mindspore.nn.MultiLabelSoftMarginLoss

class mindspore.nn.MultiLabelSoftMarginLoss(weight=None, reduction='mean')[source]

Calculates the MultiLabelSoftMarginLoss. The multi-label soft margin loss is a commonly used loss function in multi-label classification tasks where an input sample can belong to multiple classes. Given an input $x$ and binary labels $y$ of size $(N, C)$ , where $N$ denotes the number of samples and $C$ denotes the number of classes.

loss (x, y) = - \frac{1}{N} \frac{1}{C} \sum_{i = 1}^{N} \sum_{j = 1}^{C} (y_{i j} \log \frac{1}{1 + e^{- x_{i j}}} + (1 - y_{i j}) \log \frac{e^{- x_{i j}}}{1 + e^{- x_{i j}}})

where $x i j$ represents the predicted score of sample $i$ for class $j$ . $y i j$ represents the binary label of sample $i$ for class $j$ , where sample $i$ belongs to class $j$ if $y i j = 1$ , and sample $i$ does not belong to class $j$ if $y i j = 0$ . For a multi-label classification task, each sample may have multiple labels with a value of 1 in the binary label $y$ . weight will multiply to the loss of each class if given.

Parameters

weight (Union[Tensor, int, float]) – The manual rescaling weight given to each class. Default: None.
reduction (str) – Specifies which reduction to be applied to the output. It must be one of ‘none’, ‘mean’, and ‘sum’, meaning no reduction, reduce mean and sum on output, respectively. Default: ‘mean’.

Inputs:

x (Tensor) - A tensor of shape (N, C), where N is batch size and C is number of classes.
target (Tensor) - The label target Tensor which has the same shape as x.

Outputs:

Tensor, the data type is the same as x, if the reduction is ‘none’, its shape is (N), otherwise it is zero.

Raises: ValueError – If the rank of x or target is not 2.

Supported Platforms:: Ascend GPU CPU

Examples

>>> x = Tensor([[0.3, 0.6, 0.6], [0.9, 0.4, 0.2]])
>>> target = Tensor([[0.0, 0.0, 1.0], [0.0, 0.0, 1.0]])
>>> loss = nn.MultiLabelSoftMarginLoss(reduction='mean')
>>> out = loss(x, target)
>>> print(out.asnumpy())
0.84693956