SmoothL1Loss(size_average=None, reduce=None, reduction='mean', beta=1.0)¶
Creates a criterion that uses a squared term if the absolute element-wise error falls below beta and an L1 term otherwise. It is less sensitive to outliers than the
torch.nn.MSELossand in some cases prevents exploding gradients (e.g. see Fast R-CNN paper by Ross Girshick). Omitting a scaling factor of
beta, this loss is also known as the Huber loss:
where is given by:
and arbitrary shapes with a total of elements each the sum operation still operates over all the elements, and divides by .
betais an optional parameter that defaults to 1.
betais set to 0, this is equivalent to
L1Loss. Passing a negative value in for
betawill result in an exception.
The division by can be avoided if sets
reduction = 'sum'.
reduction). By default, the losses are averaged over each loss element in the batch. Note that for some losses, there are multiple elements per sample. If the field
size_averageis set to
False, the losses are instead summed for each minibatch. Ignored when
reduction). By default, the losses are averaged or summed over observations for each minibatch depending on
False, returns a loss per batch element instead and ignores
'none': no reduction will be applied,
'mean': the sum of the output will be divided by the number of elements in the output,
'sum': the output will be summed. Note:
reduceare in the process of being deprecated, and in the meantime, specifying either of those two args will override
beta (float, optional) – Specifies the threshold at which to change between L1 and L2 loss. This value defaults to 1.0.
Input: where means, any number of additional dimensions
Target: , same shape as the input
Output: scalar. If
'none', then , same shape as the input