torch.logcumsumexp — PyTorch 2.7 documentation (original) (raw)

torch.logcumsumexp(input, dim, *, out=None) → Tensor¶

Returns the logarithm of the cumulative summation of the exponentiation of elements of input in the dimension dim.

For summation index jj given by dim and other indices ii, the result is

logcumsumexp(x)ij=log⁡∑k=0jexp⁡(xik)\text{logcumsumexp}(x)_{ij} = \log \sum\limits_{k=0}^{j} \exp(x_{ik})

Parameters

• input (Tensor) – the input tensor.
• dim (int) – the dimension to do the operation over

Keyword Arguments

out (Tensor, optional) – the output tensor.

Example:

a = torch.randn(10) torch.logcumsumexp(a, dim=0) tensor([-0.42296738, -0.04462666, 0.86278635, 0.94622083, 1.05277811, 1.39202815, 1.83525007, 1.84492621, 2.06084887, 2.06844475]))