torcheval.metrics.MultilabelAccuracy

class torcheval.metrics.MultilabelAccuracy(*, threshold: float = 0.5, criteria: str = 'exact_match', device: device | None = None)

Compute multilabel accuracy score, which is the frequency of input matching target. Its functional version is torcheval.metrics.functional.multilabel_accuracy().

Parameters:

• threshold (float, Optional) – Threshold for converting input into predicted labels for each sample. torch.where(input < threshold, 0, 1) will be applied to the input.

• criteria (str, Optional) –

  • 'exact_match' [default]: The set of labels predicted for a sample must exactly match the corresponding set of labels in target. Also known as subset accuracy.

  • 'hamming': Fraction of correct labels over total number of labels.

  • 'overlap': The set of labels predicted for a sample must overlap with the corresponding set of labels in target.

  • 'contain': The set of labels predicted for a sample must contain the corresponding set of labels in target.

  • 'belong': The set of labels predicted for a sample must (fully) belong to the corresponding set of labels in target.

Examples:

>>> import torch
>>> from torcheval.metrics import MultilabelAccuracy
>>> metric = MultilabelAccuracy()
>>> input = torch.tensor([[0, 1], [1, 1], [0, 0], [0, 1]])
>>> target = torch.tensor([[0, 1], [1, 0], [0, 0], [1, 1]])
>>> metric.update(input, target)
>>> metric.compute()
tensor(0.5)  # 2 / 4

>>> metric = MultilabelAccuracy(criteria="hamming")
>>> input = torch.tensor([[0, 1], [1, 1], [0, 0], [0, 1]])
>>> target = torch.tensor([[0, 1], [1, 0], [0, 0], [1, 1]])
>>> metric.update(input, target)
>>> metric.compute()
tensor(0.75)  # 6 / 8

>>> metric = MultilabelAccuracy(criteria="overlap")
>>> input = torch.tensor([[0, 1], [1, 1], [0, 0], [0, 1]])
>>> target = torch.tensor([[0, 1], [1, 0], [0, 0], [1, 1]])
>>> metric.update(input, target)
>>> metric.compute()
tensor(1)  # 4 / 4

>>> metric = MultilabelAccuracy(criteria="contain")
>>> input = torch.tensor([[0, 1], [1, 1], [0, 0], [0, 1]])
>>> target = torch.tensor([[0, 1], [1, 0], [0, 0], [1, 1]])
>>> metric.update(input, target)
>>> metric.compute()
tensor(0.75)  # 3 / 4, input[0],input[1],input[2]

>>> metric = MultilabelAccuracy(criteria="belong")
>>> input = torch.tensor([[0, 1], [1, 1], [0, 0], [0, 1]])
>>> target = torch.tensor([[0, 1], [1, 0], [0, 0], [1, 1]])
>>> metric.update(input, target)
>>> metric.compute()
tensor(0.75)  # 3 / 4, input[0],input[1],input[3]

__init__(*, threshold: float = 0.5, criteria: str = 'exact_match', device: device | None = None) → None

Initialize a metric object and its internal states.

Use self._add_state() to initialize state variables of your metric class. The state variables should be either torch.Tensor, a list of torch.Tensor, a dictionary with torch.Tensor as values, or a deque of torch.Tensor.

Methods

__init__(*[, threshold, criteria, device])	Initialize a metric object and its internal states.
compute()	Return the accuracy score.
load_state_dict(state_dict[, strict])	Loads metric state variables from state_dict.
merge_state(metrics)	Implement this method to update the current metric's state variables to be the merged states of the current metric and input metrics.
reset()	Reset the metric state variables to their default value.
state_dict()	Save metric state variables in state_dict.
to(device, *args, **kwargs)	Move tensors in metric state variables to device.
update(input, target)	Update states with the ground truth labels and predictions.

Attributes

device

The last input device of Metric.to().