Shortcuts

torcheval.metrics.MulticlassF1Score

class torcheval.metrics.MulticlassF1Score(*, num_classes: Optional[int] = None, average: Optional[str] = 'micro', device: Optional[device] = None)[source]

Compute f1 score, which is defined as the harmonic mean of precision and recall. We convert NaN to zero when f1 score is NaN. This happens when either precision or recall is NaN or when both precision and recall are zero. Its functional version is torcheval.metrics.functional.multi_class_f1_score(). See also BinaryF1Score

Parameters:
  • num_classes (int) – Number of classes.
  • average (str, Optional) –
    • 'micro' [default]: Calculate the metrics globally.
    • 'macro': Calculate metrics for each class separately, and return their unweighted mean. Classes with 0 true and predicted instances are ignored.
    • 'weighted'” Calculate metrics for each class separately, and return their weighted sum. Weights are defined as the proportion of occurrences of each class in “target”. Classes with 0 true and predicted instances are ignored.
    • None: Calculate the metric for each class separately, and return the metric for every class.

Examples:

>>> import torch
>>> from torcheval.metrics import MulticlassF1Score
>>> metric = MulticlassF1Score(num_classes=4)
>>> input = torch.tensor([0, 2, 1, 3])
>>> target = torch.tensor([0, 1, 2, 3])
>>> metric.update(input, target)
>>> metric.compute()
tensor(0.5000)

>>> metric = MulticlassF1Score(average=None, num_classes=4)
>>> input = torch.tensor([0, 2, 1, 3])
>>> target = torch.tensor([0, 1, 2, 3])
>>> metric.update(input, target)
>>> metric.compute()
tensor([1., 0., 0., 1.])

>>> metric = MulticlassF1Score(average="macro", num_classes=2)
>>> input = torch.tensor([0, 0, 1, 1, 1])
>>> target = torch.tensor([0, 0, 0, 0, 1])
>>> metric.update(input, target)
>>> metric.compute()
tensor(0.5833)

>>> metric = MulticlassF1Score(num_classes=4)
>>> input = torch.tensor([[0.9, 0.1, 0, 0], [0.1, 0.2, 0.4, 0.3], [0, 1.0, 0, 0], [0, 0, 0.2, 0.8]])
>>> target = torch.tensor([0, 1, 2, 3])
>>> metric.update(input, target)
>>> metric.compute()
tensor(0.5)
__init__(*, num_classes: Optional[int] = None, average: Optional[str] = 'micro', device: Optional[device] = None) None[source]

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, or a dictionary with torch.Tensor as values

Methods

__init__(*[, num_classes, average, device]) Initialize a metric object and its internal states.
compute() Return the f1 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().

Docs

Access comprehensive developer documentation for PyTorch

View Docs

Tutorials

Get in-depth tutorials for beginners and advanced developers

View Tutorials

Resources

Find development resources and get your questions answered

View Resources