0.476), np.float32(0.588), np.float32(0.4392), np.float32(0.5714), np.float32(0.4702), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.5402), np.float32(0.5328), np.float32(0.5823), np.float32(0.5558), np.float32(0.0), np.float32(0.0), np.float32(0.5386), np.float32(0.4423), np.float32(0.4206), np.float32(0.3203), np.float32(0.3788), np.float32(0.4043), np.float32(0.3323), np.float32(0.4158), np.float32(0.4191), np.float32(0.3978), np.float32(0.0), np.float32(0.0), np.float32(0.3065), np.float32(0.4663), np.float32(0.0), np.float32(0.0), np.float32(0.4813), np.float32(0.4553), np.float32(0.514), np.float32(0.4808)] +2026-10-17 20:16:08.432819: Epoch time: 261.54 s +2026-10-17 20:16:09.333406: +2026-10-17 20:16:09.334219: Epoch 332 +2026-10-17 20:16:09.334702: Current learning rate: 0.00696 +2026-10-17 20:20:08.693039: train_loss -0.7089 +2026-10-17 20:20:08.693811: val_loss -0.3136 +2026-10-17 20:20:08.694617: Pseudo dice [np.float32(0.8635), np.float32(0.2574), np.float32(0.2608), np.float32(0.3364), np.float32(0.2241), np.float32(0.2927), np.float32(0.2662), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.3089), np.float32(0.2752), np.float32(0.3221), np.float32(0.3049), np.float32(0.0), np.float32(0.0), np.float32(0.2592), np.float32(0.1959), np.float32(0.2967), np.float32(0.2427), np.float32(0.1061), np.float32(0.199), np.float32(0.357), np.float32(0.3845), np.float32(0.2823), np.float32(0.3155), np.float32(0.0), np.float32(0.0), np.float32(0.1419), np.float32(0.4018), np.float32(0.0), np.float32(0.0), np.float32(0.2001), np.float32(0.1695), np.float32(0.2504), np.float32(0.2808)] +2026-10-17 20:20:08.695112: Epoch time: 239.36 s +2026-10-17 20:20:09.613096: +2026-10-17 20:20:09.613822: Epoch 333 +2026-10-17 20:20:09.614222: Current learning rate: 0.00695 +2026-10-17 20:24:08.001990: train_loss -0.7034 +2026-10-17 20:24:08.002839: val_loss -0.4675 +2026-10-17 20:24:08.003674: Pseudo dice [np.float32(0.8369), np.float32(0.54), np.float32(0.5048), np.float32(0.6257), np.float32(0.4682), np.float32(0.5665), np.float32(0.5055), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.5689), np.float32(0.5602), np.float32(0.5323), np.float32(0.4983), np.float32(0.0), np.float32(0.0), np.float32(0.5447), np.float32(0.427), np.float32(0.477), np.float32(0.3658), np.float32(0.4036), np.float32(0.4533), np.float32(0.38), np.float32(0.4974), np.float32(0.5256), np.float32(0.4275), np.float32(0.0), np.float32(0.0), np.float32(0.1545), np.float32(0.3119), np.float32(0.0), np.float32(0.0), np.float32(0.4958), np.float32(0.4754), np.float32(0.5638), np.float32(0.5317)] +2026-10-17 20:24:08.004039: Epoch time: 238.39 s +2026-10-17 20:24:08.937242: +2026-10-17 20:24:08.938009: Epoch 334 +2026-10-17 20:24:08.938509: Current learning rate: 0.00694 +2026-10-17 20:28:30.024248: train_loss -0.7225 +2026-10-17 20:28:30.025116: val_loss -0.4345 +2026-10-17 20:28:30.025707: Pseudo dice [np.float32(0.8504), np.float32(0.4739), np.float32(0.4348), np.float32(0.5654), np.float32(0.3893), np.float32(0.5175), np.float32(0.4212), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.4829), np.float32(0.4687), np.float32(0.4818), np.float32(0.4573), np.float32(0.0), np.float32(0.0), np.float32(0.4743), np.float32(0.3888), np.float32(0.4343), np.float32(0.3462), np.float32(0.3779), np.float32(0.3439), np.float32(0.243), np.float32(0.3813), np.float32(0.4949), np.float32(0.4668), np.float32(0.0), np.float32(0.0), np.float32(0.2978), np.float32(0.5066), np.float32(0.0), np.float32(0.0), np.float32(0.4885), np.float32(0.4553), np.float32(0.4647), np.float32(0.4649)] +2026-10-17 20:28:30.025999: Epoch time: 261.09 s +2026-10-17 20:28:30.908101: +2026-10-17 20:28:30.909021: Epoch 335 +2026-10-17 20:28:30.909462: Current learning rate: 0.00693 +2026-10-17 20:32:30.307551: train_loss -0.7261 +2026-10-17 20:32:30.308118: val_loss -0.5076 +2026-10-17 20:32:30.308818: Pseudo dice [np.float32(0.8512), np.float32(0.5826), np.float32(0.5524), np.float32(0.7158), np.float32(0.5295), np.float32(0.6643), np.float32(0.5511), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.6163), np.float32(0.5938), np.float32(0.5919), np.float32(0.5871), np.float32(0.0), np.float32(0.0), np.float32(0.5843), np.float32(0.4784), np.float32(0.5114), np.float32(0.4349), np.float32(0.5588), np.float32(0.4662), np.float32(0.5854), np.float32(0.6473), np.float32(0.5009), np.float32(0.4824), np.float32(0.0), np.float32(0.0), np.float32(0.3572), np.float32(0.5588), np.float32(0.0), np.float32(0.0), np.float32(0.6048), np.float32(0.5753), np.float32(0.6746), np.float32(0.635)] +2026-10-17 20:32:30.309512: Epoch time: 239.4 s +2026-10-17 20:32:31.337086: +2026-10-17 20:32:31.337624: Epoch 336 +2026-10-17 20:32:31.337933: Current learning rate: 0.00692 +2026-10-17 20:36:56.132705: train_loss -0.7269 +2026-10-17 20:36:56.133547: val_loss -0.4813 +2026-10-17 20:36:56.134434: Pseudo dice [np.float32(0.8651), np.float32(0.5016), np.float32(0.479), np.float32(0.6264), np.float32(0.4622), np.float32(0.5806), np.float32(0.4833), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.5749), np.float32(0.549), np.float32(0.5446), np.float32(0.4712), np.float32(0.0), np.float32(0.0), np.float32(0.5503), np.float32(0.4445), np.float32(0.4791), np.float32(0.3702), np.float32(0.4123), np.float32(0.4081), np.float32(0.4573), np.float32(0.5296), np.float32(0.4437), np.float32(0.4208), np.float32(0.0), np.float32(0.0), np.float32(0.2681), np.float32(0.5443), np.float32(0.0), np.float32(0.0), np.float32(0.5329), np.float32(0.4862), np.float32(0.5846), np.float32(0.5473)] +2026-10-17 20:36:56.134888: Epoch time: 264.8 s +2026-10-17 20:36:57.094594: +2026-10-17 20:36:57.095473: Epoch 337 +2026-10-17 20:36:57.095862: Current learning rate: 0.00691 +2026-10-17 20:41:22.033782: train_loss -0.7285 +2026-10-17 20:41:22.034873: val_loss -0.55 +2026-10-17 20:41:22.035456: Pseudo dice [np.float32(0.8523), np.float32(0.653), np.float32(0.6073), np.float32(0.7576), np.float32(0.5611), np.float32(0.7329), np.float32(0.6083), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.6545), np.float32(0.6554), np.float32(0.6075), np.float32(0.6365), np.float32(0.0), np.float32(0.0), np.float32(0.6398), np.float32(0.5754), np.float32(0.602), np.float32(0.4923), np.float32(0.5852), np.float32(0.504), np.float32(0.4803), np.float32(0.5372), np.float32(0.529), np.float32(0.3999), np.float32(0.0), np.float32(0.0), np.float32(0.2733), np.float32(0.5394), np.float32(0.0), np.float32(0.0), np.float32(0.6643), np.float32(0.5766), np.float32(0.686), np.float32(0.6646)] +2026-10-17 20:41:22.035911: Epoch time: 264.94 s +2026-10-17 20:41:22.940889: +2026-10-17 20:41:22.941510: Epoch 338 +2026-10-17 20:41:22.941879: Current learning rate: 0.0069 +2026-10-17 20:45:57.319517: train_loss -0.7108 +2026-10-17 20:45:57.320098: val_loss -0.4522 +2026-10-17 20:45:57.320661: Pseudo dice [np.float32(0.8721), np.float32(0.4875), np.float32(0.4727), np.float32(0.6207), np.float32(0.4448), np.float32(0.5987), np.float32(0.4693), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.5398), np.float32(0.5437), np.float32(0.548), np.float32(0.5903), np.float32(0.0), np.float32(0.0), np.float32(0.4838), np.float32(0.3892), np.float32(0.5131), np.float32(0.3864), np.float32(0.457), np.float32(0.397), np.float32(0.2919), np.float32(0.4764), np.float32(0.4983), np.float32(0.327), np.float32(0.0), np.float32(0.0), np.float32(0.3451), np.float32(0.4699), np.float32(0.0), np.float32(0.0), np.float32(0.5309), np.float32(0.5101), np.float32(0.5698), np.float32(0.5476)] +2026-10-17 20:45:57.321123: Epoch time: 274.38 s +2026-10-17 20:45:58.237525: +2026-10-17 20:45:58.238191: Epoch 339 +2026-10-17 20:45:58.238520: Current learning rate: 0.00689 +2026-10-17 20:49:50.382328: train_loss -0.7238 +2026-10-17 20:49:50.383208: val_loss -0.3543 +2026-10-17 20:49:50.384047: Pseudo dice [np.float32(0.8708), np.float32(0.3464), np.float32(0.3098), np.float32(0.3724), np.float32(0.2865), np.float32(0.3523), np.float32(0.3147), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.3215), np.float32(0.3483), np.float32(0.3681), np.float32(0.3906), np.float32(0.0), np.float32(0.0), np.float32(0.3804), np.float32(0.3368), np.float32(0.285), np.float32(0.2321), np.float32(0.2492), np.float32(0.2208), np.float32(0.3524), np.float32(0.4191), np.float32(0.2675), np.float32(0.3921), np.float32(0.0), np.float32(0.0), np.float32(0.1298), np.float32(0.3504), np.float32(0.0), np.float32(0.0), np.float32(0.2206), np.float32(0.202), np.float32(0.3084), np.float32(0.3295)] +2026-10-17 20:49:50.384622: Epoch time: 232.15 s +2026-10-17 20:49:51.313785: +2026-10-17 20:49:51.314281: Epoch 340 +2026-10-17 20:49:51.314638: Current learning rate: 0.00688 +2026-10-17 20:53:45.021122: train_loss -0.7307 +2026-10-17 20:53:45.021666: val_loss -0.3934 +2026-10-17 20:53:45.022361: Pseudo dice [np.float32(0.8339), np.float32(0.435), np.float32(0.3972), np.float32(0.5386), np.float32(0.3604), np.float32(0.4536), np.float32(0.3455), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.4546), np.float32(0.4165), np.float32(0.4375), np.float32(0.4488), np.float32(0.0), np.float32(0.0), np.float32(0.4199), np.float32(0.3787), np.float32(0.3039), np.float32(0.2778), np.float32(0.2826), np.float32(0.2863), np.float32(0.4744), np.float32(0.5307), np.float32(0.3453), np.float32(0.4007), np.float32(0.0), np.float32(0.0), np.float32(0.0968), np.float32(0.4154), np.float32(0.0), np.float32(0.0), np.float32(0.3308), np.float32(0.3081), np.float32(0.4319), np.float32(0.4273)] +2026-10-17 20:53:45.023228: Epoch time: 233.71 s +2026-10-17 20:53:45.973124: +2026-10-17 20:53:45.973810: Epoch 341 +2026-10-17 20:53:45.974167: Current learning rate: 0.00687 +2026-10-17 20:57:54.950524: train_loss -0.734 +2026-10-17 20:57:54.951174: val_loss -0.6345 +2026-10-17 20:57:54.951549: Pseudo dice [np.float32(0.8853), np.float32(0.7784), np.float32(0.7332), np.float32(0.8503), np.float32(0.7216), np.float32(0.8487), np.float32(0.7945), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.7467), np.float32(0.7381), np.float32(0.6549), np.float32(0.6677), np.float32(0.0), np.float32(0.0), np.float32(0.726), np.float32(0.6064), np.float32(0.7145), np.float32(0.6215), np.float32(0.6505), np.float32(0.4919), np.float32(0.8587), np.float32(0.8374), np.float32(0.5261), np.float32(0.5644), np.float32(0.0), np.float32(0.0), np.float32(0.2313), np.float32(0.4578), np.float32(0.0), np.float32(0.0), np.float32(0.7752), np.float32(0.7682), np.float32(0.8518), np.float32(0.847)] +2026-10-17 20:57:54.952273: Epoch time: 248.98 s +2026-10-17 20:57:55.816765: +2026-10-17 20:57:55.817329: Epoch 342 +2026-10-17 20:57:55.817736: Current learning rate: 0.00686 +2026-10-17 21:01:47.053841: train_loss -0.7329 +2026-10-17 21:01:47.054385: val_loss -0.6127 +2026-10-17 21:01:47.055235: Pseudo dice [np.float32(0.8791), np.float32(0.7399), np.float32(0.7152), np.float32(0.8196), np.float32(0.6911), np.float32(0.8126), np.float32(0.7661), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.6819), np.float32(0.6811), np.float32(0.6707), np.float32(0.7067), np.float32(0.0), np.float32(0.0), np.float32(0.7046), np.float32(0.6211), np.float32(0.6883), np.float32(0.5999), np.float32(0.5645), np.float32(0.5295), np.float32(0.709), np.float32(0.7343), np.float32(0.5267), np.float32(0.4996), np.float32(0.0), np.float32(0.0), np.float32(0.3379), np.float32(0.5582), np.float32(0.0), np.float32(0.0), np.float32(0.6677), np.float32(0.6817), np.float32(0.8049), np.float32(0.8193)] +2026-10-17 21:01:47.056145: Epoch time: 231.24 s +2026-10-17 21:01:48.002101: +2026-10-17 21:01:48.002697: Epoch 343 +2026-10-17 21:01:48.003131: Current learning rate: 0.00685 +2026-10-17 21:05:52.653968: train_loss -0.722 +2026-10-17 21:05:52.654812: val_loss -0.2548 +2026-10-17 21:05:52.655324: Pseudo dice [np.float32(0.8865), np.float32(0.2133), np.float32(0.2212), np.float32(0.2628), np.float32(0.1469), np.float32(0.2477), np.float32(0.1593), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.2173), np.float32(0.2639), np.float32(0.2824), np.float32(0.2183), np.float32(0.0), np.float32(0.0), np.float32(0.2107), np.float32(0.2002), np.float32(0.1217), np.float32(0.1152), np.float32(0.0924), np.float32(0.1379), np.float32(0.244), np.float32(0.3762), np.float32(0.2095), np.float32(0.1717), np.float32(0.0), np.float32(0.0), np.float32(0.1471), np.float32(0.3302), np.float32(0.0), np.float32(0.0), np.float32(0.1843), np.float32(0.1768), np.float32(0.2199), np.float32(0.2377)] +2026-10-17 21:05:52.655861: Epoch time: 244.65 s +2026-10-17 21:05:53.603258: +2026-10-17 21:05:53.603860: Epoch 344 +2026-10-17 21:05:53.604203: Current learning rate: 0.00684 +2026-10-17 21:09:33.777887: train_loss -0.7016 +2026-10-17 21:09:33.778790: val_loss -0.4731 +2026-10-17 21:09:33.779304: Pseudo dice [np.float32(0.8589), np.float32(0.5164), np.float32(0.5569), np.float32(0.634), np.float32(0.5062), np.float32(0.6144), np.float32(0.511), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.5333), np.float32(0.4481), np.float32(0.5747), np.float32(0.5268), np.float32(0.0), np.float32(0.0), np.float32(0.4699), np.float32(0.3726), np.float32(0.5202), np.float32(0.4224), np.float32(0.4665), np.float32(0.3765), np.float32(0.5818), np.float32(0.5638), np.float32(0.3084), np.float32(0.1103), np.float32(0.0), np.float32(0.0), np.float32(0.6475), np.float32(0.4242), np.float32(0.0), np.float32(0.0), np.float32(0.5387), np.float32(0.4904), np.float32(0.6217), np.float32(0.5762)] +2026-10-17 21:09:33.779781: Epoch time: 220.18 s +2026-10-17 21:09:35.006449: +2026-10-17 21:09:35.007123: Epoch 345 +2026-10-17 21:09:35.007643: Current learning rate: 0.00683 +2026-10-17 21:14:06.075094: train_loss -0.6868 +2026-10-17 21:14:06.075857: val_loss -0.621 +2026-10-17 21:14:06.076341: Pseudo dice [np.float32(0.8534), np.float32(0.7322), np.float32(0.6992), np.float32(0.8456), np.float32(0.7123), np.float32(0.8564), np.float32(0.7798), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.7416), np.float32(0.7419), np.float32(0.6705), np.float32(0.7452), np.float32(0.0), np.float32(0.0), np.float32(0.7129), np.float32(0.5849), np.float32(0.7131), np.float32(0.6047), np.float32(0.5893), np.float32(0.4922), np.float32(0.8392), np.float32(0.8361), np.float32(0.4963), np.float32(0.5672), np.float32(0.0), np.float32(0.0), np.float32(0.3878), np.float32(0.4736), np.float32(0.0), np.float32(0.0), np.float32(0.7092), np.float32(0.7338), np.float32(0.8631), np.float32(0.8649)] +2026-10-17 21:14:06.076600: Epoch time: 271.07 s +2026-10-17 21:14:07.066680: +2026-10-17 21:14:07.067465: Epoch 346 +2026-10-17 21:14:07.067969: Current learning rate: 0.00682 +2026-10-17 21:18:04.750284: train_loss -0.6821 +2026-10-17 21:18:04.751000: val_loss -0.442 +2026-10-17 21:18:04.751964: Pseudo dice [np.float32(0.8536), np.float32(0.4762), np.float32(0.4413), np.float32(0.5478), np.float32(0.4276), np.float32(0.5242), np.float32(0.4293), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.4767), np.float32(0.4452), np.float32(0.4457), np.float32(0.4929), np.float32(0.0), np.float32(0.0), np.float32(0.4219), np.float32(0.2816), np.float32(0.3893), np.float32(0.3303), np.float32(0.2658), np.float32(0.2203), np.float32(0.534), np.float32(0.472), np.float32(0.3611), np.float32(0.4021), np.float32(0.0), np.float32(0.0), np.float32(0.0848), np.float32(0.41), np.float32(0.0), np.float32(0.0), np.float32(0.373), np.float32(0.319), np.float32(0.4584), np.float32(0.4545)] +2026-10-17 21:18:04.752554: Epoch time: 237.69 s +2026-10-17 21:18:05.719949: +2026-10-17 21:18:05.720682: Epoch 347 +2026-10-17 21:18:05.721138: Current learning rate: 0.00681 +2026-10-17 21:22:30.488972: train_loss -0.6768 +2026-10-17 21:22:30.489888: val_loss -0.1776 +2026-10-17 21:22:30.490387: Pseudo dice [np.float32(0.8598), np.float32(0.0701), np.float32(0.0533), np.float32(0.0301), np.float32(0.0073), np.float32(0.0387), np.float32(0.0361), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.0534), np.float32(0.0363), np.float32(0.1728), np.float32(0.0745), np.float32(0.0), np.float32(0.0), np.float32(0.0166), np.float32(0.0185), np.float32(0.0622), np.float32(0.0685), np.float32(0.0124), np.float32(0.126), np.float32(0.1842), np.float32(0.0419), np.float32(0.0708), np.float32(0.0387), np.float32(0.0), np.float32(0.0), np.float32(0.0441), np.float32(0.0414), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.0289), np.float32(0.0147), np.float32(0.0237)] +2026-10-17 21:22:30.490836: Epoch time: 264.77 s +2026-10-17 21:22:31.423605: +2026-10-17 21:22:31.424442: Epoch 348 +2026-10-17 21:22:31.424810: Current learning rate: 0.0068 +2026-10-17 21:26:20.995022: train_loss -0.6906 +2026-10-17 21:26:20.995951: val_loss -0.3918 +2026-10-17 21:26:20.996492: Pseudo dice [np.float32(0.8408), np.float32(0.4313), np.float32(0.4132), np.float32(0.5546), np.float32(0.4743), np.float32(0.4495), np.float32(0.3529), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.4362), np.float32(0.4021), np.float32(0.4134), np.float32(0.4579), np.float32(0.0), np.float32(0.0), np.float32(0.4782), np.float32(0.376), np.float32(0.3993), np.float32(0.3296), np.float32(0.4328), np.float32(0.3789), np.float32(0.5312), np.float32(0.4049), np.float32(0.3274), np.float32(0.1358), np.float32(0.0), np.float32(0.0), np.float32(0.4941), np.float32(0.3929), np.float32(0.0), np.float32(0.0), np.float32(0.5134), np.float32(0.511), np.float32(0.5418), np.float32(0.3618)] +2026-10-17 21:26:20.996772: Epoch time: 229.57 s +2026-10-17 21:26:21.932571: +2026-10-17 21:26:21.933212: Epoch 349 +2026-10-17 21:26:21.933512: Current learning rate: 0.0068 +2026-10-17 21:30:27.836531: train_loss -0.6885 +2026-10-17 21:30:27.837120: val_loss -0.4493 +2026-10-17 21:30:27.837755: Pseudo dice [np.float32(0.8519), np.float32(0.4721), np.float32(0.4462), np.float32(0.5202), np.float32(0.4464), np.float32(0.4943), np.float32(0.4979), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.4418), np.float32(0.437), np.float32(0.439), np.float32(0.4099), np.float32(0.0), np.float32(0.0), np.float32(0.5097), np.float32(0.4099), np.float32(0.4194), np.float32(0.4002), np.float32(0.3924), np.float32(0.3609), np.float32(0.5545), np.float32(0.4875), np.float32(0.4107), np.float32(0.3919), np.float32(0.0), np.float32(0.0), np.float32(0.1625), np.float32(0.4547), np.float32(0.0), np.float32(0.0), np.float32(0.3849), np.float32(0.3798), np.float32(0.4921), np.float32(0.5317)] +2026-10-17 21:30:27.838818: Epoch time: 245.91 s +2026-10-17 21:30:29.439012: +2026-10-17 21:30:29.439817: Epoch 350 +2026-10-17 21:30:29.440031: Current learning rate: 0.00679 +2026-10-17 21:34:50.018269: train_loss -0.7175 +2026-10-17 21:34:50.018894: val_loss -0.3144 +2026-10-17 21:34:50.019768: Pseudo dice [np.float32(0.8553), np.float32(0.2616), np.float32(0.2785), np.float32(0.3058), np.float32(0.2527), np.float32(0.2931), np.float32(0.2565), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.2315), np.float32(0.2333), np.float32(0.2992), np.float32(0.2887), np.float32(0.0), np.float32(0.0), np.float32(0.2723), np.float32(0.1943), np.float32(0.2371), np.float32(0.2195), np.float32(0.1434), np.float32(0.1599), np.float32(0.3966), np.float32(0.2966), np.float32(0.2265), np.float32(0.2911), np.float32(0.0), np.float32(0.0), np.float32(0.1212), np.float32(0.274), np.float32(0.0), np.float32(0.0), np.float32(0.1771), np.float32(0.2044), np.float32(0.2819), np.float32(0.2901)] +2026-10-17 21:34:50.020009: Epoch time: 260.58 s +2026-10-17 21:34:50.978344: +2026-10-17 21:34:50.979090: Epoch 351 +2026-10-17 21:34:50.979611: Current learning rate: 0.00678 +2026-10-17 21:38:46.514326: train_loss -0.7202 +2026-10-17 21:38:46.514964: val_loss -0.457 +2026-10-17 21:38:46.515421: Pseudo dice [np.float32(0.8676), np.float32(0.4881), np.float32(0.4635), np.float32(0.5493), np.float32(0.4793), np.float32(0.5579), np.float32(0.5033), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.4438), np.float32(0.4696), np.float32(0.4388), np.float32(0.4361), np.float32(0.0), np.float32(0.0), np.float32(0.5183), np.float32(0.4076), np.float32(0.4664), np.float32(0.4279), np.float32(0.3584), np.float32(0.3561), np.float32(0.6003), np.float32(0.5331), np.float32(0.4425), np.float32(0.4444), np.float32(0.0), np.float32(0.0), np.float32(0.1866), np.float32(0.5488), np.float32(0.0), np.float32(0.0), np.float32(0.4458), np.float32(0.4653), np.float32(0.5628), np.float32(0.5911)] +2026-10-17 21:38:46.516541: Epoch time: 235.54 s +2026-10-17 21:38:47.497231: +2026-10-17 21:38:47.497922: Epoch 352 +2026-10-17 21:38:47.498513: Current learning rate: 0.00677 +2026-10-17 21:43:04.224396: train_loss -0.7152 +2026-10-17 21:43:04.225241: val_loss -0.6058 +2026-10-17 21:43:04.226071: Pseudo dice [np.float32(0.8849), np.float32(0.7002), np.float32(0.7096), np.float32(0.8228), np.float32(0.6837), np.float32(0.815), np.float32(0.7459), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.7013), np.float32(0.6935), np.float32(0.6574), np.float32(0.6909), np.float32(0.0), np.float32(0.0), np.float32(0.6927), np.float32(0.5719), np.float32(0.6718), np.float32(0.5573), np.float32(0.5997), np.float32(0.5623), np.float32(0.7631), np.float32(0.7513), np.float32(0.527), np.float32(0.5413), np.float32(0.0), np.float32(0.0), np.float32(0.5059), np.float32(0.5215), np.float32(0.0), np.float32(0.0), np.float32(0.7574), np.float32(0.7408), np.float32(0.8381), np.float32(0.8406)] +2026-10-17 21:43:04.226449: Epoch time: 256.73 s +2026-10-17 21:43:05.197459: +2026-10-17 21:43:05.198272: Epoch 353 +2026-10-17 21:43:05.198706: Current learning rate: 0.00676 +2026-10-17 21:47:08.091209: train_loss -0.7111 +2026-10-17 21:47:08.091851: val_loss -0.5702 +2026-10-17 21:47:08.093475: Pseudo dice [np.float32(0.8763), np.float32(0.6766), np.float32(0.6622), np.float32(0.8192), np.float32(0.6313), np.float32(0.8003), np.float32(0.7159), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.7005), np.float32(0.6676), np.float32(0.6054), np.float32(0.6751), np.float32(0.0), np.float32(0.0), np.float32(0.6569), np.float32(0.5473), np.float32(0.6544), np.float32(0.5395), np.float32(0.5627), np.float32(0.4755), np.float32(0.7948), np.float32(0.7982), np.float32(0.423), np.float32(0.3697), np.float32(0.0), np.float32(0.0), np.float32(0.5105), np.float32(0.3718), np.float32(0.0), np.float32(0.0), np.float32(0.7772), np.float32(0.7614), np.float32(0.8421), np.float32(0.8435)] +2026-10-17 21:47:08.094045: Epoch time: 242.9 s +2026-10-17 21:47:09.072482: +2026-10-17 21:47:09.073248: Epoch 354 +2026-10-17 21:47:09.073677: Current learning rate: 0.00675 +2026-10-17 21:51:23.958381: train_loss -0.658 +2026-10-17 21:51:23.959114: val_loss -0.6245 +2026-10-17 21:51:23.959795: Pseudo dice [np.float32(0.8462), np.float32(0.7883), np.float32(0.7281), np.float32(0.8687), np.float32(0.7296), np.float32(0.8675), np.float32(0.7946), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.7727), np.float32(0.7568), np.float32(0.6752), np.float32(0.7065), np.float32(0.0), np.float32(0.0), np.float32(0.704), np.float32(0.5641), np.float32(0.6982), np.float32(0.5604), np.float32(0.5827), np.float32(0.5434), np.float32(0.7923), np.float32(0.8325), np.float32(0.5786), np.float32(0.4256), np.float32(0.0), np.float32(0.0), np.float32(0.4221), np.float32(0.2694), np.float32(0.0), np.float32(0.0), np.float32(0.726), np.float32(0.6926), np.float32(0.8854), np.float32(0.8743)] +2026-10-17 21:51:23.960222: Epoch time: 254.89 s +2026-10-17 21:51:24.920352: +2026-10-17 21:51:24.921236: Epoch 355 +2026-10-17 21:51:24.921574: Current learning rate: 0.00674 +2026-10-17 21:55:31.113274: train_loss -0.6721 +2026-10-17 21:55:31.113734: val_loss -0.5117 +2026-10-17 21:55:31.114158: Pseudo dice [np.float32(0.8891), np.float32(0.5448), np.float32(0.5066), np.float32(0.6975), np.float32(0.6106), np.float32(0.6381), np.float32(0.6102), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.5542), np.float32(0.5686), np.float32(0.5235), np.float32(0.556), np.float32(0.0), np.float32(0.0), np.float32(0.5963), np.float32(0.4669), np.float32(0.5202), np.float32(0.4658), np.float32(0.3922), np.float32(0.3107), np.float32(0.7418), np.float32(0.6664), np.float32(0.4168), np.float32(0.3073), np.float32(0.0), np.float32(0.0), np.float32(0.3), np.float32(0.3646), np.float32(0.0), np.float32(0.0), np.float32(0.5723), np.float32(0.5996), np.float32(0.7131), np.float32(0.7114)] +2026-10-17 21:55:31.115103: Epoch time: 246.2 s +2026-10-17 21:55:32.086773: +2026-10-17 21:55:32.087478: Epoch 356 +2026-10-17 21:55:32.087739: Current learning rate: 0.00673 +2026-10-17 22:00:02.351767: train_loss -0.706 +2026-10-17 22:00:02.352928: val_loss -0.4991 +2026-10-17 22:00:02.353639: Pseudo dice [np.float32(0.8834), np.float32(0.5231), np.float32(0.5266), np.float32(0.6563), np.float32(0.5138), np.float32(0.6484), np.float32(0.5025), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.5493), np.float32(0.5749), np.float32(0.4585), np.float32(0.4626), np.float32(0.0), np.float32(0.0), np.float32(0.5307), np.float32(0.4195), np.float32(0.4644), np.float32(0.4199), np.float32(0.4167), np.float32(0.4195), np.float32(0.5813), np.float32(0.5871), np.float32(0.4729), np.float32(0.4444), np.float32(0.0), np.float32(0.0), np.float32(0.3201), np.float32(0.305), np.float32(0.0), np.float32(0.0), np.float32(0.5472), np.float32(0.5188), np.float32(0.6062), np.float32(0.6282)] +2026-10-17 22:00:02.353996: Epoch time: 270.27 s +2026-10-17 22:00:03.327466: +2026-10-17 22:00:03.328320: Epoch 357 +2026-10-17 22:00:03.328760: Current learning rate: 0.00672 +2026-10-17 22:04:23.945998: train_loss -0.7176 +2026-10-17 22:04:23.947252: val_loss -0.4488 +2026-10-17 22:04:23.947899: Pseudo dice [np.float32(0.8718), np.float32(0.4629), np.float32(0.5072), np.float32(0.6061), np.float32(0.5028), np.float32(0.6025), np.float32(0.5024), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.4773), np.float32(0.4293), np.float32(0.4583), np.float32(0.4613), np.float32(0.0), np.float32(0.0), np.float32(0.5168), np.float32(0.3863), np.float32(0.4716), np.float32(0.3551), np.float32(0.375), np.float32(0.4163), np.float32(0.5583), np.float32(0.4693), np.float32(0.3672), np.float32(0.1929), np.float32(0.0), np.float32(0.0), np.float32(0.465), np.float32(0.3655), np.float32(0.0), np.float32(0.0), np.float32(0.5101), np.float32(0.5535), np.float32(0.5692), np.float32(0.5567)] +2026-10-17 22:04:23.948440: Epoch time: 260.62 s +2026-10-17 22:04:24.888550: +2026-10-17 22:04:24.889310: Epoch 358 +2026-10-17 22:04:24.889658: Current learning rate: 0.00671 +2026-10-17 22:08:31.833286: train_loss -0.7109 +2026-10-17 22:08:31.834273: val_loss -0.4001 +2026-10-17 22:08:31.835023: Pseudo dice [np.float32(0.8706), np.float32(0.3915), np.float32(0.3732), np.float32(0.4883), np.float32(0.438), np.float32(0.4622), np.float32(0.3972), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.3582), np.float32(0.3804), np.float32(0.3813), np.float32(0.3999), np.float32(0.0), np.float32(0.0), np.float32(0.4002), np.float32(0.2994), np.float32(0.3345), np.float32(0.2998), np.float32(0.2564), np.float32(0.2561), np.float32(0.5382), np.float32(0.6007), np.float32(0.2387), np.float32(0.2169), np.float32(0.0), np.float32(0.0), np.float32(0.026), np.float32(0.2121), np.float32(0.0), np.float32(0.0), np.float32(0.3762), np.float32(0.3943), np.float32(0.4368), np.float32(0.4521)] +2026-10-17 22:08:31.835427: Epoch time: 246.95 s +2026-10-17 22:08:32.838653: +2026-10-17 22:08:32.839447: Epoch 359 +2026-10-17 22:08:32.839885: Current learning rate: 0.0067 +2026-10-17 22:12:48.544769: train_loss -0.6931 +2026-10-17 22:12:48.546124: val_loss -0.2897 +2026-10-17 22:12:48.546742: Pseudo dice [np.float32(0.8875), np.float32(0.2743), np.float32(0.3162), np.float32(0.3496), np.float32(0.2608), np.float32(0.2988), np.float32(0.2544), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.299), np.float32(0.2653), np.float32(0.3994), np.float32(0.3616), np.float32(0.0), np.float32(0.0), np.float32(0.2807), np.float32(0.2594), np.float32(0.2192), np.float32(0.1748), np.float32(0.1528), np.float32(0.203), np.float32(0.236), np.float32(0.285), np.float32(0.1752), np.float32(0.0886), np.float32(0.0), np.float32(0.0), np.float32(0.1143), np.float32(0.2419), np.float32(0.0), np.float32(0.0), np.float32(0.2265), np.float32(0.2569), np.float32(0.2242), np.float32(0.1985)] +2026-10-17 22:12:48.547234: Epoch time: 255.71 s +2026-10-17 22:12:49.471970: +2026-10-17 22:12:49.472704: Epoch 360 +2026-10-17 22:12:49.473150: Current learning rate: 0.00669 +2026-10-17 22:16:30.350027: train_loss -0.7162 +2026-10-17 22:16:30.350693: val_loss -0.5334 +2026-10-17 22:16:30.351498: Pseudo dice [np.float32(0.8957), np.float32(0.6325), np.float32(0.6207), np.float32(0.7224), np.float32(0.6005), np.float32(0.7021), np.float32(0.6019), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.6025), np.float32(0.5938), np.float32(0.5717), np.float32(0.6089), np.float32(0.0), np.float32(0.0), np.float32(0.6111), np.float32(0.4796), np.float32(0.5786), np.float32(0.488), np.float32(0.5379), np.float32(0.4378), np.float32(0.635), np.float32(0.5958), np.float32(0.3885), np.float32(0.327), np.float32(0.0), np.float32(0.0), np.float32(0.3925), np.float32(0.4501), np.float32(0.0), np.float32(0.0), np.float32(0.6749), np.float32(0.6635), np.float32(0.6903), np.float32(0.6644)] +2026-10-17 22:16:30.352001: Epoch time: 220.88 s +2026-10-17 22:16:31.602625: +2026-10-17 22:16:31.603232: Epoch 361 +2026-10-17 22:16:31.603585: Current learning rate: 0.00668 +2026-10-17 22:20:19.169111: train_loss -0.6988 +2026-10-17 22:20:19.170139: val_loss -0.5335 +2026-10-17 22:20:19.170835: Pseudo dice [np.float32(0.8478), np.float32(0.6056), np.float32(0.6181), np.float32(0.7655), np.float32(0.6126), np.float32(0.7496), np.float32(0.6504), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.6784), np.float32(0.6526), np.float32(0.6778), np.float32(0.6646), np.float32(0.0), np.float32(0.0), np.float32(0.6616), np.float32(0.5271), np.float32(0.6376), np.float32(0.4764), np.float32(0.4789), np.float32(0.4358), np.float32(0.6479), np.float32(0.6773), np.float32(0.3762), np.float32(0.2962), np.float32(0.0), np.float32(0.0), np.float32(0.3533), np.float32(0.4323), np.float32(0.0), np.float32(0.0), np.float32(0.7127), np.float32(0.6426), np.float32(0.7333), np.float32(0.7519)] +2026-10-17 22:20:19.171403: Epoch time: 227.57 s +2026-10-17 22:20:20.128545: +2026-10-17 22:20:20.129202: Epoch 362 +2026-10-17 22:20:20.129694: Current learning rate: 0.00667 +2026-10-17 22:24:12.541652: train_loss -0.7196 +2026-10-17 22:24:12.542187: val_loss -0.5445 +2026-10-17 22:24:12.543346: Pseudo dice [np.float32(0.8762), np.float32(0.666), np.float32(0.6441), np.float32(0.7294), np.float32(0.612), np.float32(0.7294), np.float32(0.6731), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.6155), np.float32(0.5991), np.float32(0.6156), np.float32(0.6092), np.float32(0.0), np.float32(0.0), np.float32(0.5595), np.float32(0.446), np.float32(0.5748), np.float32(0.4879), np.float32(0.5383), np.float32(0.4543), np.float32(0.6454), np.float32(0.5999), np.float32(0.4799), np.float32(0.3198), np.float32(0.0), np.float32(0.0), np.float32(0.4183), np.float32(0.36), np.float32(0.0), np.float32(0.0), np.float32(0.7212), np.float32(0.6861), np.float32(0.7139), np.float32(0.7225)] +2026-10-17 22:24:12.543847: Epoch time: 232.42 s +2026-10-17 22:24:13.438997: +2026-10-17 22:24:13.439715: Epoch 363 +2026-10-17 22:24:13.440016: Current learning rate: 0.00666 +2026-10-17 22:27:49.246141: train_loss -0.7295 +2026-10-17 22:27:49.247094: val_loss -0.5958 +2026-10-17 22:27:49.247587: Pseudo dice [np.float32(0.8897), np.float32(0.7115), np.float32(0.7049), np.float32(0.8378), np.float32(0.6737), np.float32(0.8315), np.float32(0.7517), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.6907), np.float32(0.6834), np.float32(0.6389), np.float32(0.6628), np.float32(0.0), np.float32(0.0), np.float32(0.6338), np.float32(0.5352), np.float32(0.6454), np.float32(0.5448), np.float32(0.5665), np.float32(0.5264), np.float32(0.7846), np.float32(0.6955), np.float32(0.5233), np.float32(0.4293), np.float32(0.0), np.float32(0.0), np.float32(0.4121), np.float32(0.454), np.float32(0.0), np.float32(0.0), np.float32(0.7693), np.float32(0.7165), np.float32(0.826), np.float32(0.831)] +2026-10-17 22:27:49.247940: Epoch time: 215.81 s +2026-10-17 22:27:50.117782: +2026-10-17 22:27:50.118669: Epoch 364 +2026-10-17 22:27:50.118932: Current learning rate: 0.00665 +2026-10-17 22:31:23.267245: train_loss -0.7309 +2026-10-17 22:31:23.268109: val_loss -0.6002 +2026-10-17 22:31:23.268707: Pseudo dice [np.float32(0.8566), np.float32(0.6897), np.float32(0.7101), np.float32(0.8262), np.float32(0.6768), np.float32(0.8452), np.float32(0.7535), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.6989), np.float32(0.6778), np.float32(0.6113), np.float32(0.6762), np.float32(0.0), np.float32(0.0), np.float32(0.6553), np.float32(0.5449), np.float32(0.6566), np.float32(0.5735), np.float32(0.6304), np.float32(0.4925), np.float32(0.5671), np.float32(0.5864), np.float32(0.5648), np.float32(0.4834), np.float32(0.0), np.float32(0.0), np.float32(0.453), np.float32(0.4586), np.float32(0.0), np.float32(0.0), np.float32(0.7794), np.float32(0.7247), np.float32(0.8218), np.float32(0.8335)] +2026-10-17 22:31:23.269273: Epoch time: 213.15 s +2026-10-17 22:31:24.165111: +2026-10-17 22:31:24.165849: Epoch 365 +2026-10-17 22:31:24.166198: Current learning rate: 0.00665 +2026-10-17 22:35:05.870534: train_loss -0.7322 +2026-10-17 22:35:05.871150: val_loss -0.6353 +2026-10-17 22:35:05.871644: Pseudo dice [np.float32(0.8898), np.float32(0.7984), np.float32(0.7626), np.float32(0.8429), np.float32(0.7193), np.float32(0.8745), np.float32(0.8089), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.7546), np.float32(0.7517), np.float32(0.6845), np.float32(0.6969), np.float32(0.0), np.float32(0.0), np.float32(0.7371), np.float32(0.6441), np.float32(0.7164), np.float32(0.6287), np.float32(0.6772), np.float32(0.6015), np.float32(0.801), np.float32(0.7847), np.float32(0.5998), np.float32(0.5279), np.float32(0.0), np.float32(0.0), np.float32(0.4433), np.float32(0.5447), np.float32(0.0), np.float32(0.0), np.float32(0.7872), np.float32(0.7727), np.float32(0.8846), np.float32(0.8819)] +2026-10-17 22:35:05.872614: Epoch time: 221.71 s +2026-10-17 22:35:06.830111: +2026-10-17 22:35:06.830715: Epoch 366 +2026-10-17 22:35:06.830906: Current learning rate: 0.00664 +2026-10-17 22:38:49.631479: train_loss -0.7289 +2026-10-17 22:38:49.632472: val_loss -0.6324 +2026-10-17 22:38:49.633056: Pseudo dice [np.float32(0.9062), np.float32(0.7664), np.float32(0.7317), np.float32(0.8602), np.float32(0.7263), np.float32(0.8777), np.float32(0.8253), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.754), np.float32(0.7634), np.float32(0.7051), np.float32(0.7553), np.float32(0.0), np.float32(0.0), np.float32(0.7322), np.float32(0.6222), np.float32(0.7441), np.float32(0.6342), np.float32(0.6478), np.float32(0.5797), np.float32(0.7433), np.float32(0.6875), np.float32(0.5078), np.float32(0.4902), np.float32(0.0), np.float32(0.0), np.float32(0.3775), np.float32(0.5389), np.float32(0.0), np.float32(0.0), np.float32(0.7865), np.float32(0.7563), np.float32(0.8365), np.float32(0.8401)] +2026-10-17 22:38:49.633459: Epoch time: 222.8 s +2026-10-17 22:38:49.633865: Yayy! New best EMA pseudo Dice: 0.42800000309944153 +2026-10-17 22:38:51.460315: +2026-10-17 22:38:51.460931: Epoch 367 +2026-10-17 22:38:51.461339: Current learning rate: 0.00663 +2026-10-17 22:42:21.814523: train_loss -0.7408 +2026-10-17 22:42:21.815202: val_loss -0.5959 +2026-10-17 22:42:21.815892: Pseudo dice [np.float32(0.9022), np.float32(0.7221), np.float32(0.6935), np.float32(0.8096), np.float32(0.6435), np.float32(0.8056), np.float32(0.692), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.6738), np.float32(0.6635), np.float32(0.6599), np.float32(0.6719), np.float32(0.0), np.float32(0.0), np.float32(0.6686), np.float32(0.5949), np.float32(0.6606), np.float32(0.5639), np.float32(0.6365), np.float32(0.5243), np.float32(0.6647), np.float32(0.6369), np.float32(0.5303), np.float32(0.446), np.float32(0.0), np.float32(0.0), np.float32(0.4586), np.float32(0.5144), np.float32(0.0), np.float32(0.0), np.float32(0.7097), np.float32(0.6699), np.float32(0.8076), np.float32(0.7586)] +2026-10-17 22:42:21.816398: Epoch time: 210.36 s +2026-10-17 22:42:21.816702: Yayy! New best EMA pseudo Dice: 0.43459999561309814 +2026-10-17 22:42:23.381119: +2026-10-17 22:42:23.381807: Epoch 368 +2026-10-17 22:42:23.392426: Current learning rate: 0.00662 +2026-10-17 22:45:54.518515: train_loss -0.7464 +2026-10-17 22:45:54.519517: val_loss -0.5934 +2026-10-17 22:45:54.520079: Pseudo dice [np.float32(0.8825), np.float32(0.7226), np.float32(0.6972), np.float32(0.8188), np.float32(0.6662), np.float32(0.8124), np.float32(0.7292), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.7075), np.float32(0.6983), np.float32(0.6563), np.float32(0.701), np.float32(0.0), np.float32(0.0), np.float32(0.7074), np.float32(0.5963), np.float32(0.6716), np.float32(0.5778), np.float32(0.6527), np.float32(0.5592), np.float32(0.7932), np.float32(0.727), np.float32(0.5515), np.float32(0.4622), np.float32(0.0), np.float32(0.0), np.float32(0.4345), np.float32(0.3054), np.float32(0.0), np.float32(0.0), np.float32(0.7264), np.float32(0.7098), np.float32(0.8006), np.float32(0.7973)] +2026-10-17 22:45:54.520389: Epoch time: 211.14 s +2026-10-17 22:45:54.520687: Yayy! New best EMA pseudo Dice: 0.4415999948978424 +2026-10-17 22:45:57.471386: +2026-10-17 22:45:57.472120: Epoch 369 +2026-10-17 22:45:57.472647: Current learning rate: 0.00661 +2026-10-17 22:49:47.741871: train_loss -0.7281 +2026-10-17 22:49:47.743055: val_loss -0.6443 +2026-10-17 22:49:47.743680: Pseudo dice [np.float32(0.8134), np.float32(0.8099), np.float32(0.7431), np.float32(0.8725), np.float32(0.7715), np.float32(0.8636), np.float32(0.841), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.7214), np.float32(0.7414), np.float32(0.6703), np.float32(0.6856), np.float32(0.0), np.float32(0.0), np.float32(0.7697), np.float32(0.6427), np.float32(0.7421), np.float32(0.6153), np.float32(0.6286), np.float32(0.5215), np.float32(0.6698), np.float32(0.8598), np.float32(0.5958), np.float32(0.5478), np.float32(0.0), np.float32(0.0), np.float32(0.2807), np.float32(0.5594), np.float32(0.0), np.float32(0.0), np.float32(0.7774), np.float32(0.757), np.float32(0.8097), np.float32(0.854)] +2026-10-17 22:49:47.744090: Epoch time: 230.27 s +2026-10-17 22:49:47.744608: Yayy! New best EMA pseudo Dice: 0.4507000148296356 +2026-10-17 22:49:49.472972: +2026-10-17 22:49:49.473749: Epoch 370 +2026-10-17 22:49:49.474174: Current learning rate: 0.0066 +2026-10-17 22:53:39.143672: train_loss -0.6821 +2026-10-17 22:53:39.144729: val_loss -0.5165 +2026-10-17 22:53:39.145602: Pseudo dice [np.float32(0.8982), np.float32(0.6228), np.float32(0.5273), np.float32(0.6096), np.float32(0.536), np.float32(0.6218), np.float32(0.533), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.5011), np.float32(0.5132), np.float32(0.4576), np.float32(0.5566), np.float32(0.0), np.float32(0.0), np.float32(0.551), np.float32(0.4762), np.float32(0.4882), np.float32(0.4478), np.float32(0.3763), np.float32(0.3229), np.float32(0.7782), np.float32(0.7353), np.float32(0.4958), np.float32(0.5209), np.float32(0.0), np.float32(0.0), np.float32(0.0204), np.float32(0.4035), np.float32(0.0), np.float32(0.0), np.float32(0.4955), np.float32(0.4823), np.float32(0.621), np.float32(0.6331)] +2026-10-17 22:53:39.146008: Epoch time: 229.67 s +2026-10-17 22:53:40.106464: +2026-10-17 22:53:40.107206: Epoch 371 +2026-10-17 22:53:40.107586: Current learning rate: 0.00659 +2026-10-17 22:57:35.631156: train_loss -0.7157 +2026-10-17 22:57:35.632071: val_loss -0.5884 +2026-10-17 22:57:35.632663: Pseudo dice [np.float32(0.8996), np.float32(0.7494), np.float32(0.7308), np.float32(0.8021), np.float32(0.6219), np.float32(0.7986), np.float32(0.7138), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.6551), np.float32(0.6312), np.float32(0.6546), np.float32(0.6753), np.float32(0.0), np.float32(0.0), np.float32(0.6606), np.float32(0.5484), np.float32(0.6603), np.float32(0.5516), np.float32(0.596), np.float32(0.5055), np.float32(0.7929), np.float32(0.7449), np.float32(0.5156), np.float32(0.5116), np.float32(0.0), np.float32(0.0), np.float32(0.2241), np.float32(0.4471), np.float32(0.0), np.float32(0.0), np.float32(0.7276), np.float32(0.6977), np.float32(0.7801), np.float32(0.782)] +2026-10-17 22:57:35.633121: Epoch time: 235.53 s +2026-10-17 22:57:36.532660: +2026-10-17 22:57:36.533484: Epoch 372 +2026-10-17 22:57:36.533717: Current learning rate: 0.00658 +2026-10-17 23:01:31.333556: train_loss -0.7066 +2026-10-17 23:01:31.334494: val_loss -0.6308 +2026-10-17 23:01:31.335047: Pseudo dice [np.float32(0.8737), np.float32(0.764), np.float32(0.7115), np.float32(0.8515), np.float32(0.7378), np.float32(0.845), np.float32(0.7846), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.7141), np.float32(0.7152), np.float32(0.6473), np.float32(0.6471), np.float32(0.0), np.float32(0.0), np.float32(0.7401), np.float32(0.6186), np.float32(0.7193), np.float32(0.6076), np.float32(0.6235), np.float32(0.467), np.float32(0.7629), np.float32(0.7661), np.float32(0.552), np.float32(0.5125), np.float32(0.0), np.float32(0.0), np.float32(0.4263), np.float32(0.5996), np.float32(0.0), np.float32(0.0), np.float32(0.752), np.float32(0.7564), np.float32(0.8185), np.float32(0.7997)] +2026-10-17 23:01:31.335425: Epoch time: 234.8 s +2026-10-17 23:01:31.335620: Yayy! New best EMA pseudo Dice: 0.4569999873638153 +2026-10-17 23:01:33.399623: +2026-10-17 23:01:33.400333: Epoch 373 +2026-10-17 23:01:33.400710: Current learning rate: 0.00657 +2026-10-17 23:05:02.581642: train_loss -0.7266 +2026-10-17 23:05:02.582563: val_loss -0.5262 +2026-10-17 23:05:02.583077: Pseudo dice [np.float32(0.8791), np.float32(0.595), np.float32(0.6109), np.float32(0.7348), np.float32(0.5771), np.float32(0.7143), np.float32(0.6422), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.6198), np.float32(0.6161), np.float32(0.627), np.float32(0.6315), np.float32(0.0), np.float32(0.0), np.float32(0.6252), np.float32(0.5097), np.float32(0.6269), np.float32(0.5173), np.float32(0.5209), np.float32(0.4683), np.float32(0.3955), np.float32(0.5044), np.float32(0.4685), np.float32(0.4331), np.float32(0.0), np.float32(0.0), np.float32(0.4035), np.float32(0.5613), np.float32(0.0), np.float32(0.0), np.float32(0.5664), np.float32(0.6098), np.float32(0.623), np.float32(0.6221)] +2026-10-17 23:05:02.583553: Epoch time: 209.18 s +2026-10-17 23:05:03.559368: +2026-10-17 23:05:03.560062: Epoch 374 +2026-10-17 23:05:03.560461: Current learning rate: 0.00656 +2026-10-17 23:08:40.238003: train_loss -0.6937 +2026-10-17 23:08:40.238890: val_loss -0.3138 +2026-10-17 23:08:40.239442: Pseudo dice [np.float32(0.8413), np.float32(0.2585), np.float32(0.2932), np.float32(0.456), np.float32(0.2107), np.float32(0.3433), np.float32(0.2971), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.4053), np.float32(0.365), np.float32(0.44), np.float32(0.404), np.float32(0.0), np.float32(0.0), np.float32(0.1614), np.float32(0.0826), np.float32(0.2696), np.float32(0.2715), np.float32(0.0875), np.float32(0.1549), np.float32(0.1848), np.float32(0.3034), np.float32(0.1771), np.float32(0.1667), np.float32(0.0), np.float32(0.0), np.float32(0.0659), np.float32(0.2478), np.float32(0.0), np.float32(0.0), np.float32(0.2017), np.float32(0.1856), np.float32(0.3291), np.float32(0.226)] +2026-10-17 23:08:40.240159: Epoch time: 216.68 s +2026-10-17 23:08:41.151401: +2026-10-17 23:08:41.152066: Epoch 375 +2026-10-17 23:08:41.152407: Current learning rate: 0.00655 +2026-10-17 23:12:11.128692: train_loss -0.6868 +2026-10-17 23:12:11.129682: val_loss -0.4649 +2026-10-17 23:12:11.130508: Pseudo dice [np.float32(0.8626), np.float32(0.5426), np.float32(0.5188), np.float32(0.6631), np.float32(0.4687), np.float32(0.6199), np.float32(0.4642), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.5551), np.float32(0.5493), np.float32(0.525), np.float32(0.5137), np.float32(0.0), np.float32(0.0), np.float32(0.5431), np.float32(0.4664), np.float32(0.4925), np.float32(0.4067), np.float32(0.4482), np.float32(0.4739), np.float32(0.4093), np.float32(0.4812), np.float32(0.4507), np.float32(0.479), np.float32(0.0), np.float32(0.0), np.float32(0.3463), np.float32(0.4534), np.float32(0.0), np.float32(0.0), np.float32(0.5422), np.float32(0.4911), np.float32(0.6215), np.float32(0.5796)] +2026-10-17 