Struct WordErrorRate 

pub struct WordErrorRate<B: Backend> { /* private fields */ }

The word error rate (WER) metric, similar to the CER, is defined as the edit distance (e.g. Levenshtein distance) between the predicted and reference word sequences, divided by the total number of words in the reference. Here, the “units” within the sequences are words.

Implementations

impl<B: Backend> WordErrorRate<B>

pub fn new() -> Self

Creates the metric.

pub fn with_pad_token(self, index: usize) -> Self

Sets the pad token.

Trait Implementations

impl<B: Clone + Backend> Clone for WordErrorRate<B>

fn clone(&self) -> WordErrorRate<B>

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl<B: Backend> Default for WordErrorRate<B>

fn default() -> Self

Returns the “default value” for a type.

impl<B: Backend> Metric for WordErrorRate<B>

type Input = WerInput<B>

The input type of the metric.

fn update( &mut self, input: &WerInput<B>, _metadata: &MetricMetadata, ) -> SerializedEntry

Update the metric state and returns the current metric entry.

fn name(&self) -> MetricName

The parameterized name of the metric.

fn clear(&mut self)

Clear the metric state.

fn attributes(&self) -> MetricAttributes

Attributes of the metric.

fn description(&self) -> Option<String>

A short description of the metric.

impl<B: Backend> Numeric for WordErrorRate<B>

fn value(&self) -> NumericEntry

Returns the numeric value of the metric.

fn running_value(&self) -> NumericEntry

Returns the current aggregated value of the metric over the global step (epoch).