This only applies for num_conv_per_stage and convolutional_upsampling=True not real vram consumption. just a constant term to which t

:param spacing: :param patch_size: :param min_feature_map_size: min edge length of feature maps in bottleneck :return:

unpacks all npz files in a folder to npy (whatever you want to have unpacked must be saved unter key) :param folder: :param threads:

This is what actually configures the entire experiment. The plans file is generated by experiment planning :return:

:param spacing: :param patch_size: :param min_feature_map_size: min edge length of feature maps in bottleneck :return:

net_output must be (b, c, x, y(, z))) gt must be a label map (shape (b, 1, x, y(, z)) OR shape (b, x, y(, z))) or one hot encoding (b, c, x,

test = predicted image :param test_ref_pairs: :param evaluator: :param labels: must be a dict of int-> str or a list of int :para

returns a dictionary that contains pointers to conv, nonlin and norm ops and the default kwargs I like to use :return:

:param data: :param seg: :param original_spacing: :param target_spacing: :param order_data: :param order_seg: :param forc

Set the reference segmentation.

Set the test segmentation.

square_volumes will square the weight term. The paper recommends square_volumes=True; I don't (just an intuition)

The default split is a 5 fold CV on all available training cases. nnU-Net will create a split (it is seeded, so always the same) and

We need to wrap this because we need to enforce self.network.do_ds = False for prediction

:param base: :param gt_labels_folder: subfolder of base with niftis of ground truth labels :param raw_subfolder_name: subfolder of base w

this is only valid for the brats data in here where the labels are 1, 2, and 3. The original brats data have a different labeling convention!

Retrieves the correct output directory for the nnU-Net model described by the input parameters :param model: :param task: :param tra

- replaced get_default_augmentation with get_moreDA_augmentation - enforce to only run this code once - loss function wrapper

used for if the checkpoint is already in ram :param checkpoint: :param train: :return:

loads the relevant part of the pkl file that is needed for applying postprocessing :param pkl_file: :return:

separate_z=True will resample with order 0 along z :param data: :param new_shape: :param is_seg: :param axis: :param order:

Reads an image (must be a format that it recognized by skimage.io.imread) and converts it into a series of niftis. The image can have an arbi

- momentum 0.99 - SGD instead of Adam - self.lr_scheduler = None because we do poly_lr - deep supervision = True

This is a utility for writing segmentations to nifto and npz. It requires the data to have been preprocessed by GenericPreprocessor because i

This one does fully convolutional inference. No sliding window

applies removing of all but the largest connected component to all niftis in a folder :param min_valid_object_size: :param min_valid_obje

ref and gt are binary integer numpy.ndarray s spacing is assumed to be (1, 1, 1) :param ref: :param pred: :return:

ref and gt are binary integer numpy.ndarray s :param ref: :param gt: :return:

Used to determine the postprocessing for an experiment after all five folds have been completed. In the validation of each fold, the postproc

Construct label set from unique entries in segmentations.

reads all prediction files (nifti) in the input folder, converts the labels back to BraTS convention and saves the result in output_folder