giagrad.Tensor.softmax#

Tensor.softmax(axis) → Tensor[source]#

Applies Softmax function to every 1-D slice defined by axis. See Softmax.

The elements of the n-dimensinal output Tensor will lie in the range \([0, 1]\) and sum to \(1\) for the specified 1-D slices defined by axis.

Softmax for a one-dimensional slice is defined as:

\[\text{Softmax}(x_i) = \frac{\exp(x_i)}{\sum_j \exp(x_j)}\]

Parameters:: axis¶ (int) – The dimension along which Softmax will be computed (so every slice along axis will sum to 1).

Examples

>>> t = Tensor.empty(2, 3).uniform(-1, 1)
>>> t
tensor: [[ 0.27639335  0.7524293   0.69203097]
         [ 0.37772807 -0.9291505  -0.80418533]]
>>> t.softmax(axis=1)
tensor: [[0.24242324 0.390224   0.36735278]
         [0.6339727  0.17159334 0.19443396]] fn: Softmax(axis=1)