mindspore.ops.smooth_l1_loss

mindspore.ops.smooth_l1_loss(input, target, beta=1.0, reduction='none')

Computes smooth L1 loss, a robust L1 loss.

SmoothL1Loss is a Loss similar to MSELoss but less sensitive to outliers as described in the Fast R-CNN by Ross Girshick.

Given two input \(x,\ y\) of length \(N\), the unreduced SmoothL1Loss can be described as follows:

\[\begin{split}L_{i} = \begin{cases} \frac{0.5 (x_i - y_i)^{2}}{\text{beta}}, & \text{if } |x_i - y_i| < \text{beta} \\ |x_i - y_i| - 0.5 * \text{beta}, & \text{otherwise. } \end{cases}\end{split}\]

If reduction is not none, then:

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

Here \(\text{beta}\) controls the point where the loss function changes from quadratic to linear. \(\text{beta}>0\) , its default value is 1.0 . \(N\) is the batch size.

Warning

This API has poor performance on CPU and it is recommended to run it on the Ascend/GPU.

Parameters

• input (Tensor) –

  Tensor of shape \((N, *)\) where \(*\) means, any number of additional dimensions.Supported dtypes:

  • Ascend: float16, float32, bfloat16.

  • CPU/GPU: float16, float32, float64.

• target (Tensor) –

  Ground truth data, tensor of shape \((N, *)\).

  • CPU/Ascend: has the same shape as the input, target and input comply with the implicit type conversion rules to make the data types consistent.

  • GPU: has the same shape and dtype as the input.

• beta (number, optional) –

  A parameter used to control the point where the function will change between L1 to L2 loss. Default: 1.0 .

  • Ascend: The value should be equal to or greater than zero.

  • CPU/GPU: The value should be greater than zero.

• reduction (str, optional) –

  Apply specific reduction method to the output: 'none' , 'mean' , 'sum' . Default: 'none' .

  • 'none': no reduction will be applied.

  • 'mean': compute and return the mean of elements in the output.

  • 'sum': the output elements will be summed.

Returns

Tensor, if reduction is 'none', then output is a tensor with the same shape as input. Otherwise, the shape of output tensor is \(()\).

Raises

• TypeError – If input input, target is not Tensor.

• RuntimeError – If dtype of input or target is not one of float16, float32, float64, bfloat16.

• ValueError – If shape of input is not the same as target.

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

• TypeError – If beta is not a float, int or bool.

• RuntimeError – If beta is less than or equal to 0.

Supported Platforms:

Ascend GPU CPU

Examples

>>> import mindspore
>>> import numpy as np
>>> from mindspore import Tensor, ops
>>> logits = Tensor(np.array([1, 2, 3]), mindspore.float32)
>>> labels = Tensor(np.array([1, 2, 2]), mindspore.float32)
>>> output = ops.smooth_l1_loss(logits, labels)
>>> print(output)
[0.  0.  0.5]