mindspore.ops.hinge_embedding_loss

mindspore.ops.hinge_embedding_loss(inputs, targets, margin=1.0, reduction='mean')[source]

Measures Hinge Embedding Loss given an input Tensor intputs and a labels Tensor targets (containing 1 or -1).

The loss function for \(n\)-th sample in the mini-batch is

\[\begin{split}l_n = \begin{cases} x_n, & \text{if}\; y_n = 1,\\ \max \{0, \Delta - x_n\}, & \text{if}\; y_n = -1, \end{cases}\end{split}\]

and the total loss functions is

\[\begin{split}\ell(x, y) = \begin{cases} \operatorname{mean}(L), & \text{if reduction} = \text{'mean';}\\ \operatorname{sum}(L), & \text{if reduction} = \text{'sum'.} \end{cases}\end{split}\]

where \(L = \{l_1,\dots,l_N\}^\top\).

Parameters

  • inputs (Tensor) – Predicted values, represented as \(x\) in the formula.

  • targets (Tensor) – Label values, represented as \(y\) in the formula. Has the same shape as inputs, contains -1 or 1.

  • margin (float, int) – Threshold defined by Hinge Embedding Loss \(margin\). Represented as \(\Delta\) in the formula. Default: 1.0.

  • reduction (str) – Specify the computing method to be applied to the outputs: ‘none’, ‘mean’, or ‘sum’. Default: ‘mean’.

Returns

Tensor or Tensor scalar, the computed loss depending on \(reduction\).

Raises

  • TypeError – If inputs is not a Tensor.

  • TypeError – If targets is not a Tensor.

  • TypeError – If margin is not a float or int.

  • ValueError – If targets does not have the same shape as inputs or they could not broadcast to each other.

  • ValueError – If reduction is not one of ‘none’, ‘mean’, ‘sum’.

Supported Platforms:

Ascend GPU CPU

Examples

>>> import numpy as np
>>> import mindspore.common.dtype as mstype
>>> import mindspore.ops as ops
>>> from mindspore import Tensor
>>> arr1 = np.array([0.9, -1.2, 2, 0.8, 3.9, 2, 1, 0, -1]).reshape((3, 3))
>>> arr2 = np.array([1, 1, -1, 1, -1, 1, -1, 1, 1]).reshape((3, 3))
>>> logits = Tensor(arr1, mstype.float32)
>>> labels = Tensor(arr2, mstype.float32)
>>> loss = ops.hinge_embedding_loss(logits, labels, margin=1.0, reduction='mean')
>>> print(loss)
0.16666666