Struct ConfusionStatsState 

pub struct ConfusionStatsState { /* private fields */ }

Accumulates confusion-matrix counts (TP, FP, FN) across an epoch.

Used by metrics derived from confusion-matrix ratios (precision, recall, etc.) which must accumulate raw counts and compute the ratio over the summed counts. Averaging the per-batch ratios (using NumericMetricState) is not equivalent; it is biased whenever per-batch class support varies. Accumulates confusion-matrix counts (TP, FP, FN) across an epoch using native tensors.

Implementations

impl ConfusionStatsState

pub fn new() -> ConfusionStatsState

Create a new ConfusionStatsState.

pub fn reset(&mut self)

Reset the state.

pub fn update( &mut self, tp: Option<Tensor<1>>, fp: Option<Tensor<1>>, fn_: Option<Tensor<1>>, sample_size: usize, )

Accumulate running totals and capture the batch snapshot.

pub fn compute_update( &mut self, class_reduction: ClassReduction, format: FormatOptions, compute_fn: impl Fn(Option<Tensor<1>>, Option<Tensor<1>>, Option<Tensor<1>>) -> Tensor<1>, ) -> Result<SerializedEntry, TensorReadError>

Compute the batch-level value and the running epoch-level value (from all accumulated counts)

Errors

Returns a TensorReadError if the computed metric tensors cannot be read back.

pub fn compute_final(&mut self, format: FormatOptions) -> SerializedEntry

Compute the final metric for the accumulated global state.

pub fn current_value(&self) -> Option<NumericEntry>

Get the current batch value.

pub fn running_value(&self) -> Option<NumericEntry>

Get the running aggregated value.

pub fn final_value(&self) -> NumericEntry

Get the final value of the metric.

Trait Implementations

impl Clone for ConfusionStatsState

fn clone(&self) -> ConfusionStatsState

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Default for ConfusionStatsState

fn default() -> ConfusionStatsState

Returns the “default value” for a type.