Check and ensure the numpy array or PyTorch Tensor in data to be contiguous in memory. Args: data: input data to convert, will recur

Utility to correct the crop center if the crop size and centers are not compatible with the image size. Args: centers: pre-computed

Factory method to create MedNeXt variants. Args: variant (str): The MedNeXt variant to create ('S', 'B', 'M', or 'L'). spati

Based on a latent space sample, forwards it through the Decoder. Args: z: Bx[Z_CHANNELS]x[LATENT SPACE SHAPE]

From a set of original boxes and encoded relative box offsets, Args: rel_codes: encoded boxes, Nx(4*num_box_reg) or Nx(6

Apply transform on `batch` and `output`. If `batch` and `output` are dictionaries, temporarily combine them for the transform, otherwise,

An in-place function. We expect ``head_outputs`` to be Dict[str, List[Tensor]]. Yet if it is Dict[str, Tensor], this func converts it to Dict

Execute the current config content and return the result if it is expression, based on Python `eval()`. For more details: https://doc

Save the cache state as ``cache.yaml`` in the working directory

Similar to np.fft.fftshift but applies to PyTorch Tensors Args: x: input data (k-space or image) that can be 1) real-val

Finalize merging by dividing values by counts and return the merged tensor. Notes: To avoid creating a new tensor for th

Args: x: input tensor weight: weights for different operations.

Returns a dict of losses during training, or a list predicted dict of boxes and labels during inference. Args: input_ima

Args: input: the shape should be BNH[WD], where N is the number of classes. target: the shape should be BNH[WD] or B1

Args: input: the shape should be BNH[WD]. target: the shape should be BNH[WD]. mask: the shape should B1H

Args: input: the shape should be BNH[WD]. target: the shape should be BNH[WD]. Raises: ValueErro

Generate the spatial bounding box of foreground in the image with start-end positions (inclusive). Users can define arbitrary function to sel

Get all information (include "name" and "browser_download_url") of a bundle with the specified bundle name and version which is stored in the

Compute confusion matrix. A tensor with the shape [BC4] will be returned. Where, the third dimension represents the number of true positive,

Generate dataset based on the specified fold indices in the cross validation group. Args: folds: index of folds for trai

Get next frame. For a file, this will be the next frame, whereas for a camera source, it will be the next available frame.

Get image. If is dictionary, extract key. If is list, stack. If both dictionary and list, do both. Also return the image size as string to be used

Gets the largest connected component mask of an image. Args: img: Image to get largest connected component from. Shape is (spatial_d

Computes the log-likelihoods of the latent representations of the input. Args: inputs: input images, NxCxHxW[xD]

Computes the log-likelihoods of the latent representations of the input. Args: inputs: input images, NxCxHxW[xD]

Return the filename of the dumped algo object.

Get final path and child model probabilities from architecture weights `log_alpha_a`. This is used in forward pass, getting training

This is equivalent to self.inner_blocks[idx](x), but torchscript doesn't support this yet

Returns the size (height, width) of the whole slide image at a given level. Args: wsi: a whole slide image object loaded

Get the most recently applied transformation matrix

Computing window size based on: "Liu et al., Swin Transformer: Hierarchical Vision Transformer using Shifted Windows <https://arxiv.org/abs/21

Pytorch-based ifft for spatial_dims-dim signals. "centered" means this function automatically takes care of the required ifft and fft shifts.

Pytorch-based ifft for spatial_dims-dim signals. "centered" means this function automatically takes care of the required ifft and fft shifts.

Similar to np.fft.ifftshift but applies to PyTorch Tensors Args: x: input data (k-space or image) that can be 1) real-va

Check if the data is an affine matrix.

Returns whether `val` indicates "no_channel", for MetaKeys.ORIGINAL_CHANNEL_DIM.

Yield successive tuples of upper left corner of patches of size `patch_size` from an array of dimensions `image_size`. The iteration starts f

Iterate over the sub-configs of the input config, the output `sub_id` uses `cls.sep` to denote substructure. Args: id: i

Compute the foreground and background of input label data, return the indices after fattening. For example: ``label = np.array([[[0, 1, 1

Filter out indices of every class of the input label data, return the indices after fattening. It can handle both One-Hot format label and Ar

Create a 3D volume figure as a grid of images. Args: volume: 3D volume to display. data shape can be `BCHWD`, `CHWD` or `HWD`.

`np.nonzero` with equivalent implementation for torch. Args: x: array/tensor. Returns: Index unravelled for given shape

Compute an affine matrix according to the input shape. The transform normalizes the homogeneous image coordinates to the range of `[-1, 1

Return number of anchor shapes for each feature map.

Args: item: data item to be hashed protocol: protocol version used for pickling, defaults to `pickle.HIGHEST_PROTOCO

Return a + if we don't already have one, else return a .

Return a + if we don't already have one, else return a .

Apply a set of preprocessing operations to the input data before the training. Args: overwrite_plans_name: [OPTIONAL] Yo

Creates a context to profile, placing a timing result onto the queue when it exits.

Find best n idle devices and return a string of device ids using the `nvidia-smi` command.

Encodes and decodes an input image. Args: x: BxCx[SPATIAL DIMENSIONS] tensor. Returns: reconstructe

Register a given class to the encoder dict. Please notice that input class must be a subclass of BaseEncoder.

Temporarily replace sub-module(s) in a parent module (context manager). See :py:class:`monai.networks.utils.replace_modules`.

find MetaTensors in list or dict `data` and (in-place) set ``TraceKeys.ID`` to ``Tracekeys.NONE``.

Constructs a feature extractor network with a ResNet-FPN backbone, used as feature_extractor in RetinaNet. Reference: `"Focal Loss for Dense

Run the AutoRunner pipeline

Run the bundle workflow, it can be a training, evaluation or inference. Before run, we add bundle root directory to Python search dir

Set the input data and delete all the out-dated cache content.

Set the training params for all algos. Args: params: a dict that defines the overriding key-value pairs during training.

Applies fourier transform and shifts the zero-frequency component to the center of the spectrum. Only the spatial dimensions get tran

`np.stack` with equivalent implementation for torch. Args: x: array/tensor. dim: dimension along which to perform the stack (refe

Predict the sample at the previous timestep by reversing the SDE. Core function to propagate the diffusion process from the learned m

Store a given object with the given key name. Args: key: key of the data object to store. data_obj: data obj

Convert a string to a list. Useful with argparse commandline arguments: parser.add_argument("--blocks", default=[1,2,3], type=str2list)

Load the run function in the training script of each model. Training parameter is predefined by the algo_config.yaml file, which is p

Set network(s) to train mode and then return to original state at the end. Args: nets: Input network(s) Examples .. code-b

Instruments model or submodule(s) with TrtCompiler and replaces its forward() with TRT hook. Note: TRT 10.3 is recommended for best performan

`torch.unique` with equivalent implementation for numpy. Args: x: array/tensor.

Replace the content of `self.config` with new `config`. A typical usage is to modify the initial config content at runtime.

Args: img: channel first array, must have shape: (num_channels, H[, W, ..., ]) lazy: a flag to indicate whether this

Args: spatial_dims: number of spatial dimensions. in_chns: number of input channels. out_chns: number of