23:12:11.131045: Epoch time: 209.98 s +2026-10-17 23:12:12.068668: +2026-10-17 23:12:12.069292: Epoch 376 +2026-10-17 23:12:12.069621: Current learning rate: 0.00654 +2026-10-17 23:15:50.630292: train_loss -0.6822 +2026-10-17 23:15:50.631138: val_loss -0.3392 +2026-10-17 23:15:50.631691: Pseudo dice [np.float32(0.8778), np.float32(0.3469), np.float32(0.3789), np.float32(0.3783), np.float32(0.2718), np.float32(0.3341), np.float32(0.3114), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.3158), np.float32(0.2782), np.float32(0.3527), np.float32(0.2999), np.float32(0.0), np.float32(0.0), np.float32(0.2126), np.float32(0.1737), np.float32(0.2796), np.float32(0.2357), np.float32(0.1647), np.float32(0.2721), np.float32(0.6457), np.float32(0.4585), np.float32(0.1208), np.float32(0.1927), np.float32(0.0), np.float32(0.0), np.float32(0.1472), np.float32(0.0789), np.float32(0.0), np.float32(0.0), np.float32(0.2123), np.float32(0.2168), np.float32(0.2937), np.float32(0.2651)] +2026-10-17 23:15:50.632200: Epoch time: 218.56 s +2026-10-17 23:15:51.492288: +2026-10-17 23:15:51.493082: Epoch 377 +2026-10-17 23:15:51.493516: Current learning rate: 0.00653 +2026-10-17 23:19:45.368457: train_loss -0.6607 +2026-10-17 23:19:45.369070: val_loss -0.453 +2026-10-17 23:19:45.369539: Pseudo dice [np.float32(0.8866), np.float32(0.4828), np.float32(0.4748), np.float32(0.6094), np.float32(0.4483), np.float32(0.6001), np.float32(0.4714), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.5576), np.float32(0.5542), np.float32(0.5075), np.float32(0.4425), np.float32(0.0), np.float32(0.0), np.float32(0.4762), np.float32(0.3692), np.float32(0.4735), np.float32(0.3738), np.float32(0.4005), np.float32(0.4469), np.float32(0.3287), np.float32(0.4543), np.float32(0.5055), np.float32(0.4267), np.float32(0.0), np.float32(0.0), np.float32(0.272), np.float32(0.5074), np.float32(0.0), np.float32(0.0), np.float32(0.3553), np.float32(0.4237), np.float32(0.4752), np.float32(0.4816)] +2026-10-17 23:19:45.370451: Epoch time: 233.88 s +2026-10-17 23:19:46.261600: +2026-10-17 23:19:46.262424: Epoch 378 +2026-10-17 23:19:46.262905: Current learning rate: 0.00652 +2026-10-17 23:23:56.282310: train_loss -0.7108 +2026-10-17 23:23:56.283316: val_loss -0.5004 +2026-10-17 23:23:56.283931: Pseudo dice [np.float32(0.8799), np.float32(0.5682), np.float32(0.5994), np.float32(0.6895), np.float32(0.5038), np.float32(0.6672), np.float32(0.533), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.6221), np.float32(0.5878), np.float32(0.5973), np.float32(0.5592), np.float32(0.0), np.float32(0.0), np.float32(0.552), np.float32(0.4584), np.float32(0.5329), np.float32(0.4729), np.float32(0.4447), np.float32(0.4853), np.float32(0.4947), np.float32(0.4711), np.float32(0.4898), np.float32(0.4436), np.float32(0.0), np.float32(0.0), np.float32(0.4322), np.float32(0.4895), np.float32(0.0), np.float32(0.0), np.float32(0.5909), np.float32(0.5982), np.float32(0.6048), np.float32(0.6155)] +2026-10-17 23:23:56.284533: Epoch time: 250.02 s +2026-10-17 23:23:57.491306: +2026-10-17 23:23:57.491975: Epoch 379 +2026-10-17 23:23:57.492586: Current learning rate: 0.00651 +2026-10-17 23:28:17.997476: train_loss -0.7208 +2026-10-17 23:28:17.998539: val_loss -0.468 +2026-10-17 23:28:17.999265: Pseudo dice [np.float32(0.8725), np.float32(0.5561), np.float32(0.5387), np.float32(0.6447), np.float32(0.4843), np.float32(0.6258), np.float32(0.5413), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.5821), np.float32(0.5914), np.float32(0.5661), np.float32(0.531), np.float32(0.0), np.float32(0.0), np.float32(0.5579), np.float32(0.4411), np.float32(0.5178), np.float32(0.4237), np.float32(0.4401), np.float32(0.4146), np.float32(0.3744), np.float32(0.4506), np.float32(0.4133), np.float32(0.3558), np.float32(0.0), np.float32(0.0), np.float32(0.4236), np.float32(0.4993), np.float32(0.0), np.float32(0.0), np.float32(0.531), np.float32(0.5276), np.float32(0.5717), np.float32(0.553)] +2026-10-17 23:28:17.999858: Epoch time: 260.51 s +2026-10-17 23:28:18.950012: +2026-10-17 23:28:18.950836: Epoch 380 +2026-10-17 23:28:18.951214: Current learning rate: 0.0065 +2026-10-17 23:32:20.384789: train_loss -0.7302 +2026-10-17 23:32:20.385784: val_loss -0.4454 +2026-10-17 23:32:20.386504: Pseudo dice [np.float32(0.8633), np.float32(0.4574), np.float32(0.4559), np.float32(0.6027), np.float32(0.4239), np.float32(0.5753), np.float32(0.4561), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.5238), np.float32(0.498), np.float32(0.5463), np.float32(0.5035), np.float32(0.0), np.float32(0.0), np.float32(0.4524), np.float32(0.361), np.float32(0.464), np.float32(0.3672), np.float32(0.389), np.float32(0.3725), np.float32(0.3543), np.float32(0.4038), np.float32(0.3758), np.float32(0.2601), np.float32(0.0), np.float32(0.0), np.float32(0.4295), np.float32(0.4351), np.float32(0.0), np.float32(0.0), np.float32(0.5153), np.float32(0.4959), np.float32(0.4722), np.float32(0.4911)] +2026-10-17 23:32:20.386764: Epoch time: 241.44 s +2026-10-17 23:32:21.319264: +2026-10-17 23:32:21.320131: Epoch 381 +2026-10-17 23:32:21.320520: Current learning rate: 0.00649 +2026-10-17 23:36:22.016254: train_loss -0.7258 +2026-10-17 23:36:22.016829: val_loss -0.4433 +2026-10-17 23:36:22.017426: Pseudo dice [np.float32(0.8356), np.float32(0.5085), np.float32(0.4833), np.float32(0.6258), np.float32(0.4402), np.float32(0.6006), np.float32(0.4932), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.584), np.float32(0.564), np.float32(0.6012), np.float32(0.5689), np.float32(0.0), np.float32(0.0), np.float32(0.4957), np.float32(0.4067), np.float32(0.5065), np.float32(0.417), np.float32(0.5207), np.float32(0.4445), np.float32(0.4304), np.float32(0.5218), np.float32(0.422), np.float32(0.3382), np.float32(0.0), np.float32(0.0), np.float32(0.3101), np.float32(0.5101), np.float32(0.0), np.float32(0.0), np.float32(0.5264), np.float32(0.5181), np.float32(0.5278), np.float32(0.5392)] +2026-10-17 23:36:22.018312: Epoch time: 240.7 s +2026-10-17 23:36:22.969037: +2026-10-17 23:36:22.970168: Epoch 382 +2026-10-17 23:36:22.970532: Current learning rate: 0.00648 +2026-10-17 23:40:23.969305: train_loss -0.7279 +2026-10-17 23:40:23.970115: val_loss -0.2239 +2026-10-17 23:40:23.970981: Pseudo dice [np.float32(0.8797), np.float32(0.1448), np.float32(0.2085), np.float32(0.2116), np.float32(0.1528), np.float32(0.1817), np.float32(0.1888), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.1701), np.float32(0.142), np.float32(0.2602), np.float32(0.2061), np.float32(0.0), np.float32(0.0), np.float32(0.1838), np.float32(0.1222), np.float32(0.1928), np.float32(0.1485), np.float32(0.0687), np.float32(0.1439), np.float32(0.1195), np.float32(0.1534), np.float32(0.0766), np.float32(0.0833), np.float32(0.0), np.float32(0.0), np.float32(0.0483), np.float32(0.2298), np.float32(0.0), np.float32(0.0), np.float32(0.1434), np.float32(0.1688), np.float32(0.1206), np.float32(0.1182)] +2026-10-17 23:40:23.971536: Epoch time: 241.0 s +2026-10-17 23:40:24.950787: +2026-10-17 23:40:24.951477: Epoch 383 +2026-10-17 23:40:24.951942: Current learning rate: 0.00648 +2026-10-17 23:44:09.530309: train_loss -0.6922 +2026-10-17 23:44:09.530896: val_loss -0.2925 +2026-10-17 23:44:09.531420: Pseudo dice [np.float32(0.8406), np.float32(0.2106), np.float32(0.2665), np.float32(0.3626), np.float32(0.2179), np.float32(0.3109), np.float32(0.2076), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.2879), np.float32(0.2354), np.float32(0.3972), np.float32(0.2233), np.float32(0.0), np.float32(0.0), np.float32(0.1965), np.float32(0.1409), np.float32(0.2025), np.float32(0.145), np.float32(0.1746), np.float32(0.2039), np.float32(0.1407), np.float32(0.2651), np.float32(0.1882), np.float32(0.0765), np.float32(0.0), np.float32(0.0), np.float32(0.3644), np.float32(0.2908), np.float32(0.0), np.float32(0.0), np.float32(0.3228), np.float32(0.2516), np.float32(0.2818), np.float32(0.2533)] +2026-10-17 23:44:09.532251: Epoch time: 224.58 s +2026-10-17 23:44:10.494517: +2026-10-17 23:44:10.495310: Epoch 384 +2026-10-17 23:44:10.495780: Current learning rate: 0.00647 +2026-10-17 23:48:12.312599: train_loss -0.7011 +2026-10-17 23:48:12.313528: val_loss -0.3115 +2026-10-17 23:48:12.313911: Pseudo dice [np.float32(0.8908), np.float32(0.2699), np.float32(0.2799), np.float32(0.427), np.float32(0.2957), np.float32(0.3804), np.float32(0.2912), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.3755), np.float32(0.234), np.float32(0.3647), np.float32(0.3182), np.float32(0.0), np.float32(0.0), np.float32(0.3187), np.float32(0.2104), np.float32(0.2683), np.float32(0.1812), np.float32(0.1103), np.float32(0.1825), np.float32(0.0975), np.float32(0.3102), np.float32(0.1841), np.float32(0.0839), np.float32(0.0), np.float32(0.0), np.float32(0.3887), np.float32(0.3726), np.float32(0.0), np.float32(0.0), np.float32(0.3626), np.float32(0.3279), np.float32(0.2926), np.float32(0.271)] +2026-10-17 23:48:12.314370: Epoch time: 241.82 s +2026-10-17 23:48:13.280934: +2026-10-17 23:48:13.281794: Epoch 385 +2026-10-17 23:48:13.282148: Current learning rate: 0.00646 +2026-10-17 23:52:09.411384: train_loss -0.7224 +2026-10-17 23:52:09.412081: val_loss -0.4309 +2026-10-17 23:52:09.412875: Pseudo dice [np.float32(0.8655), np.float32(0.4837), np.float32(0.4667), np.float32(0.6011), np.float32(0.4199), np.float32(0.5525), np.float32(0.4379), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.525), np.float32(0.5284), np.float32(0.4854), np.float32(0.4435), np.float32(0.0), np.float32(0.0), np.float32(0.5365), np.float32(0.4177), np.float32(0.4522), np.float32(0.3632), np.float32(0.4937), np.float32(0.4138), np.float32(0.3251), np.float32(0.5036), np.float32(0.4461), np.float32(0.3964), np.float32(0.0), np.float32(0.0), np.float32(0.2757), np.float32(0.457), np.float32(0.0), np.float32(0.0), np.float32(0.4095), np.float32(0.3836), np.float32(0.4973), np.float32(0.4888)] +2026-10-17 23:52:09.413354: Epoch time: 236.13 s +2026-10-17 23:52:10.379326: +2026-10-17 23:52:10.380004: Epoch 386 +2026-10-17 23:52:10.380413: Current learning rate: 0.00645 +2026-10-17 23:55:51.335989: train_loss -0.7139 +2026-10-17 23:55:51.336587: val_loss -0.423 +2026-10-17 23:55:51.337126: Pseudo dice [np.float32(0.8797), np.float32(0.4901), np.float32(0.4923), np.float32(0.6065), np.float32(0.4378), np.float32(0.5409), np.float32(0.4681), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.5181), np.float32(0.4981), np.float32(0.5523), np.float32(0.5317), np.float32(0.0), np.float32(0.0), np.float32(0.5216), np.float32(0.3786), np.float32(0.4732), np.float32(0.3713), np.float32(0.4068), np.float32(0.39), np.float32(0.3392), np.float32(0.3991), np.float32(0.3668), np.float32(0.2676), np.float32(0.0), np.float32(0.0), np.float32(0.4545), np.float32(0.4234), np.float32(0.0), np.float32(0.0), np.float32(0.4956), np.float32(0.4554), np.float32(0.4966), np.float32(0.4753)] +2026-10-17 23:55:51.337819: Epoch time: 220.96 s +2026-10-17 23:55:52.344809: +2026-10-17 23:55:52.345645: Epoch 387 +2026-10-17 23:55:52.346084: Current learning rate: 0.00644 +2026-10-18 00:00:00.102261: train_loss -0.7109 +2026-10-18 00:00:00.102918: val_loss -0.4044 +2026-10-18 00:00:00.103588: Pseudo dice [np.float32(0.8856), np.float32(0.4295), np.float32(0.4436), np.float32(0.548), np.float32(0.4075), np.float32(0.5151), np.float32(0.4543), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.4761), np.float32(0.4648), np.float32(0.5183), np.float32(0.5092), np.float32(0.0), np.float32(0.0), np.float32(0.4573), np.float32(0.3455), np.float32(0.4483), np.float32(0.3427), np.float32(0.3917), np.float32(0.3346), np.float32(0.2832), np.float32(0.2578), np.float32(0.3438), np.float32(0.2709), np.float32(0.0), np.float32(0.0), np.float32(0.4062), np.float32(0.3188), np.float32(0.0), np.float32(0.0), np.float32(0.4895), np.float32(0.444), np.float32(0.451), np.float32(0.4293)] +2026-10-18 00:00:00.104199: Epoch time: 247.76 s +2026-10-18 00:00:01.007827: +2026-10-18 00:00:01.008592: Epoch 388 +2026-10-18 00:00:01.008980: Current learning rate: 0.00643 +2026-10-18 00:03:55.740840: train_loss -0.7283 +2026-10-18 00:03:55.742409: val_loss -0.4529 +2026-10-18 00:03:55.743148: Pseudo dice [np.float32(0.8674), np.float32(0.54), np.float32(0.5112), np.float32(0.603), np.float32(0.4576), np.float32(0.565), np.float32(0.4751), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.5276), np.float32(0.5097), np.float32(0.5523), np.float32(0.5048), np.float32(0.0), np.float32(0.0), np.float32(0.5556), np.float32(0.4676), np.float32(0.4903), np.float32(0.3603), np.float32(0.4719), np.float32(0.3901), np.float32(0.5382), np.float32(0.5424), np.float32(0.4574), np.float32(0.3684), np.float32(0.0), np.float32(0.0), np.float32(0.4066), np.float32(0.4408), np.float32(0.0), np.float32(0.0), np.float32(0.5256), np.float32(0.4585), np.float32(0.5844), np.float32(0.5401)] +2026-10-18 00:03:55.743483: Epoch time: 234.74 s +2026-10-18 00:03:56.649049: +2026-10-18 00:03:56.649830: Epoch 389 +2026-10-18 00:03:56.650260: Current learning rate: 0.00642 +2026-10-18 00:08:08.530201: train_loss -0.7358 +2026-10-18 00:08:08.530787: val_loss -0.4861 +2026-10-18 00:08:08.531327: Pseudo dice [np.float32(0.872), np.float32(0.5981), np.float32(0.5675), np.float32(0.693), np.float32(0.5224), np.float32(0.6551), np.float32(0.5299), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.5693), np.float32(0.5839), np.float32(0.5256), np.float32(0.5112), np.float32(0.0), np.float32(0.0), np.float32(0.6094), np.float32(0.5085), np.float32(0.509), np.float32(0.4108), np.float32(0.4679), np.float32(0.3854), np.float32(0.5074), np.float32(0.5012), np.float32(0.4761), np.float32(0.4597), np.float32(0.0), np.float32(0.0), np.float32(0.4799), np.float32(0.513), np.float32(0.0), np.float32(0.0), np.float32(0.5692), np.float32(0.5584), np.float32(0.656), np.float32(0.6342)] +2026-10-18 00:08:08.532405: Epoch time: 251.88 s +2026-10-18 00:08:09.501041: +2026-10-18 00:08:09.501737: Epoch 390 +2026-10-18 00:08:09.502028: Current learning rate: 0.00641 +2026-10-18 00:12:09.257878: train_loss -0.7371 +2026-10-18 00:12:09.258796: val_loss -0.5286 +2026-10-18 00:12:09.259443: Pseudo dice [np.float32(0.8755), np.float32(0.6218), np.float32(0.5989), np.float32(0.6812), np.float32(0.5385), np.float32(0.6546), np.float32(0.6185), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.5646), np.float32(0.5416), np.float32(0.5624), np.float32(0.5865), np.float32(0.0), np.float32(0.0), np.float32(0.5965), np.float32(0.5144), np.float32(0.5829), np.float32(0.4701), np.float32(0.498), np.float32(0.4417), np.float32(0.5524), np.float32(0.6693), np.float32(0.4973), np.float32(0.4682), np.float32(0.0), np.float32(0.0), np.float32(0.3427), np.float32(0.4738), np.float32(0.0), np.float32(0.0), np.float32(0.5037), np.float32(0.5122), np.float32(0.6556), np.float32(0.6333)] +2026-10-18 00:12:09.259952: Epoch time: 239.76 s +2026-10-18 00:12:10.196892: +2026-10-18 00:12:10.197698: Epoch 391 +2026-10-18 00:12:10.198059: Current learning rate: 0.0064 +2026-10-18 00:16:13.666108: train_loss -0.7446 +2026-10-18 00:16:13.666692: val_loss -0.4895 +2026-10-18 00:16:13.667474: Pseudo dice [np.float32(0.8709), np.float32(0.5719), np.float32(0.5322), np.float32(0.6266), np.float32(0.5086), np.float32(0.6189), np.float32(0.5687), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.5164), np.float32(0.5549), np.float32(0.4833), np.float32(0.5333), np.float32(0.0), np.float32(0.0), np.float32(0.5702), np.float32(0.4874), np.float32(0.5456), np.float32(0.4744), np.float32(0.5091), np.float32(0.4616), np.float32(0.5644), np.float32(0.6014), np.float32(0.5249), np.float32(0.4904), np.float32(0.0), np.float32(0.0), np.float32(0.1888), np.float32(0.5361), np.float32(0.0), np.float32(0.0), np.float32(0.5301), np.float32(0.5221), np.float32(0.6002), np.float32(0.6249)] +2026-10-18 00:16:13.667999: Epoch time: 243.47 s +2026-10-18 00:16:14.665310: +2026-10-18 00:16:14.666219: Epoch 392 +2026-10-18 00:16:14.666696: Current learning rate: 0.00639 +2026-10-18 00:19:58.399143: train_loss -0.7409 +2026-10-18 00:19:58.399827: val_loss -0.2925 +2026-10-18 00:19:58.400841: Pseudo dice [np.float32(0.8311), np.float32(0.2468), np.float32(0.272), np.float32(0.2968), np.float32(0.2054), np.float32(0.2938), np.float32(0.2782), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.2305), np.float32(0.2096), np.float32(0.3042), np.float32(0.272), np.float32(0.0), np.float32(0.0), np.float32(0.2453), np.float32(0.1971), np.float32(0.2353), np.float32(0.2067), np.float32(0.1433), np.float32(0.1783), np.float32(0.2636), np.float32(0.3639), np.float32(0.1705), np.float32(0.1801), np.float32(0.0), np.float32(0.0), np.float32(0.1162), np.float32(0.3122), np.float32(0.0), np.float32(0.0), np.float32(0.1965), np.float32(0.2061), np.float32(0.2261), np.float32(0.2411)] +2026-10-18 00:19:58.401329: Epoch time: 223.74 s +2026-10-18 00:19:59.289128: +2026-10-18 00:19:59.290131: Epoch 393 +2026-10-18 00:19:59.290492: Current learning rate: 0.00638 +2026-10-18 00:23:49.363364: train_loss -0.7441 +2026-10-18 00:23:49.363964: val_loss -0.4706 +2026-10-18 00:23:49.364635: Pseudo dice [np.float32(0.8554), np.float32(0.5995), np.float32(0.5364), np.float32(0.6682), np.float32(0.5588), np.float32(0.6342), np.float32(0.5956), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.5761), np.float32(0.5831), np.float32(0.5524), np.float32(0.5707), np.float32(0.0), np.float32(0.0), np.float32(0.6531), np.float32(0.5639), np.float32(0.5617), np.float32(0.4818), np.float32(0.5093), np.float32(0.3976), np.float32(0.327), np.float32(0.3229), np.float32(0.4565), np.float32(0.4252), np.float32(0.0), np.float32(0.0), np.float32(0.2644), np.float32(0.488), np.float32(0.0), np.float32(0.0), np.float32(0.5605), np.float32(0.5489), np.float32(0.5549), np.float32(0.5259)] +2026-10-18 00:23:49.365149: Epoch time: 230.08 s +2026-10-18 00:23:50.247556: +2026-10-18 00:23:50.248312: Epoch 394 +2026-10-18 00:23:50.248686: Current learning rate: 0.00637 +2026-10-18 00:27:58.991381: train_loss -0.7372 +2026-10-18 00:27:58.992055: val_loss -0.5897 +2026-10-18 00:27:58.992854: Pseudo dice [np.float32(0.8086), np.float32(0.7386), np.float32(0.6981), np.float32(0.7993), np.float32(0.6722), np.float32(0.7847), np.float32(0.7229), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.6571), np.float32(0.6443), np.float32(0.6345), np.float32(0.6791), np.float32(0.0), np.float32(0.0), np.float32(0.6835), np.float32(0.6055), np.float32(0.6857), np.float32(0.6006), np.float32(0.622), np.float32(0.4811), np.float32(0.627), np.float32(0.7354), np.float32(0.5304), np.float32(0.5147), np.float32(0.0), np.float32(0.0), np.float32(0.416), np.float32(0.5774), np.float32(0.0), np.float32(0.0), np.float32(0.7087), np.float32(0.684), np.float32(0.7818), np.float32(0.7619)] +2026-10-18 00:27:58.993510: Epoch time: 248.75 s +2026-10-18 00:27:59.973021: +2026-10-18 00:27:59.973761: Epoch 395 +2026-10-18 00:27:59.974167: Current learning rate: 0.00636 +2026-10-18 00:32:26.083900: train_loss -0.7384 +2026-10-18 00:32:26.084556: val_loss -0.5899 +2026-10-18 00:32:26.085095: Pseudo dice [np.float32(0.8801), np.float32(0.7377), np.float32(0.6797), np.float32(0.7687), np.float32(0.6626), np.float32(0.7688), np.float32(0.7311), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.6303), np.float32(0.6266), np.float32(0.62), np.float32(0.7011), np.float32(0.0), np.float32(0.0), np.float32(0.716), np.float32(0.6124), np.float32(0.681), np.float32(0.5937), np.float32(0.6055), np.float32(0.5307), np.float32(0.6426), np.float32(0.6829), np.float32(0.5665), np.float32(0.5428), np.float32(0.0), np.float32(0.0), np.float32(0.2313), np.float32(0.5918), np.float32(0.0), np.float32(0.0), np.float32(0.6994), np.float32(0.6621), np.float32(0.7333), np.float32(0.7386)] +2026-10-18 00:32:26.085823: Epoch time: 266.11 s +2026-10-18 00:32:27.055026: +2026-10-18 00:32:27.055625: Epoch 396 +2026-10-18 00:32:27.056047: Current learning rate: 0.00635 +2026-10-18 00:36:16.541553: train_loss -0.7449 +2026-10-18 00:36:16.542198: val_loss -0.5455 +2026-10-18 00:36:16.542810: Pseudo dice [np.float32(0.8772), np.float32(0.626), np.float32(0.6024), np.float32(0.6897), np.float32(0.5967), np.float32(0.7057), np.float32(0.6315), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.5805), np.float32(0.6137), np.float32(0.589), np.float32(0.6134), np.float32(0.0), np.float32(0.0), np.float32(0.6244), np.float32(0.5466), np.float32(0.5719), np.float32(0.5086), np.float32(0.5172), np.float32(0.4198), np.float32(0.6368), np.float32(0.6457), np.float32(0.4729), np.float32(0.5093), np.float32(0.0), np.float32(0.0), np.float32(0.372), np.float32(0.499), np.float32(0.0), np.float32(0.0), np.float32(0.6474), np.float32(0.6467), np.float32(0.6634), np.float32(0.6514)] +2026-10-18 00:36:16.543418: Epoch time: 229.49 s +2026-10-18 00:36:17.541529: +2026-10-18 00:36:17.542334: Epoch 397 +2026-10-18 00:36:17.542604: Current learning rate: 0.00634 +2026-10-18 00:40:03.881021: train_loss -0.7303 +2026-10-18 00:40:03.881810: val_loss -0.4815 +2026-10-18 00:40:03.882415: Pseudo dice [np.float32(0.8192), np.float32(0.5704), np.float32(0.5537), np.float32(0.6869), np.float32(0.5816), np.float32(0.665), np.float32(0.6032), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.5179), np.float32(0.5011), np.float32(0.5332), np.float32(0.5653), np.float32(0.0), np.float32(0.0), np.float32(0.6302), np.float32(0.4932), np.float32(0.6049), np.float32(0.4522), np.float32(0.4804), np.float32(0.4601), np.float32(0.435), np.float32(0.5791), np.float32(0.3362), np.float32(0.3201), np.float32(0.0), np.float32(0.0), np.float32(0.3047), np.float32(0.4485), np.float32(0.0), np.float32(0.0), np.float32(0.5484), np.float32(0.551), np.float32(0.6128), np.float32(0.623)] +2026-10-18 00:40:03.882853: Epoch time: 226.34 s +2026-10-18 00:40:05.082127: +2026-10-18 00:40:05.082921: Epoch 398 +2026-10-18 00:40:05.083441: Current learning rate: 0.00633 +2026-10-18 00:44:13.984216: train_loss -0.7143 +2026-10-18 00:44:13.984896: val_loss -0.5536 +2026-10-18 00:44:13.985434: Pseudo dice [np.float32(0.8523), np.float32(0.6076), np.float32(0.5938), np.float32(0.6625), np.float32(0.6117), np.float32(0.6691), np.float32(0.631), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.5478), np.float32(0.5526), np.float32(0.616), np.float32(0.632), np.float32(0.0), np.float32(0.0), np.float32(0.6372), np.float32(0.5428), np.float32(0.6279), np.float32(0.5115), np.float32(0.5529), np.float32(0.5081), np.float32(0.6292), np.float32(0.5941), np.float32(0.4251), np.float32(0.3457), np.float32(0.0), np.float32(0.0), np.float32(0.5227), np.float32(0.4413), np.float32(0.0), np.float32(0.0), np.float32(0.6343), np.float32(0.6175), np.float32(0.6177), np.float32(0.6211)] +2026-10-18 00:44:13.986197: Epoch time: 248.9 s +2026-10-18 00:44:14.950442: +2026-10-18 00:44:14.951260: Epoch 399 +2026-10-18 00:44:14.951699: Current learning rate: 0.00632 +2026-10-18 00:48:02.803111: train_loss -0.7048 +2026-10-18 00:48:02.803652: val_loss -0.4311 +2026-10-18 00:48:02.804310: Pseudo dice [np.float32(0.8795), np.float32(0.4671), np.float32(0.4788), np.float32(0.5343), np.float32(0.4548), np.float32(0.4954), np.float32(0.4979), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.4589), np.float32(0.4328), np.float32(0.5756), np.float32(0.568), np.float32(0.0), np.float32(0.0), np.float32(0.4713), np.float32(0.3777), np.float32(0.4858), np.float32(0.409), np.float32(0.3462), np.float32(0.339), np.float32(0.4002), np.float32(0.4305), np.float32(0.2505), np.float32(0.1175), np.float32(0.0), np.float32(0.0), np.float32(0.2335), np.float32(0.2522), np.float32(0.0), np.float32(0.0), np.float32(0.4577), np.float32(0.4806), np.float32(0.4732), np.float32(0.4371)] +2026-10-18 00:48:02.804895: Epoch time: 227.85 s +2026-10-18 00:48:06.487482: +2026-10-18 00:48:06.488304: Epoch 400 +2026-10-18 00:48:06.488752: Current learning rate: 0.00631 +2026-10-18 00:51:51.688258: train_loss -0.7146 +2026-10-18 00:51:51.689251: val_loss -0.5133 +2026-10-18 00:51:51.689937: Pseudo dice [np.float32(0.8764), np.float32(0.5423), np.float32(0.5564), np.float32(0.6541), np.float32(0.4686), np.float32(0.6686), np.float32(0.576), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.5238), np.float32(0.5407), np.float32(0.4633), np.float32(0.5318), np.float32(0.0), np.float32(0.0), np.float32(0.5332), np.float32(0.4461), np.float32(0.5371), np.float32(0.4653), np.float32(0.5367), np.float32(0.5284), np.float32(0.6202), np.float32(0.6446), np.float32(0.5172), np.float32(0.4869), np.float32(0.0), np.float32(0.0), np.float32(0.4352), np.float32(0.3962), np.float32(0.0), np.float32(0.0), np.float32(0.5346), np.float32(0.6001), np.float32(0.657), np.float32(0.7007)] +2026-10-18 00:51:51.690297: Epoch time: 225.2 s +2026-10-18 00:51:52.678127: +2026-10-18 00:51:52.679007: Epoch 401 +2026-10-18 00:51:52.679460: Current learning rate: 0.0063 +2026-10-18 00:55:59.892662: train_loss -0.7211 +2026-10-18 00:55:59.893188: val_loss -0.3742 +2026-10-18 00:55:59.893718: Pseudo dice [np.float32(0.8901), np.float32(0.3687), np.float32(0.3158), np.float32(0.408), np.float32(0.2796), np.float32(0.3693), np.float32(0.3344), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.3098), np.float32(0.3549), np.float32(0.3563), np.float32(0.3485), np.float32(0.0), np.float32(0.0), np.float32(0.3362), np.float32(0.295), np.float32(0.2982), np.float32(0.2856), np.float32(0.2028), np.float32(0.2523), np.float32(0.3499), np.float32(0.4677), np.float32(0.305), np.float32(0.3083), np.float32(0.0), np.float32(0.0), np.float32(0.1052), np.float32(0.4409), np.float32(0.0), np.float32(0.0), np.float32(0.264), np.float32(0.2424), np.float32(0.3582), np.float32(0.4004)] +2026-10-18 00:55:59.894684: Epoch time: 247.22 s +2026-10-18 00:56:00.773864: +2026-10-18 00:56:00.774645: Epoch 402 +2026-10-18 00:56:00.775126: Current learning rate: 0.0063 +2026-10-18 01:00:13.545436: train_loss -0.6736 +2026-10-18 01:00:13.546206: val_loss -0.4388 +2026-10-18 01:00:13.546723: Pseudo dice [np.float32(0.8559), np.float32(0.4489), np.float32(0.455), np.float32(0.5994), np.float32(0.4168), np.float32(0.5727), np.float32(0.4478), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.5483), np.float32(0.5258), np.float32(0.5405), np.float32(0.5322), np.float32(0.0), np.float32(0.0), np.float32(0.5131), np.float32(0.423), np.float32(0.4855), np.float32(0.3476), np.float32(0.3726), np.float32(0.3818), np.float32(0.3514), np.float32(0.4412), np.float32(0.4814), np.float32(0.4026), np.float32(0.0), np.float32(0.0), np.float32(0.4622), np.float32(0.4185), np.float32(0.0), np.float32(0.0), np.float32(0.4491), np.float32(0.4843), np.float32(0.5002), np.float32(0.5065)] +2026-10-18 01:00:13.547328: Epoch time: 252.77 s +2026-10-18 01:00:14.453346: +2026-10-18 01:00:14.454300: Epoch 403 +2026-10-18 01:00:14.454670: Current learning rate: 0.00629 +2026-10-18 01:04:28.777966: train_loss -0.6779 +2026-10-18 01:04:28.778826: val_loss -0.4841 +2026-10-18 01:04:28.779376: Pseudo dice [np.float32(0.8838), np.float32(0.6012), np.float32(0.538), np.float32(0.6938), np.float32(0.5573), np.float32(0.6836), np.float32(0.569), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.5965), np.float32(0.5392), np.float32(0.5527), np.float32(0.5391), np.float32(0.0), np.float32(0.0), np.float32(0.6091), np.float32(0.4992), np.float32(0.5943), np.float32(0.4754), np.float32(0.5859), np.float32(0.5082), np.float32(0.313), np.float32(0.3185), np.float32(0.4523), np.float32(0.4633), np.float32(0.0), np.float32(0.0), np.float32(0.3975), np.float32(0.4511), np.float32(0.0), np.float32(0.0), np.float32(0.5675), np.float32(0.5611), np.float32(0.5202), np.float32(0.5486)] +2026-10-18 01:04:28.779886: Epoch time: 254.33 s +2026-10-18 01:04:29.772819: +2026-10-18 01:04:29.773674: Epoch 404 +2026-10-18 01:04:29.774019: Current learning rate: 0.00628 +2026-10-18 01:08:47.778085: train_loss -0.6625 +2026-10-18 01:08:47.778672: val_loss -0.4286 +2026-10-18 01:08:47.779267: Pseudo dice [np.float32(0.8512), np.float32(0.4319), np.float32(0.4665), np.float32(0.5598), np.float32(0.4705), np.float32(0.5503), np.float32(0.4571), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.5303), np.float32(0.493), np.float32(0.4561), np.float32(0.4418), np.float32(0.0), np.float32(0.0), np.float32(0.4923), np.float32(0.3788), np.float32(0.4602), np.float32(0.381), np.float32(0.4017), np.float32(0.3508), np.float32(0.3021), np.float32(0.2612), np.float32(0.2363), np.float32(0.1288), np.float32(0.0), np.float32(0.0), np.float32(0.3457), np.float32(0.2568), np.float32(0.0), np.float32(0.0), np.float32(0.4959), np.float32(0.5064), np.float32(0.456), np.float32(0.4472)] +2026-10-18 01:08:47.780263: Epoch time: 258.01 s +2026-10-18 01:08:48.726803: +2026-10-18 01:08:48.727674: Epoch 405 +2026-10-18 01:08:48.728078: Current learning rate: 0.00627 +2026-10-18 01:13:00.701530: train_loss -0.6631 +2026-10-18 01:13:00.702088: val_loss -0.4222 +2026-10-18 01:13:00.703142: Pseudo dice [np.float32(0.9105), np.float32(0.4072), np.float32(0.4007), np.float32(0.559), np.float32(0.4646), np.float32(0.5332), np.float32(0.4608), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.4727), np.float32(0.4454), np.float32(0.5446), np.float32(0.5727), np.float32(0.0), np.float32(0.0), np.float32(0.4282), np.float32(0.3335), np.float32(0.4524), np.float32(0.3702), np.float32(0.294), np.float32(0.2362), np.float32(0.2208), np.float32(0.3872), np.float32(0.2785), np.float32(0.1775), np.float32(0.0), np.float32(0.0), np.float32(0.3102), np.float32(0.3338), np.float32(0.0), np.float32(0.0), np.float32(0.4978), np.float32(0.4762), np.float32(0.4227), np.float32(0.4342)] +2026-10-18 01:13:00.703610: Epoch time: 251.98 s +2026-10-18 01:13:01.625792: +2026-10-18 01:13:01.626562: Epoch 406 +2026-10-18 01:13:01.626919: Current learning rate: 0.00626 +2026-10-18 01:17:00.569058: train_loss -0.707 +2026-10-18 01:17:00.569832: val_loss -0.2857 +2026-10-18 01:17:00.570344: Pseudo dice [np.float32(0.9023), np.float32(0.2431), np.float32(0.3042), np.float32(0.3511), np.float32(0.2768), np.float32(0.3459), np.float32(0.2781), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.3235), np.float32(0.3022), np.float32(0.4253), np.float32(0.3548), np.float32(0.0), np.float32(0.0), np.float32(0.2862), np.float32(0.2275), np.float32(0.2706), np.float32(0.213), np.float32(0.25), np.float32(0.2188), np.float32(0.1227), np.float32(0.2361), np.float32(0.1542), np.float32(0.0835), np.float32(0.0), np.float32(0.0), np.float32(0.21), np.float32(0.2284), np.float32(0.0), np.float32(0.0), np.float32(0.2878), np.float32(0.3166), np.float32(0.2668), np.float32(0.2381)] +2026-10-18 01:17:00.570663: Epoch time: 238.95 s +2026-10-18 01:17:01.521226: +2026-10-18 01:17:01.521904: Epoch 407 +2026-10-18 01:17:01.522380: Current learning rate: 0.00625 +2026-10-18 01:20:56.923292: train_loss -0.6999 +2026-10-18 01:20:56.924227: val_loss -0.3741 +2026-10-18 01:20:56.924726: Pseudo dice [np.float32(0.8451), np.float32(0.3351), np.float32(0.4011), np.float32(0.5117), np.float32(0.3957), np.float32(0.4825), np.float32(0.4349), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.404), np.float32(0.3664), np.float32(0.4195), np.float32(0.3661), np.float32(0.0), np.float32(0.0), np.float32(0.4679), np.float32(0.3933), np.float32(0.4104), np.float32(0.3056), np.float32(0.4249), np.float32(0.3667), np.float32(0.2253), np.float32(0.2945), np.float32(0.2574), np.float32(0.1199), np.float32(0.0), np.float32(0.0), np.float32(0.3339), np.float32(0.2168), np.float32(0.0), np.float32(0.0), np.float32(0.4412), np.float32(0.4238), np.float32(0.3841), np.float32(0.3672)] +2026-10-18 01:20:56.925190: Epoch time: 235.4 s +2026-10-18 01:20:57.831372: +2026-10-18 01:20:57.831913: Epoch 408 +2026-10-18 01:20:57.832314: Current learning rate: 0.00624 +2026-10-18 01:25:08.472197: train_loss -0.7182 +2026-10-18 01:25:08.473061: val_loss -0.2386 +2026-10-18 01:25:08.473818: Pseudo dice [np.float32(0.8784), np.float32(0.1345), np.float32(0.1805), np.float32(0.2072), np.float32(0.1351), np.float32(0.1951), np.float32(0.1801), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.1526), np.float32(0.1739), np.float32(0.2744), np.float32(0.269), np.float32(0.0), np.float32(0.0), np.float32(0.1457), np.float32(0.0994), np.float32(0.2027), np.float32(0.1482), np.float32(0.0474), np.float32(0.1221), np.float32(0.2019), np.float32(0.3302), np.float32(0.0723), np.float32(0.0786), np.float32(0.0), np.float32(0.0), np.float32(0.1715), np.float32(0.2248), np.float32(0.0), np.float32(0.0), np.float32(0.1105), np.float32(0.1821), np.float32(0.1511), np.float32(0.1619)] +2026-10-18 01:25:08.474276: Epoch time: 250.64 s +2026-10-18 01:25:09.457849: +2026-10-18 01:25:09.458701: Epoch 409 +2026-10-18 01:25:09.458942: Current learning rate: 0.00623 +2026-10-18 01:29:09.950972: train_loss -0.7291 +2026-10-18 01:29:09.951831: val_loss -0.4236 +2026-10-18 01:29:09.952380: Pseudo dice [np.float32(0.8803), np.float32(0.4579), np.float32(0.4728), np.float32(0.5824), np.float32(0.3998), np.float32(0.543), np.float32(0.4654), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.4986), np.float32(0.4246), np.float32(0.5131), np.float32(0.4306), np.float32(0.0), np.float32(0.0), np.float32(0.4579), np.float32(0.3558), np.float32(0.4691), np.float32(0.3581), np.float32(0.4394), np.float32(0.3825), np.float32(0.2923), np.float32(0.5304), np.float32(0.4109), np.float32(0.2916), np.float32(0.0), np.float32(0.0), np.float32(0.4824), np.float32(0.4413), np.float32(0.0), np.float32(0.0), np.float32(0.4771), np.float32(0.4821), np.float32(0.4983), np.float32(0.496)] +2026-10-18 01:29:09.952945: Epoch time: 240.5 s +2026-10-18 01:29:10.909651: +2026-10-18 01:29:10.910483: Epoch 410 +2026-10-18 01:29:10.910875: Current learning rate: 0.00622 +2026-10-18 01:33:11.581883: train_loss -0.7344 +2026-10-18 01:33:11.582650: val_loss -0.4744 +2026-10-18 01:33:11.583225: Pseudo dice [np.float32(0.8512), np.float32(0.5027), np.float32(0.5465), np.float32(0.6442), np.float32(0.5426), np.float32(0.6394), np.float32(0.5521), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.5013), np.float32(0.4948), np.float32(0.4965), np.float32(0.5173), np.float32(0.0), np.float32(0.0), np.float32(0.5175), np.float32(0.4096), np.float32(0.5038), np.float32(0.4454), np.float32(0.4968), np.float32(0.4125), np.float32(0.4042), np.float32(0.4968), np.float32(0.2643), np.float32(0.2077), np.float32(0.0), np.float32(0.0), np.float32(0.617), np.float32(0.3537), np.float32(0.0), np.float32(0.0), np.float32(0.6465), np.float32(0.6525), np.float32(0.6285), np.float32(0.6197)] +2026-10-18 01:33:11.583707: Epoch time: 240.67 s +2026-10-18 01:33:12.462440: +2026-10-18 01:33:12.463055: Epoch 411 +2026-10-18 01:33:12.463392: Current learning rate: 0.00621 +2026-10-18 01:37:08.495401: train_loss -0.7228 +2026-10-18 01:37:08.495896: val_loss -0.528 +2026-10-18 01:37:08.496834: Pseudo dice [np.float32(0.8915), np.float32(0.5914), np.float32(0.5695), np.float32(0.7037), np.float32(0.5926), np.float32(0.6968), np.float32(0.6195), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.5589), np.float32(0.5435), np.float32(0.5785), np.float32(0.6187), np.float32(0.0), np.float32(0.0), np.float32(0.6234), np.float32(0.5025), np.float32(0.5926), np.float32(0.5085), np.float32(0.5069), np.float32(0.443), np.float32(0.5783), np.float32(0.5414), np.float32(0.4147), np.float32(0.4027), np.float32(0.0), np.float32(0.0), np.float32(0.2935), np.float32(0.3922), np.float32(0.0), np.float32(0.0), np.float32(0.659), np.float32(0.6509), np.float32(0.6649), np.float32(0.6789)] +2026-10-18 01:37:08.497133: Epoch time: 236.04 s +2026-10-18 01:37:09.433054: +2026-10-18 01:37:09.433636: Epoch 412 +2026-10-18 01:37:09.433985: Current learning rate: 0.0062 +2026-10-18 01:41:15.926059: train_loss -0.7327 +2026-10-18 01:41:15.926934: val_loss -0.5168 +2026-10-18 01:41:15.927571: Pseudo dice [np.float32(0.8913), np.float32(0.6379), np.float32(0.5926), np.float32(0.7021), np.float32(0.5704), np.float32(0.6828), np.float32(0.6339), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.5585), np.float32(0.5317), np.float32(0.5384), np.float32(0.5755), np.float32(0.0), np.float32(0.0), np.float32(0.6195), np.float32(0.5332), np.float32(0.5775), np.float32(0.5302), np.float32(0.5699), np.float32(0.486), np.float32(0.6343), np.float32(0.662), np.float32(0.3917), np.float32(0.4502), np.float32(0.0), np.float32(0.0), np.float32(0.2541), np.float32(0.4736), np.float32(0.0), np.float32(0.0), np.float32(0.6002), np.float32(0.5926), np.float32(0.6405), np.float32(0.6712)] +2026-10-18 01:41:15.927971: Epoch time: 246.5 s +2026-10-18 01:41:16.810242: +2026-10-18 01:41:16.810980: Epoch 413 +2026-10-18 01:41:16.811247: Current learning rate: 0.00619 +2026-10-18 01:45:13.221254: train_loss -0.7421 +2026-10-18 01:45:13.221996: val_loss -0.4244 +2026-10-18 01:45:13.222539: Pseudo dice [np.float32(0.8985), np.float32(0.4158), np.float32(0.4335), np.float32(0.574), np.float32(0.4532), np.float32(0.5466), np.float32(0.4765), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.4676), np.float32(0.4202), np.float32(0.4712), np.float32(0.4024), np.float32(0.0), np.float32(0.0), np.float32(0.5263), np.float32(0.4422), np.float32(0.4382), np.float32(0.3582), np.float32(0.432), np.float32(0.3872), np.float32(0.3657), np.float32(0.4359), np.float32(0.3524), np.float32(0.3182), np.float32(0.0), np.float32(0.0), np.float32(0.3473), np.float32(0.3548), np.float32(0.0), np.float32(0.0), np.float32(0.4912), np.float32(0.5104), np.float32(0.4907), np.float32(0.4803)] +2026-10-18 01:45:13.222847: Epoch time: 236.41 s +2026-10-18 01:45:14.157002: +2026-10-18 01:45:14.157806: Epoch 414 +2026-10-18 01:45:14.158145: Current learning rate: 0.00618 +2026-10-18 01:49:22.748724: train_loss -0.7402 +2026-10-18 01:49:22.749363: val_loss -0.3178 +2026-10-18 01:49:22.750252: Pseudo dice [np.float32(0.8866), np.float32(0.3308), np.float32(0.307), np.float32(0.4296), np.float32(0.3068), np.float32(0.3713), np.float32(0.2983), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.3545), np.float32(0.2932), np.float32(0.3844), np.float32(0.3117), np.float32(0.0), np.float32(0.0), np.float32(0.4071), np.float32(0.352), np.float32(0.2833), np.float32(0.2312), np.float32(0.2975), np.float32(0.2746), np.float32(0.2202), np.float32(0.2446), np.float32(0.2433), np.float32(0.294), np.float32(0.0), np.float32(0.0), np.float32(0.2273), np.float32(0.2649), np.float32(0.0), np.float32(0.0), np.float32(0.3202), np.float32(0.2779), np.float32(0.2913), np.float32(0.2674)] +2026-10-18 01:49:22.750548: Epoch time: 248.59 s +2026-10-18 01:49:23.694522: +2026-10-18 01:49:23.695287: Epoch 415 +2026-10-18 01:49:23.695692: Current learning rate: 0.00617 +2026-10-18 01:53:40.085551: train_loss -0.7375 +2026-10-18 01:53:40.086484: val_loss -0.3394 +2026-10-18 01:53:40.087269: Pseudo dice [np.float32(0.8595), np.float32(0.3218), np.float32(0.3132), np.float32(0.3968), np.float32(0.2852), np.float32(0.3895), np.float32(0.3321), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.3066), np.float32(0.3105), np.float32(0.3503), np.float32(0.3226), np.float32(0.0), np.float32(0.0), np.float32(0.3716), np.float32(0.2706), np.float32(0.3276), np.float32(0.2632), np.float32(0.1617), np.float32(0.2671), np.float32(0.2422), np.float32(0.4125), np.float32(0.2954), np.float32(0.3394), np.float32(0.0), np.float32(0.0), np.float32(0.0374), np.float32(0.3975), np.float32(0.0), np.float32(0.0), np.float32(0.2579), np.float32(0.2523), np.float32(0.308), np.float32(0.3312)] +2026-10-18 01:53:40.087738: Epoch time: 256.39 s +2026-10-18 01:53:41.045707: +2026-10-18 01:53:41.046366: Epoch 416 +2026-10-18 01:53:41.046699: Current learning rate: 0.00616 +2026-10-18 01:57:44.709212: train_loss -0.7426 +2026-10-18 01:57:44.709786: val_loss -0.3258 +2026-10-18 01:57:44.710471: Pseudo dice [np.float32(0.8783), np.float32(0.3243), np.float32(0.3058), np.float32(0.399), np.float32(0.2815), np.float32(0.3784), np.float32(0.2971), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.3164), np.float32(0.3403), np.float32(0.41), np.float32(0.3754), np.float32(0.0), np.float32(0.0), np.float32(0.4017), np.float32(0.3368), np.float32(0.2763), np.float32(0.2294), np.float32(0.1957), np.float32(0.2355), np.float32(0.3014), np.float32(0.3561), np.float32(0.3542), np.float32(0.3895), np.float32(0.0), np.float32(0.0), np.float32(0.157), np.float32(0.3828), np.float32(0.0), np.float32(0.0), np.float32(0.2758), np.float32(0.2112), np.float32(0.31), np.float32(0.3283)] +2026-10-18 01:57:44.710977: Epoch time: 243.67 s +2026-10-18 01:57:45.898015: +2026-10-18 01:57:45.898853: Epoch 417 +2026-10-18 01:57:45.899261: Current learning rate: 0.00615 +2026-10-18 02:01:42.074656: train_loss -0.7382 +2026-10-18 02:01:42.075450: val_loss -0.2425 +2026-10-18 02:01:42.076440: Pseudo dice [np.float32(0.8727), np.float32(0.1769), np.float32(0.2068), np.float32(0.2426), np.float32(0.1682), np.float32(0.2445), np.float32(0.2002), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.1621), np.float32(0.2013), np.float32(0.2777), np.float32(0.2698), np.float32(0.0), np.float32(0.0), np.float32(0.22), np.float32(0.1743), np.float32(0.213), np.float32(0.1931), np.float32(0.0279), np.float32(0.1592), np.float32(0.0594), np.float32(0.1986), np.float32(0.0985), np.float32(0.1142), np.float32(0.0), np.float32(0.0), np.float32(0.0882), np.float32(0.2064), np.float32(0.0), np.float32(0.0), np.float32(0.1451), np.float32(0.1762), np.float32(0.1471), np.float32(0.1506)] +2026-10-18 02:01:42.076879: Epoch time: 236.18 s +2026-10-18 02:01:43.044949: +2026-10-18 02:01:43.045734: Epoch 418 +2026-10-18 02:01:43.046135: Current learning rate: 0.00614 +2026-10-18 02:05:45.197248: train_loss -0.6916 +2026-10-18 02:05:45.198022: val_loss -0.4362 +2026-10-18 02:05:45.199032: Pseudo dice [np.float32(0.8666), np.float32(0.5213), np.float32(0.4951), np.float32(0.5774), np.float32(0.4765), np.float32(0.5371), np.float32(0.4995), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.5016), np.float32(0.503), np.float32(0.5579), np.float32(0.5207), np.float32(0.0), np.float32(0.0), np.float32(0.4941), np.float32(0.4174), np.float32(0.4923), np.float32(0.4452), np.float32(0.4735), np.float32(0.4529), np.float32(0.4414), np.float32(0.5118), np.float32(0.3605), np.float32(0.1933), np.float32(0.0), np.float32(0.0), np.float32(0.3226), np.float32(0.2925), np.float32(0.0), np.float32(0.0), np.float32(0.5185), np.float32(0.4707), np.float32(0.4987), np.float32(0.4438)] +2026-10-18 02:05:45.199742: Epoch time: 242.15 s +2026-10-18 02:05:46.135181: +2026-10-18 02:05:46.135847: Epoch 419 +2026-10-18 02:05:46.136053: Current learning rate: 0.00613 +2026-10-18 02:09:32.888106: train_loss -0.7223 +2026-10-18 02:09:32.889095: val_loss -0.2315 +2026-10-18 02:09:32.889706: Pseudo dice [np.float32(0.8999), np.float32(0.1901), np.float32(0.203), np.float32(0.2259), np.float32(0.1567), np.float32(0.2174), np.float32(0.1802), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.1853), np.float32(0.1524), np.float32(0.2501), np.float32(0.2505), np.float32(0.0), np.float32(0.0), np.float32(0.1754), np.float32(0.1273), np.float32(0.1669), np.float32(0.1332), np.float32(0.0367), np.float32(0.1052), np.float32(0.1359), np.float32(0.2387), np.float32(0.1332), np.float32(0.1484), np.float32(0.0), np.float32(0.0), np.float32(0.0605), np.float32(0.2758), np.float32(0.0), np.float32(0.0), np.float32(0.1418), np.float32(0.1258), np.float32(0.1535), np.float32(0.1533)] +2026-10-18 02:09:32.890136: Epoch time: 226.76 s +2026-10-18 02:09:33.862504: +2026-10-18 02:09:33.863223: Epoch 420 +2026-10-18 02:09:33.863657: Current learning rate: 0.00612 +2026-10-18 02:13:18.715401: train_loss -0.7252 +2026-10-18 02:13:18.716371: val_loss -0.4321 +2026-10-18 02:13:18.716906: Pseudo dice [np.float32(0.894), np.float32(0.4166), np.float32(0.4286), np.float32(0.4989), np.float32(0.3658), np.float32(0.4925), np.float32(0.4622), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.422), np.float32(0.3828), np.float32(0.4839), np.float32(0.474), np.float32(0.0), np.float32(0.0), np.float32(0.3314), np.float32(0.2749), np.float32(0.4277), np.float32(0.3842), np.float32(0.1973), np.float32(0.2839), np.float32(0.3267), np.float32(0.5263), np.float32(0.2634), np.float32(0.2752), np.float32(0.0), np.float32(0.0), np.float32(0.0864), np.float32(0.476), np.float32(0.0), np.float32(0.0), np.float32(0.3122), np.float32(0.3552), np.float32(0.4052), np.float32(0.4447)] +2026-10-18 02:13:18.717478: Epoch time: 224.86 s +2026-10-18 02:13:19.684162: +2026-10-18 02:13:19.684961: Epoch 421 +2026-10-18 02:13:19.685471: Current learning rate: 0.00612 +2026-10-18 02:17:26.957973: train_loss -0.7188 +2026-10-18 02:17:26.958827: val_loss -0.4942 +2026-10-18 02:17:26.959451: Pseudo dice [np.float32(0.8601), np.float32(0.5081), np.float32(0.5222), np.float32(0.6696), np.float32(0.5463), np.float32(0.5803), np.float32(0.6218), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.581), np.float32(0.5578), np.float32(0.5919), np.float32(0.5722), np.float32(0.0), np.float32(0.0), np.float32(0.5987), np.float32(0.4913), np.float32(0.5907), np.float32(0.5234), np.float32(0.4599), np.float32(0.445), np.float32(0.5098), np.float32(0.5511), np.float32(0.4121), np.float32(0.4165), np.float32(0.0), np.float32(0.0), np.float32(0.3667), np.float32(0.5293), np.float32(0.0), np.float32(0.0), np.float32(0.6049), np.float32(0.6069), np.float32(0.6282), np.float32(0.6284)] +2026-10-18 02:17:26.959774: Epoch time: 247.28 s +2026-10-18 02:17:27.867985: +2026-10-18 02:17:27.868725: Epoch 422 +2026-10-18 02:17:27.869097: Current learning rate: 0.00611 +2026-10-18 02:21:41.979764: train_loss -0.6935 +2026-10-18 02:21:41.980333: val_loss -0.5978 +2026-10-18 02:21:41.981190: Pseudo dice [np.float32(0.904), np.float32(0.7163), np.float32(0.6678), np.float32(0.8091), np.float32(0.6765), np.float32(0.8011), np.float32(0.7227), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.6354), np.float32(0.6386), np.float32(0.6262), np.float32(0.7179), np.float32(0.0), np.float32(0.0), np.float32(0.704), np.float32(0.5663), np.float32(0.6935), np.float32(0.6074), np.float32(0.5828), np.float32(0.5056), np.float32(0.7644), np.float32(0.7385), np.float32(0.3429), np.float32(0.4115), np.float32(0.0), np.float32(0.0), np.float32(0.3456), np.float32(0.3389), np.float32(0.0), np.float32(0.0), np.float32(0.747), np.float32(0.7411), np.float32(0.7773), np.float32(0.7867)] +2026-10-18 02:21:41.981565: Epoch time: 254.11 s +2026-10-18 02:21:42.924675: +2026-10-18 02:21:42.925485: Epoch 423 +2026-10-18 02:21:42.925858: Current learning rate: 0.0061 +2026-10-18 02:25:48.820301: train_loss -0.7257 +2026-10-18 02:25:48.821192: val_loss -0.3614 +2026-10-18 02:25:48.821721: Pseudo dice [np.float32(0.8952), np.float32(0.3826), np.float32(0.3752), np.float32(0.3939), np.float32(0.3526), np.float32(0.3526), np.float32(0.3874), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.3534), np.float32(0.2757), np.float32(0.3696), np.float32(0.3334), np.float32(0.0), np.float32(0.0), np.float32(0.4308), np.float32(0.3554), np.float32(0.3516), np.float32(0.3261), np.float32(0.395), np.float32(0.3215), np.float32(0.4663), np.float32(0.3866), np.float32(0.3163), np.float32(0.2698), np.float32(0.0), np.float32(0.0), np.float32(0.3063), np.float32(0.3916), np.float32(0.0), np.float32(0.0), np.float32(0.3286), np.float32(0.3613), np.float32(0.3995), np.float32(0.4329)] +2026-10-18 02:25:48.822314: Epoch time: 245.9 s +2026-10-18 02:25:49.743800: +2026-10-18 02:25:49.744425: Epoch 424 +2026-10-18 02:25:49.744855: Current learning rate: 0.00609 +2026-10-18 02:29:21.508482: train_loss -0.7072 +2026-10-18 02:29:21.509429: val_loss -0.6267 +2026-10-18 02:29:21.510000: Pseudo dice [np.float32(0.8617), np.float32(0.7479), np.float32(0.7302), np.float32(0.8237), np.float32(0.6875), np.float32(0.822), np.float32(0.7753), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.7053), np.float32(0.6676), np.float32(0.6701), np.float32(0.7147), np.float32(0.0), np.float32(0.0), np.float32(0.7124), np.float32(0.5836), np.float32(0.6915), np.float32(0.5629), np.float32(0.5055), np.float32(0.4682), np.float32(0.8698), np.float32(0.8155), np.float32(0.4127), np.float32(0.4974), np.float32(0.0), np.float32(0.0), np.float32(0.1844), np.float32(0.5422), np.float32(0.0), np.float32(0.0), np.float32(0.7455), np.float32(0.7214), np.float32(0.8268), np.float32(0.8339)] +2026-10-18 02:29:21.510504: Epoch time: 211.77 s +2026-10-18 02:29:22.447120: +2026-10-18 02:29:22.447638: Epoch 425 +2026-10-18 02:29:22.448144: Current learning rate: 0.00608 +2026-10-18 02:33:04.950082: train_loss -0.7177 +2026-10-18 02:33:04.951018: val_loss -0.5157 +2026-10-18 02:33:04.951351: Pseudo dice [np.float32(0.8811), np.float32(0.5586), np.float32(0.578), np.float32(0.6738), np.float32(0.5627), np.float32(0.6833), np.float32(0.6038), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.5818), np.float32(0.5701), np.float32(0.5473), np.float32(0.553), np.float32(0.0), np.float32(0.0), np.float32(0.5923), np.float32(0.5079), np.float32(0.5724), np.float32(0.4937), np.float32(0.4576), np.float32(0.3815), np.float32(0.7108), np.float32(0.6519), np.float32(0.3523), np.float32(0.4523), np.float32(0.0), np.float32(0.0), np.float32(0.2303), np.float32(0.4337), np.float32(0.0), np.float32(0.0), np.float32(0.5583), np.float32(0.5991), np.float32(0.6102), np.float32(0.6288)] +2026-10-18 02:33:04.952057: Epoch time: 222.51 s +2026-10-18 02:33:05.857443: +2026-10-18 02:33:05.858277: Epoch 426 +2026-10-18 02:33:05.858705: Current learning rate: 0.00607 +2026-10-18 02:36:52.872958: train_loss -0.7425 +2026-10-18 02:36:52.873576: val_loss -0.5908 +2026-10-18 02:36:52.874545: Pseudo dice [np.float32(0.8802), np.float32(0.656), np.float32(0.6758), np.float32(0.8021), np.float32(0.6707), np.float32(0.8014), np.float32(0.7346), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.6993), np.float32(0.6893), np.float32(0.63), np.float32(0.5934), np.float32(0.0), np.float32(0.0), np.float32(0.7108), np.float32(0.5815), np.float32(0.7029), np.float32(0.5744), np.float32(0.6482), np.float32(0.54), np.float32(0.7148), np.float32(0.6769), np.float32(0.4479), np.float32(0.5396), np.float32(0.0), np.float32(0.0), np.float32(0.4838), np.float32(0.5068), np.float32(0.0), np.float32(0.0), np.float32(0.7316), np.float32(0.75), np.float32(0.7706), np.float32(0.7675)] +2026-10-18 02:36:52.875068: Epoch time: 227.02 s +2026-10-18 02:36:53.851481: +2026-10-18 02:36:53.852261: Epoch 427 +2026-10-18 02:36:53.852709: Current learning rate: 0.00606 +2026-10-18 02:40:49.815949: train_loss -0.7424 +2026-10-18 02:40:49.816780: val_loss -0.4749 +2026-10-18 02:40:49.817436: Pseudo dice [np.float32(0.8832), np.float32(0.5339), np.float32(0.518), np.float32(0.604), np.float32(0.5079), np.float32(0.6076), np.float32(0.5385), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.5238), np.float32(0.5355), np.float32(0.5594), np.float32(0.5653), np.float32(0.0), np.float32(0.0), np.float32(0.559), np.float32(0.4698), np.float32(0.5216), np.float32(0.4782), np.float32(0.4697), np.float32(0.4225), np.float32(0.5477), np.float32(0.4753), np.float32(0.4598), np.float32(0.4715), np.float32(0.0), np.float32(0.0), np.float32(0.267), np.float32(0.4724), np.float32(0.0), np.float32(0.0), np.float32(0.5203), np.float32(0.5543), np.float32(0.5646), np.float32(0.5749)] +2026-10-18 02:40:49.817790: Epoch time: 235.97 s +2026-10-18 02:40:50.710313: +2026-10-18 02:40:50.711096: Epoch 428 +2026-10-18 02:40:50.711634: Current learning rate: 0.00605 +2026-10-18 02:45:01.054953: train_loss -0.7417 +2026-10-18 02:45:01.056034: val_loss -0.3849 +2026-10-18 02:45:01.056577: Pseudo dice [np.float32(0.8594), np.float32(0.3938), np.float32(0.3419), np.float32(0.4345), np.float32(0.3662), np.float32(0.4141), np.float32(0.3607), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.3436), np.float32(0.3564), np.float32(0.384), np.float32(0.4085), np.float32(0.0), np.float32(0.0), np.float32(0.44), np.float32(0.384), np.float32(0.3315), np.float32(0.3312), np.float32(0.2597), np.float32(0.2758), np.float32(0.3847), np.float32(0.3741), np.float32(0.3052), np.float32(0.384), np.float32(0.0), np.float32(0.0), np.float32(0.0618), np.float32(0.4133), np.float32(0.0), np.float32(0.0), np.float32(0.3127), np.float32(0.2873), np.float32(0.369), np.float32(0.3789)] +2026-10-18 02:45:01.056838: Epoch time: 250.35 s +2026-10-18 02:45:02.016215: +2026-10-18 02:45:02.017039: Epoch 429 +2026-10-18 02:45:02.017497: Current learning rate: 0.00604 +2026-10-18 02:48:55.483542: train_loss -0.7281 +2026-10-18 02:48:55.484381: val_loss -0.5623 +2026-10-18 02:48:55.484782: Pseudo dice [np.float32(0.8625), np.float32(0.6559), np.float32(0.6213), np.float32(0.7548), np.float32(0.6253), np.float32(0.7451), np.float32(0.6797), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.5957), np.float32(0.6166), np.float32(0.558), np.float32(0.5934), np.float32(0.0), np.float32(0.0), np.float32(0.6444), np.float32(0.5129), np.float32(0.6415), np.float32(0.554), np.float32(0.5846), np.float32(0.4926), np.float32(0.6942), np.float32(0.7168), np.float32(0.4891), np.float32(0.5667), np.float32(0.0), np.float32(0.0), np.float32(0.3249), np.float32(0.567), np.float32(0.0), np.float32(0.0), np.float32(0.6276), np.float32(0.6546), np.float32(0.7183), np.float32(0.7244)] +2026-10-18 02:48:55.485253: Epoch time: 233.47 s +2026-10-18 02:48:56.442529: +2026-10-18 02:48:56.443008: Epoch 430 +2026-10-18 02:48:56.443289: Current learning rate: 0.00603 +2026-10-18 02:53:00.464507: train_loss -0.7199 +2026-10-18 02:53:00.465328: val_loss -0.5518 +2026-10-18 02:53:00.465964: Pseudo dice [np.float32(0.8766), np.float32(0.6589), np.float32(0.6284), np.float32(0.7487), np.float32(0.6107), np.float32(0.7486), np.float32(0.6882), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.5984), np.float32(0.5966), np.float32(0.599), np.float32(0.6029), np.float32(0.0), np.float32(0.0), np.float32(0.6762), np.float32(0.5318), np.float32(0.6564), np.float32(0.5605), np.float32(0.5956), np.float32(0.5128), np.float32(0.5119), np.float32(0.5838), np.float32(0.5612), np.float32(0.5159), np.float32(0.0), np.float32(0.0), np.float32(0.206), np.float32(0.5451), np.float32(0.0), np.float32(0.0), np.float32(0.5613), np.float32(0.6116), np.float32(0.6949), np.float32(0.7097)] +2026-10-18 02:53:00.466536: Epoch time: 244.03 s +2026-10-18 02:53:01.426449: +2026-10-18 02:53:01.427226: Epoch 431 +2026-10-18 02:53:01.427758: Current learning rate: 0.00602 +2026-10-18 02:57:32.756357: train_loss -0.694 +2026-10-18 02:57:32.757356: val_loss -0.3699 +2026-10-18 02:57:32.757887: Pseudo dice [np.float32(0.8409), np.float32(0.4759), np.float32(0.4179), np.float32(0.3667), np.float32(0.3701), np.float32(0.398), np.float32(0.3502), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.2686), np.float32(0.2856), np.float32(0.2551), np.float32(0.2877), np.float32(0.0), np.float32(0.0), np.float32(0.4133), np.float32(0.3446), np.float32(0.2372), np.float32(0.2147), np.float32(0.3057), np.float32(0.3589), np.float32(0.4215), np.float32(0.4902), np.float32(0.3309), np.float32(0.2707), np.float32(0.0), np.float32(0.0), np.float32(0.097), np.float32(0.3227), np.float32(0.0), np.float32(0.0), np.float32(0.218), np.float32(0.2123), np.float32(0.3301), np.float32(0.3683)] +2026-10-18 02:57:32.758233: Epoch time: 271.33 s +2026-10-18 02:57:33.668172: +2026-10-18 02:57:33.668806: Epoch 432 +2026-10-18 02:57:33.669171: Current learning rate: 0.00601 +2026-10-18 03:01:35.594782: train_loss -0.6844 +2026-10-18 03:01:35.595563: val_loss -0.3552 +2026-10-18 03:01:35.596113: Pseudo dice [np.float32(0.8611), np.float32(0.3063), np.float32(0.352), np.float32(0.3988), np.float32(0.2949), np.float32(0.3187), np.float32(0.3109), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.3148), np.float32(0.3075), np.float32(0.3859), np.float32(0.3473), np.float32(0.0), np.float32(0.0), np.float32(0.3442), np.float32(0.3009), np.float32(0.2821), np.float32(0.2543), np.float32(0.3116), np.float32(0.334), np.float32(0.1076), np.float32(0.2182), np.float32(0.1525), np.float32(0.1002), np.float32(0.0), np.float32(0.0), np.float32(0.3259), np.float32(0.0086), np.float32(0.0), np.float32(0.0), np.float32(0.3119), np.float32(0.3146), np.float32(0.2682), np.float32(0.2437)] +2026-10-18 03:01:35.596670: Epoch time: 241.93 s +2026-10-18 03:01:36.534786: +2026-10-18 03:01:36.535470: Epoch 433 +2026-10-18 03:01:36.535835: Current learning rate: 0.006 +2026-10-18 03:05:43.715729: train_loss -0.715 +2026-10-18 03:05:43.716708: val_loss -0.2197 +2026-10-18 03:05:43.717261: Pseudo dice [np.float32(0.8942), np.float32(0.1674), np.float32(0.1886), np.float32(0.2157), np.float32(0.1285), np.float32(0.1828), np.float32(0.1564), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.1508), np.float32(0.1504), np.float32(0.2644), np.float32(0.1953), np.float32(0.0), np.float32(0.0), np.float32(0.1871), np.float32(0.1579), np.float32(0.1053), np.float32(0.0778), np.float32(0.0808), np.float32(0.1359), np.float32(0.1242), np.float32(0.2661), np.float32(0.1187), np.float32(0.1228), np.float32(0.0), np.float32(0.0), np.float32(0.0911), np.float32(0.3003), np.float32(0.0), np.float32(0.0), np.float32(0.1174), np.float32(0.1283), np.float32(0.1509), np.float32(0.1367)] +2026-10-18 03:05:43.717691: Epoch time: 247.18 s +2026-10-18 03:05:44.665823: +2026-10-18 03:05:44.666828: Epoch 434 +2026-10-18 03:05:44.667299: Current learning rate: 0.00599 +2026-10-18 03:09:24.577393: train_loss -0.7336 +2026-10-18 03:09:24.578429: val_loss -0.2046 +2026-10-18 03:09:24.579217: Pseudo dice [np.float32(0.8643), np.float32(0.1629), np.float32(0.1352), np.float32(0.1585), np.float32(0.102), np.float32(0.137), np.float32(0.0933), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.1075), np.float32(0.1274), np.float32(0.1825), np.float32(0.1854), np.float32(0.0), np.float32(0.0), np.float32(0.116), np.float32(0.12), np.float32(0.0683), np.float32(0.0566), np.float32(0.0237), np.float32(0.1152), np.float32(0.1234), np.float32(0.1489), np.float32(0.0996), np.float32(0.0775), np.float32(0.0), np.float32(0.0), np.float32(0.0632), np.float32(0.2106), np.float32(0.0), np.float32(0.0), np.float32(0.0781), np.float32(0.0809), np.float32(0.1094), np.float32(0.0885)] +2026-10-18 03:09:24.579619: Epoch time: 219.91 s +2026-10-18 03:09:25.502715: +2026-10-18 03:09:25.503411: Epoch 435 +2026-10-18 03:09:25.503776: Current learning rate: 0.00598 +2026-10-18 03:13:33.103999: train_loss -0.7393 +2026-10-18 03:13:33.104850: val_loss -0.2702 +2026-10-18 