Weighted MAPE¶

Module Interface¶

class torchmetrics.WeightedMeanAbsolutePercentageError(**kwargs)[source]

Computes weighted mean absolute percentage error (WMAPE). The output of WMAPE metric is a non-negative floating point, where the optimal value is 0. It is computes as:

$\text{WMAPE} = \frac{\sum_{t=1}^n | y_t - \hat{y}_t | }{\sum_{t=1}^n |y_t| }$

Where $y$ is a tensor of target values, and $\hat{y}$ is a tensor of predictions.

As input to forward and update the metric accepts the following input:

preds (Tensor): Predictions from model
target (Tensor): Ground truth float tensor with shape (N,d)

As output of forward and compute the metric returns the following output:

wmape (Tensor): A tensor with non-negative floating point wmape value between 0 and 1

Parameters: kwargs¶ (Any) – Additional keyword arguments, see Advanced metric settings for more info.

Example

>>> import torch
>>> _ = torch.manual_seed(42)
>>> preds = torch.randn(20,)
>>> target = torch.randn(20,)
>>> wmape = WeightedMeanAbsolutePercentageError()
>>> wmape(preds, target)
tensor(1.3967)

Initializes internal Module state, shared by both nn.Module and ScriptModule.

Functional Interface¶

torchmetrics.functional.weighted_mean_absolute_percentage_error(preds, target)[source]

Computes weighted mean absolute percentage error (WMAPE).

The output of WMAPE metric is a non-negative floating point, where the optimal value is 0. It is computes as:

$\text{WMAPE} = \frac{\sum_{t=1}^n | y_t - \hat{y}_t | }{\sum_{t=1}^n |y_t| }$

Where $y$ is a tensor of target values, and $\hat{y}$ is a tensor of predictions.

Parameters

preds¶ (Tensor) – estimated labels
target¶ (Tensor) – ground truth labels

Return type

Tensor

Returns

Tensor with WMAPE.

Example

>>> import torch
>>> _ = torch.manual_seed(42)
>>> preds = torch.randn(20,)
>>> target = torch.randn(20,)
>>> weighted_mean_absolute_percentage_error(preds, target)
tensor(1.3967)