MCPcopy Create free account
hub / github.com/MegEngine/MegEngine / logaddexp

Function logaddexp

imperative/python/megengine/functional/elemwise.py:827–849  ·  view source on GitHub ↗

r"""Element-wise :math:`\log(e^x + e^y)` function. This function is useful in statistics where the calculated probabilities of events may be so small as to exceed the range of normal floating point numbers. In such cases the logarithm of the calculated probability is stored. This f

(x: Tensor, y: Tensor)

Source from the content-addressed store, hash-verified

825
826
827def logaddexp(x: Tensor, y: Tensor) -> Tensor:
828 r"""Element-wise :math:`\log(e^x + e^y)` function.
829
830 This function is useful in statistics where the calculated probabilities of events may be so small
831 as to exceed the range of normal floating point numbers.
832 In such cases the logarithm of the calculated probability is stored.
833 This function allows adding probabilities stored in such a fashion.
834
835 Args:
836 x: input tensor. Should have a floating-point data type.
837 y: input tensor. Must be compatible with :math:`x`` (see :ref:`broadcasting-rule` ). Should have a floating-point data type.
838
839 Returns:
840 a tensor containing the evaluated result for each element in :math:`x` and :math:`y`.
841 The returned tensor must have a floating-point data type determined by :ref:`dtype-promotion`.
842
843 Examples:
844 >>> prob1 = F.log(1e-10)
845 >>> prob2 = F.log(2e-10)
846 >>> F.logaddexp(prob1, prob2)
847 Tensor(-21.927238, device=xpux:0)
848 """
849 return _elwise(x, y, mode=Elemwise.Mode.LOG_SUM_EXP)
850
851
852def round(x):

Callers 1

ctc_lossFunction · 0.85

Calls 1

_elwiseFunction · 0.50

Tested by

no test coverage detected