Operators
Operators¶
Function |
Supported Device |
Description |
---|---|---|
HPU, CPU |
Performs element-wise addition and supports |
|
CPU |
Crops input tensor along width, height and depth. |
|
HPU |
Computes the output by doing bitwise |
|
HPU |
Computes the output by doing bitwise |
|
HPU |
Computes the output by doing bitwise |
|
HPU |
Defines the brightness of the image. |
|
HPU |
Changes the data type of the input tensor. |
|
HPU |
Clamps all elements in input into the range [lowerBound, upperBound]. |
|
CPU |
Reader for SSD Coco dataset which produces outputs as images and metadata. |
|
HPU |
Transforms the image from one color representation to another color representation. |
|
HPU |
Concatenates tensors along one dimension. |
|
HPU, CPU |
Outputs a tensor with random boolean values using a Bernoulli distribution. |
|
HPU, CPU |
Constant operator generates scalar constant value tensor. |
|
HPU |
Defines the contrast of the image. |
|
HPU, CPU |
Crops input tensor with specified window dimensions and window position. |
|
HPU |
Performs fused cropping, mirroring, normalization, and type casting. |
|
CPU |
Allows user to add existing CPU operators. |
|
HPU |
Allows user to add existing TPC operators to Media Pipeline. |
|
HPU |
This operator flips selected dimensions (horizontal, vertical, depthwise). |
|
HPU |
Gathers slices of input tensor into an output tensor of a specific rank. |
|
HPU |
Applies 1D Gaussian Blur on input tensor by applying grouped 1D convolution with input Gaussian kernel. |
|
HPU |
Modifies the hue of an image by specified degrees. |
|
HPU |
Decodes and resizes batch of images. |
|
CPU |
Generates constant data as per given shape and data type on every iteration. |
|
CPU |
Integrates external data reader with MediaPipe. |
|
CPU |
Allows user to generate new tensor data on every iteration. |
|
HPU |
Copies data from the input tensor to the output tensor. |
|
HPU, CPU |
Performs element-wise multiplication and supports |
|
HPU |
Resultant output tensor has the sign flipped for each element in |
|
HPU |
Normalizes the input tensor by subtracting mean and dividing by standard deviation. |
|
HPU |
Pads a tensor with zeros or any numbers. . |
|
CPU |
Performs random crop operation based on label data. |
|
HPU, CPU |
Flips images in a selected direction (horizontal or vertical) with a random predicate. |
|
HPU, CPU |
Generates random numbers from a normal distribution. |
|
HPU, CPU |
Generates random numbers from an uniform distribution. |
|
CPU |
Reader designed for image classification tasks. |
|
CPU |
Reads numpy data files and numpy label files either from given directory or file list and returns the batch of numpy images and labels. |
|
CPU |
Reader designed for video classification tasks. |
|
HPU |
Computes the maximum of the input tensor’s elements along the provided dimension. |
|
HPU |
Computes the minimum of the input tensor’s elements along the provided dimension. |
|
HPU |
Returns a new tensor that has the same values as input tensor in the same order, except with a new shape given by |
|
HPU |
Resizes an image by a scale factor using either bilinear, cubic, nearest neighbor or gaussian sampling. |
|
HPU |
Modifies the saturation level of an image by a specified change factor. |
|
HPU |
Produces a slice of the input tensor along multiple axes. |
|
HPU |
Splits a tensor into a list of tensors along the specified |
|
CPU |
Takes bbox tensor and perform horizontal flip. |
|
CPU |
Takes the metadata output of Coco reader and generate crop window. |
|
CPU |
Takes boxes, labels and encode using predefined 8732 SSD anchor boxes. |
|
CPU |
Takes the metadata output of Coco reader and performs operations such as crop, flip and encoding of bounding boxes. |
|
HPU |
Performs element-wise subtraction and supports |
|
HPU |
Produces a transposed tensor of the input tensor along multiple axes. |
|
HPU |
Decodes, resizes and crops batch of videos. |
|
HPU, CPU |
Outputs a tensor of elements selected from either input0 or input1. |
|
CPU |
Zoom operator supports zoom-in operation on the given images and labels. |