With specified property name and information, get the expected property value. Args: name: the name of target property.

Helper for getting padding (nn.ConstantPadNd) to be used to get SAME padding conv operations similar to Tensorflow's SAME padding. This

Compute scan interval according to the image size, roi size and overlap. Scan interval will be `int((1 - overlap) * roi_size)`, if interval i

Get the array data and metadata of the segmentation image. Aegs: img: a Pydicom dataset object that has attribute "Segme

Get the spatial shape of `img`. Args: img: an ITK image object loaded from an image file.

Get the spatial shape of image data, it doesn't contain the channel dim. Args: img: a Nibabel image object loaded from a

Get the spatial shape of image data, it doesn't contain the channel dim. Args: img: a PIL Image object loaded from an ima

In each up layer, the input is passed through a convolution block before upsampled, unless this is the top layer in which case the in

Returns the decoding (up) part of a layer of the network. This typically will upsample data at some point in its structure. Its outpu

Args: net_input: the original input of the network. vae_input: the input of VAE module, which is also the output of t

Parse the expression and discover what variables are present in it based on ast module. Args: expr: source expression to parse.

Compute the variance of the posterior at timestep t. Args: timestep: current timestep. predicted_variance: v

Helper function to get the window index.

utility function to handle batched MetaTensors.

Implementation based on PyTorch ignite package, it can support more kinds of backends.

Function to actually create the animated gif. Args: tag: Data identifier image: 3D image tensors expected to be in `HWD` format

Create cache of the identity grid if cache_grid=True and spatial_size is known.

similar to monai/networks/blocks/patchembedding.py for the decoder positional encoding and for mask and classification tokens

Args: None, initializes weights for conv/linear/batchnorm layers following weight init methods from `offi

Return True if the read thread should be still active.

Formats the sequence of intensities ranges to Sequence[Sequence[float]].

Matching result with not ground truth in image Args: iou_thresholds: defined which IoU thresholds should be evaluated dt_sco

Matching result with no predictions Args: iou_thresholds: defined which IoU thresholds should be evaluated gt_ignore: specif

Adapted from https://github.com/cocodataset/cocoapi/blob/master/PythonAPI/pycocotools/cocoeval.py Args: iou_fn: compute overlap for

Compute the MSE loss of the gradients of the horizontal and vertical centroid distance maps

Tensor initialization with truncated normal distribution. Based on: https://people.sc.fsu.edu/~jburkardt/presentations/truncated_normal.pdf

Utility function based on `decollate_batch`, to identify the largest batch size from the collated data. returns batch_size, the list of non-i

Builds TRT engine from ONNX file at onnx_path and saves to self.plan_path

This function takes an ONNX model as input, exports it to a TensorRT engine, wraps the TensorRT engine to a TensorRT engine-based TorchScript

s -> (s0,s1), (s1,s2), (s2, s3), ...

Log artifacts to mlflow. Given a path, all files in the path will be logged recursively. Given a file, it will be logged to mlflow.

parse single import statement such as "from monai.transforms import Resize

Process the data from before the first random transform to the final state ready for evaluation. Args: item_transformed:

Register the decollate operation for batch data, will execute after model forward and loss forward.

Register the handlers to the engine, supports ignite Handlers with `attach` API.

Register the key metric and additional metrics to the engine, supports ignite Metrics.

Register the postprocessing logic to the engine, will execute them as a chain when iteration completed.

Concat network output map list to a single Tensor. This function is used in both training and inference. Args: r

Restart background thread if killed for some reason.

Re-calculate module's repeat number of a block based on depth coefficient multiplier. Args: repeats: number of original repeats.

Execute the training command with target devices information.

Load the object from the given bytes data, this must be loadable as weights only using `torch.load`.

Serialize the tensor/array `val` using the pickle protocol, and return its bytes object.

Silhouette-based inter-class separation score. Requires scikit-learn. Returns ``None`` if unavailable. Args: emb: ``[N, D]`` float t

Sets default intensity ranges to be sampled. Args: img: image to transform.

Set learning rate(s) for optimizer.

With specified property name and information, set value for the expected property. Args: name: the name of target proper

Set the WSI reader object based on the input reader Args: reader: the module to be used for loading whole slide imaging.