cupy.matmul

cupy.matmul()

Returns the matrix product of two arrays and is the implementation of the @ operator introduced in Python 3.5 following PEP465.

The main difference against cupy.dot are the handling of arrays with more than 2 dimensions. For more information see numpy.matmul().

Note

Differences to numpy or missing features:

Currently the output must be real (float16, float32, uint8, ...), complex64 and complex128 follow later. This means, that numpy.result_type(a.dtype, b.dtype) have to be real.

The out array as input is currently not supported.

Parameters:
Returns:

Output array.
Return type:

cupy.ndarray

See also

numpy.matmul()