torch.frexp#

torch.frexp(input,*,out=None)->(Tensormantissa,Tensorexponent)#

Decomposesinput into mantissa and exponent tensorssuch that$\text{input}=\text{mantissa}\times 2^{\text{exponent}}$input=mantissa×2exponent.

The range of mantissa is the open interval (-1, 1).

Supports float inputs.

Parameters

input (Tensor) – the input tensor

Keyword Arguments

out (tuple,optional) – the output tensors

Example:

>>>x=torch.arange(9.)>>>mantissa,exponent=torch.frexp(x)>>>mantissatensor([0.0000, 0.5000, 0.5000, 0.7500, 0.5000, 0.6250, 0.7500, 0.8750, 0.5000])>>>exponenttensor([0, 1, 2, 2, 3, 3, 3, 3, 4], dtype=torch.int32)>>>torch.ldexp(mantissa,exponent)tensor([0., 1., 2., 3., 4., 5., 6., 7., 8.])