03:13:33.105512: Pseudo dice [np.float32(0.8907), np.float32(0.167), np.float32(0.2231), np.float32(0.2149), np.float32(0.1366), np.float32(0.225), np.float32(0.1911), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.1511), np.float32(0.1342), np.float32(0.2985), np.float32(0.1486), np.float32(0.0), np.float32(0.0), np.float32(0.1613), np.float32(0.142), np.float32(0.1724), np.float32(0.1256), np.float32(0.066), np.float32(0.1673), np.float32(0.1303), np.float32(0.2764), np.float32(0.1253), np.float32(0.1213), np.float32(0.0), np.float32(0.0), np.float32(0.0265), np.float32(0.2467), np.float32(0.0), np.float32(0.0), np.float32(0.1228), np.float32(0.1773), np.float32(0.1712), np.float32(0.1842)] +2026-10-18 03:13:33.105972: Epoch time: 247.6 s +2026-10-18 03:13:34.060857: +2026-10-18 03:13:34.061512: Epoch 436 +2026-10-18 03:13:34.061901: Current learning rate: 0.00597 +2026-10-18 03:17:28.748678: train_loss -0.7414 +2026-10-18 03:17:28.749623: val_loss -0.1978 +2026-10-18 03:17:28.750143: Pseudo dice [np.float32(0.8986), np.float32(0.1138), np.float32(0.1542), np.float32(0.162), np.float32(0.1088), np.float32(0.148), np.float32(0.1389), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.1101), np.float32(0.0944), np.float32(0.1687), np.float32(0.1618), np.float32(0.0), np.float32(0.0), np.float32(0.1234), np.float32(0.0926), np.float32(0.1022), np.float32(0.0858), np.float32(0.0023), np.float32(0.1072), np.float32(0.1112), np.float32(0.2526), np.float32(0.0741), np.float32(0.0757), np.float32(0.0), np.float32(0.0), np.float32(0.0455), np.float32(0.2198), np.float32(0.0), np.float32(0.0), np.float32(0.0816), np.float32(0.1181), np.float32(0.1319), np.float32(0.1124)] +2026-10-18 03:17:28.750707: Epoch time: 234.69 s +2026-10-18 03:17:29.677957: +2026-10-18 03:17:29.678537: Epoch 437 +2026-10-18 03:17:29.678920: Current learning rate: 0.00596 +2026-10-18 03:21:37.366629: train_loss -0.739 +2026-10-18 03:21:37.367141: val_loss -0.124 +2026-10-18 03:21:37.367770: Pseudo dice [np.float32(0.9045), np.float32(0.0244), np.float32(0.0429), np.float32(0.0239), np.float32(0.0156), np.float32(0.0423), np.float32(0.0314), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.0272), np.float32(0.0743), np.float32(0.1045), np.float32(0.0507), np.float32(0.0), np.float32(0.0), np.float32(0.0204), np.float32(0.0127), np.float32(0.0317), np.float32(0.0509), np.float32(0.0002), np.float32(0.0376), np.float32(0.1153), np.float32(0.0451), np.float32(0.0114), np.float32(0.0215), np.float32(0.0), np.float32(0.0), np.float32(0.0206), np.float32(0.0509), np.float32(0.0), np.float32(0.0), np.float32(0.012), np.float32(0.0359), np.float32(0.023), np.float32(0.0345)] +2026-10-18 03:21:37.368609: Epoch time: 247.69 s +2026-10-18 03:21:38.260566: +2026-10-18 03:21:38.261210: Epoch 438 +2026-10-18 03:21:38.261581: Current learning rate: 0.00595 +2026-10-18 03:25:24.450655: train_loss -0.7411 +2026-10-18 03:25:24.451391: val_loss -0.2422 +2026-10-18 03:25:24.451992: Pseudo dice [np.float32(0.862), np.float32(0.1174), np.float32(0.1945), np.float32(0.1935), np.float32(0.1294), np.float32(0.1915), np.float32(0.1888), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.1446), np.float32(0.1271), np.float32(0.2518), np.float32(0.227), np.float32(0.0), np.float32(0.0), np.float32(0.1518), np.float32(0.1139), np.float32(0.1536), np.float32(0.1335), np.float32(0.0573), np.float32(0.1204), np.float32(0.189), np.float32(0.2528), np.float32(0.0566), np.float32(0.0858), np.float32(0.0), np.float32(0.0), np.float32(0.0865), np.float32(0.1959), np.float32(0.0), np.float32(0.0), np.float32(0.1095), np.float32(0.1316), np.float32(0.1416), np.float32(0.1225)] +2026-10-18 03:25:24.452888: Epoch time: 226.19 s +2026-10-18 03:25:25.357230: +2026-10-18 03:25:25.357783: Epoch 439 +2026-10-18 03:25:25.358199: Current learning rate: 0.00594 +2026-10-18 03:29:26.179222: train_loss -0.7481 +2026-10-18 03:29:26.180145: val_loss -0.1913 +2026-10-18 03:29:26.180688: Pseudo dice [np.float32(0.8973), np.float32(0.1068), np.float32(0.1465), np.float32(0.1777), np.float32(0.1084), np.float32(0.1543), np.float32(0.1366), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.1188), np.float32(0.1096), np.float32(0.1993), np.float32(0.1293), np.float32(0.0), np.float32(0.0), np.float32(0.1369), np.float32(0.1083), np.float32(0.1093), np.float32(0.0749), np.float32(0.0002), np.float32(0.1077), np.float32(0.0692), np.float32(0.1591), np.float32(0.0575), np.float32(0.0763), np.float32(0.0), np.float32(0.0), np.float32(0.028), np.float32(0.2109), np.float32(0.0), np.float32(0.0), np.float32(0.0781), np.float32(0.0954), np.float32(0.1151), np.float32(0.1093)] +2026-10-18 03:29:26.181242: Epoch time: 240.82 s +2026-10-18 03:29:27.398270: +2026-10-18 03:29:27.399043: Epoch 440 +2026-10-18 03:29:27.399541: Current learning rate: 0.00593 +2026-10-18 03:33:33.061846: train_loss -0.744 +2026-10-18 03:33:33.062919: val_loss -0.3082 +2026-10-18 03:33:33.063580: Pseudo dice [np.float32(0.8834), np.float32(0.2505), np.float32(0.2606), np.float32(0.3116), np.float32(0.1908), np.float32(0.2852), np.float32(0.2618), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.2312), np.float32(0.2122), np.float32(0.3012), np.float32(0.2456), np.float32(0.0), np.float32(0.0), np.float32(0.2702), np.float32(0.2216), np.float32(0.2326), np.float32(0.1931), np.float32(0.0652), np.float32(0.2079), np.float32(0.2421), np.float32(0.3613), np.float32(0.2288), np.float32(0.2364), np.float32(0.0), np.float32(0.0), np.float32(0.0325), np.float32(0.3067), np.float32(0.0), np.float32(0.0), np.float32(0.1673), np.float32(0.1751), np.float32(0.2046), np.float32(0.2345)] +2026-10-18 03:33:33.063852: Epoch time: 245.67 s +2026-10-18 03:33:33.966146: +2026-10-18 03:33:33.966964: Epoch 441 +2026-10-18 03:33:33.967338: Current learning rate: 0.00592 +2026-10-18 03:37:47.764162: train_loss -0.7503 +2026-10-18 03:37:47.765191: val_loss -0.3409 +2026-10-18 03:37:47.765851: Pseudo dice [np.float32(0.8986), np.float32(0.3523), np.float32(0.3159), np.float32(0.3871), np.float32(0.2949), np.float32(0.3613), np.float32(0.3126), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.3015), np.float32(0.309), np.float32(0.3353), np.float32(0.3191), np.float32(0.0), np.float32(0.0), np.float32(0.3663), np.float32(0.2964), np.float32(0.2426), np.float32(0.2107), np.float32(0.1974), np.float32(0.2379), np.float32(0.2532), np.float32(0.3262), np.float32(0.2919), np.float32(0.2852), np.float32(0.0), np.float32(0.0), np.float32(0.0536), np.float32(0.3497), np.float32(0.0), np.float32(0.0), np.float32(0.27), np.float32(0.2277), np.float32(0.2766), np.float32(0.2907)] +2026-10-18 03:37:47.766220: Epoch time: 253.8 s +2026-10-18 03:37:48.702439: +2026-10-18 03:37:48.703205: Epoch 442 +2026-10-18 03:37:48.703641: Current learning rate: 0.00592 +2026-10-18 03:42:00.744878: train_loss -0.7541 +2026-10-18 03:42:00.745720: val_loss -0.2671 +2026-10-18 03:42:00.746262: Pseudo dice [np.float32(0.8784), np.float32(0.2335), np.float32(0.2316), np.float32(0.2776), np.float32(0.1707), np.float32(0.2532), np.float32(0.2149), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.1994), np.float32(0.209), np.float32(0.2775), np.float32(0.2882), np.float32(0.0), np.float32(0.0), np.float32(0.2477), np.float32(0.1952), np.float32(0.174), np.float32(0.1229), np.float32(0.088), np.float32(0.1749), np.float32(0.1985), np.float32(0.2788), np.float32(0.145), np.float32(0.1611), np.float32(0.0), np.float32(0.0), np.float32(0.0523), np.float32(0.3047), np.float32(0.0), np.float32(0.0), np.float32(0.1609), np.float32(0.1575), np.float32(0.1921), np.float32(0.188)] +2026-10-18 03:42:00.746870: Epoch time: 252.04 s +2026-10-18 03:42:01.701259: +2026-10-18 03:42:01.701948: Epoch 443 +2026-10-18 03:42:01.702380: Current learning rate: 0.00591 +2026-10-18 03:46:02.987016: train_loss -0.7486 +2026-10-18 03:46:02.987565: val_loss -0.3618 +2026-10-18 03:46:02.988565: Pseudo dice [np.float32(0.9047), np.float32(0.3194), np.float32(0.3529), np.float32(0.4178), np.float32(0.3289), np.float32(0.396), np.float32(0.3594), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.318), np.float32(0.3049), np.float32(0.2968), np.float32(0.2656), np.float32(0.0), np.float32(0.0), np.float32(0.3719), np.float32(0.2991), np.float32(0.3327), np.float32(0.2803), np.float32(0.2896), np.float32(0.2746), np.float32(0.4322), np.float32(0.5317), np.float32(0.2418), np.float32(0.3409), np.float32(0.0), np.float32(0.0), np.float32(0.0923), np.float32(0.3571), np.float32(0.0), np.float32(0.0), np.float32(0.2731), np.float32(0.2654), np.float32(0.3614), np.float32(0.3606)] +2026-10-18 03:46:02.989052: Epoch time: 241.29 s +2026-10-18 03:46:03.873766: +2026-10-18 03:46:03.874259: Epoch 444 +2026-10-18 03:46:03.875647: Current learning rate: 0.0059 +2026-10-18 03:50:06.583655: train_loss -0.7262 +2026-10-18 03:50:06.584360: val_loss -0.2686 +2026-10-18 03:50:06.584953: Pseudo dice [np.float32(0.8839), np.float32(0.1657), np.float32(0.2433), np.float32(0.0997), np.float32(0.2266), np.float32(0.1782), np.float32(0.2875), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.1722), np.float32(0.1622), np.float32(0.2658), np.float32(0.2831), np.float32(0.0), np.float32(0.0), np.float32(0.2722), np.float32(0.2414), np.float32(0.2593), np.float32(0.2213), np.float32(0.2312), np.float32(0.281), np.float32(0.2094), np.float32(0.2145), np.float32(0.0739), np.float32(0.0783), np.float32(0.0), np.float32(0.0), np.float32(0.0143), np.float32(0.3089), np.float32(0.0), np.float32(0.0), np.float32(0.1642), np.float32(0.1756), np.float32(0.1488), np.float32(0.199)] +2026-10-18 03:50:06.585199: Epoch time: 242.71 s +2026-10-18 03:50:07.477435: +2026-10-18 03:50:07.478193: Epoch 445 +2026-10-18 03:50:07.478608: Current learning rate: 0.00589 +2026-10-18 03:54:17.581663: train_loss -0.7406 +2026-10-18 03:54:17.582606: val_loss -0.4204 +2026-10-18 03:54:17.583340: Pseudo dice [np.float32(0.8921), np.float32(0.3746), np.float32(0.4153), np.float32(0.4807), np.float32(0.4065), np.float32(0.4391), np.float32(0.4313), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.4127), np.float32(0.3598), np.float32(0.4181), np.float32(0.3303), np.float32(0.0), np.float32(0.0), np.float32(0.4586), np.float32(0.3966), np.float32(0.4288), np.float32(0.3521), np.float32(0.3859), np.float32(0.3692), np.float32(0.3313), np.float32(0.5548), np.float32(0.2559), np.float32(0.2298), np.float32(0.0), np.float32(0.0), np.float32(0.1793), np.float32(0.4164), np.float32(0.0), np.float32(0.0), np.float32(0.3365), np.float32(0.284), np.float32(0.3923), np.float32(0.3656)] +2026-10-18 03:54:17.583922: Epoch time: 250.11 s +2026-10-18 03:54:18.464479: +2026-10-18 03:54:18.465128: Epoch 446 +2026-10-18 03:54:18.465522: Current learning rate: 0.00588 +2026-10-18 03:58:38.824267: train_loss -0.7485 +2026-10-18 03:58:38.825107: val_loss -0.2218 +2026-10-18 03:58:38.825752: Pseudo dice [np.float32(0.8558), np.float32(0.1957), np.float32(0.1724), np.float32(0.2414), np.float32(0.177), np.float32(0.1758), np.float32(0.1724), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.1928), np.float32(0.134), np.float32(0.2824), np.float32(0.2112), np.float32(0.0), np.float32(0.0), np.float32(0.2458), np.float32(0.2015), np.float32(0.1573), np.float32(0.1018), np.float32(0.1055), np.float32(0.124), np.float32(0.06), np.float32(0.1711), np.float32(0.0574), np.float32(0.0536), np.float32(0.0), np.float32(0.0), np.float32(0.0244), np.float32(0.1829), np.float32(0.0), np.float32(0.0), np.float32(0.1372), np.float32(0.1345), np.float32(0.1191), np.float32(0.0949)] +2026-10-18 03:58:38.826321: Epoch time: 260.36 s +2026-10-18 03:58:39.763605: +2026-10-18 03:58:39.764344: Epoch 447 +2026-10-18 03:58:39.764658: Current learning rate: 0.00587 +2026-10-18 04:02:32.604039: train_loss -0.7432 +2026-10-18 04:02:32.605157: val_loss -0.2991 +2026-10-18 04:02:32.605731: Pseudo dice [np.float32(0.8864), np.float32(0.2635), np.float32(0.2667), np.float32(0.222), np.float32(0.2055), np.float32(0.19), np.float32(0.2279), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.1948), np.float32(0.2322), np.float32(0.2968), np.float32(0.3077), np.float32(0.0), np.float32(0.0), np.float32(0.3111), np.float32(0.2316), np.float32(0.2627), np.float32(0.2855), np.float32(0.1611), np.float32(0.2427), np.float32(0.3617), np.float32(0.3429), np.float32(0.2375), np.float32(0.215), np.float32(0.0), np.float32(0.0), np.float32(0.0764), np.float32(0.3596), np.float32(0.0), np.float32(0.0), np.float32(0.1667), np.float32(0.1335), np.float32(0.2402), np.float32(0.2448)] +2026-10-18 04:02:32.606115: Epoch time: 232.84 s +2026-10-18 04:02:33.493236: +2026-10-18 04:02:33.493891: Epoch 448 +2026-10-18 04:02:33.494202: Current learning rate: 0.00586 +2026-10-18 04:06:40.568619: train_loss -0.7536 +2026-10-18 04:06:40.569256: val_loss -0.4089 +2026-10-18 04:06:40.570437: Pseudo dice [np.float32(0.8989), np.float32(0.4308), np.float32(0.4331), np.float32(0.4742), np.float32(0.4202), np.float32(0.4787), np.float32(0.4168), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.3743), np.float32(0.422), np.float32(0.377), np.float32(0.4075), np.float32(0.0), np.float32(0.0), np.float32(0.4533), np.float32(0.3916), np.float32(0.3432), np.float32(0.3204), np.float32(0.3058), np.float32(0.2826), np.float32(0.5916), np.float32(0.5875), np.float32(0.3256), np.float32(0.4173), np.float32(0.0), np.float32(0.0), np.float32(0.1078), np.float32(0.3978), np.float32(0.0), np.float32(0.0), np.float32(0.3322), np.float32(0.3516), np.float32(0.4604), np.float32(0.471)] +2026-10-18 04:06:40.570854: Epoch time: 247.08 s +2026-10-18 04:06:41.496461: +2026-10-18 04:06:41.497374: Epoch 449 +2026-10-18 04:06:41.497904: Current learning rate: 0.00585 +2026-10-18 04:10:46.853838: train_loss -0.7541 +2026-10-18 04:10:46.854704: val_loss -0.3314 +2026-10-18 04:10:46.855193: Pseudo dice [np.float32(0.8778), np.float32(0.3018), np.float32(0.3059), np.float32(0.3279), np.float32(0.293), np.float32(0.3248), np.float32(0.3122), np.float32(0.0), np.float32(0.0), np.float32(0.0), np.float32(0.2721), np.float32(0.3209), np.float32(0.3238), np.float32(0.2436), np.float32(0.0), np.float32(0.0), np.float32(0.3417), np.float32(0.2905), np.float32(0.2561), np.float32(0.2578), np.float32(0.192), np.float32(0.1799), np.float32(0.4495), np.float32(0.35), np.float32(0.2433), np.float32(0.3301), np.float32(0.0), np.float32(0.0), np.float32(0.0811), np.float32(0.2855), np.float32(0.0), np.float32(0.0), np.float32(0.2254), np.float32(0.2121), np.float32(0.2799), np.float32(0.2869)] +2026-10-18 04:10:46.855645: Epoch time: 245.36 s +2026-10-18 04:10:48.590670: +2026-10-18 04:10:48.591512: Epoch 450 +2026-10-18 04:10:48.592068: Current learning rate: 0.00584