Transpose

Function Usage

Multi-dimensional transposition API of the tensor class. It performs transposition based on the given axes dimension arrangement. Asynchronous calling is supported.

It is supported by the Atlas inference product and Atlas 200I/500 A2 inference product.

For the Atlas 200I/500 A2 inference product, preloading is supported (see Example of the Preloading File of the Initialization Operator).

The following conditions must be met:

The input and output tensors must be on the device or DVPP side, and the parameters (stream and data memory) must be on the same device.
For synchronization, the device where the data memory is located must be the same as the initialized device.
The input and output parameters cannot exceed four dimensions, and must match the tensor types and shapes.

Prototype

APP_ERROR Transpose(const Tensor &src, Tensor &dst, std::vector<int> axes, AscendStream &stream = AscendStream::DefaultStream());

Parameters

Parameter	Input/Output	Description
src	Input	Tensor class, supporting float16, float32, and uint8 data types.
dst	Output	Tensor class, supporting float16, float32, and uint8 data types. An empty tensor can be passed. If dst is not an empty tensor, call Tensor.Malloc() to allocate memory in advance. The shape must be the shape of the src after the axis is exchanged according to the axes setting. For example, if the src shape is {1, 480, 640, 3} and axes value is {0, 2, 1, 3}, the shape of dst must be {1, 640, 480, 3}.
axes	Input	std::vector<int> class. It specifies how the dimensions to be transposed are arranged. The length of axes must be the same as the dimension of the input tensor. For example, if src has four dimensions, the length of axes must be 4. Elements in axes must contain numbers in the range of [0, size - 1], and each number can appear only once. If an empty vector is passed, the vector is transposed in reverse order.
stream	Input	AscendStream type. The default value is AscendStream::DefaultStream(). When the parameter value is the default value, the API calling is a synchronous operation. In other cases, the API calling is an asynchronous operation.

Response Parameters

Data Structure	Description
APP_ERROR	For details about the returned error codes, see APP_ERROR Description.

Parent topic: TensorOperations