Accuracy

Accuracy metric.

The accuracy metric calculates how often a set of predictions matches a corresponding set of targets. The metric creates two local variables, total and count that are used to compute the frequency with which the predictions match the targets. This frequency is ultimately returned as an idempotent operation that simply divides total by count.

For estimation of the metric over a stream of data, the function creates an update operation that updates these variables and returns the accuracy. Internally, an isCorrect operation computes a tensor with elements equal to 1 where the corresponding elements of the predictions and the targets match, and 0 otherwise. update increments total with the reduced sum of the product of isCorrect and weights, and increments count with the reduced sum of weights.

If weights is None, the weights default to 1. Use weights of 0 to mask values.

Linear Supertypes

Metric[(Output, Output), Output], AnyRef, Any

Instance Constructors

new Accuracy(nameScope: String, defaultWeights: Option[tensors.Tensor[types.FLOAT32]] = None, variablesCollections: Set[Key[variables.Variable]] = Set(METRIC_VARIABLES), valuesCollections: Set[Key[Output]] = Set(METRIC_VALUES), updatesCollections: Set[Key[Output]] = Set(METRIC_UPDATES), resetsCollections: Set[Key[Op]] = Set(METRIC_RESETS))

nameScope
Name prefix for the created ops.
defaultWeights
Default weights with which all computed metric values are multiplied.
variablesCollections
Graph collections in which to add the metric variables (for streaming metrics).
valuesCollections
Graph collections in which to add the metric values.
updatesCollections
Graph collections in which to add the metric updates.
resetsCollections
Graph collections in which to add the metric resets.

Value Members

final def !=(arg0: Any): Boolean

Definition Classes
AnyRef → Any
final def ##(): Int

Definition Classes
AnyRef → Any
final def ==(arg0: Any): Boolean

Definition Classes
AnyRef → Any
final def asInstanceOf[T0]: T0

Definition Classes
Any
def clone(): AnyRef

Attributes
protected[java.lang]
Definition Classes
AnyRef
Annotations
@throws( ... )
def compute(values: (Output, Output), weights: Option[Output] = None, name: String = s"$name/Compute"): Output

Computes the value of this metric for the provided predictions and targets, optionally weighted by weights.
Computes the value of this metric for the provided predictions and targets, optionally weighted by weights.
values
Tuple containing the predictions tensor and the targets tensor.
weights
Optional tensor containing weights for the values.
name
Name prefix for the created ops.
returns
Created output containing the metric value.

Definition Classes
Accuracy → Metric
val defaultWeights: Option[tensors.Tensor[types.FLOAT32]]

Default weights with which all computed metric values are multiplied.
Default weights with which all computed metric values are multiplied.

Attributes
protected
final def eq(arg0: AnyRef): Boolean

Definition Classes
AnyRef
def equals(arg0: Any): Boolean

Definition Classes
AnyRef → Any
def finalize(): Unit

Attributes
protected[java.lang]
Definition Classes
AnyRef
Annotations
@throws( classOf[java.lang.Throwable] )
final def getClass(): Class[_]

Definition Classes
AnyRef → Any
def getWeights(providedWeights: Option[Output]): Option[Output]

Attributes
protected
Definition Classes
Metric
def hashCode(): Int

Definition Classes
AnyRef → Any
final def isInstanceOf[T0]: Boolean

Definition Classes
Any
def name: String

Name of this metric.
Name of this metric.

Definition Classes
Accuracy → Metric
val nameScope: String

Name prefix for the created ops.
final def ne(arg0: AnyRef): Boolean

Definition Classes
AnyRef
final def notify(): Unit

Definition Classes
AnyRef
final def notifyAll(): Unit

Definition Classes
AnyRef
val resetsCollections: Set[Key[Op]]

Graph collections in which to add the metric resets.
def streaming(values: (Output, Output), weights: Option[Output] = None, name: String = s"$name/Streaming"): StreamingInstance[Output]

Creates ops for computing the value of this metric in a streaming fashion.
Creates ops for computing the value of this metric in a streaming fashion. This function returns an op for obtaining the value of this metric, as well as a pair of ops to update its accumulated value and reset it.
values
Tuple containing the predictions tensor and the targets tensor.
weights
Optional tensor containing weights for the predictions.
name
Name prefix for the created ops.
returns
Tuple containing: (i) an output representing the current value of the metric, (ii) an op used to update its current value and obtain the new value, and (iii) an op used to reset its value.

Definition Classes
Accuracy → Metric
final def synchronized[T0](arg0: ⇒ T0): T0

Definition Classes
AnyRef
def toString(): String

Definition Classes
Metric → AnyRef → Any
val updatesCollections: Set[Key[Output]]

Graph collections in which to add the metric updates.
val valuesCollections: Set[Key[Output]]

Graph collections in which to add the metric values.
val variablesCollections: Set[Key[variables.Variable]]

Graph collections in which to add the metric variables (for streaming metrics).
final def wait(): Unit

Definition Classes
AnyRef
Annotations
@throws( ... )
final def wait(arg0: Long, arg1: Int): Unit

Definition Classes
AnyRef
Annotations
@throws( ... )
final def wait(arg0: Long): Unit

Definition Classes
AnyRef
Annotations
@throws( ... )
def weights: Option[tensors.Tensor[types.FLOAT32]]

Weights to multiply the provided values with when computing the value of this metric.
Weights to multiply the provided values with when computing the value of this metric.

Definition Classes
Accuracy → Metric

Related Docs: object Accuracy | package metrics

class Accuracy extends Metric[(Output, Output), Output]

Instance Constructors

Value Members

final def !=(arg0: Any): Boolean

final def ##(): Int

final def ==(arg0: Any): Boolean

final def asInstanceOf[T0]: T0

def clone(): AnyRef

def compute(values: (Output, Output), weights: Option[Output] = None, name: String = s"$name/Compute"): Output

val defaultWeights: Option[tensors.Tensor[types.FLOAT32]]

final def eq(arg0: AnyRef): Boolean

def equals(arg0: Any): Boolean

def finalize(): Unit

final def getClass(): Class[_]

def getWeights(providedWeights: Option[Output]): Option[Output]

def hashCode(): Int

final def isInstanceOf[T0]: Boolean

def name: String

val nameScope: String

final def ne(arg0: AnyRef): Boolean

final def notify(): Unit

final def notifyAll(): Unit

val resetsCollections: Set[Key[Op]]

def streaming(values: (Output, Output), weights: Option[Output] = None, name: String = s"$name/Streaming"): StreamingInstance[Output]

final def synchronized[T0](arg0: ⇒ T0): T0

def toString(): String

val updatesCollections: Set[Key[Output]]

val valuesCollections: Set[Key[Output]]

val variablesCollections: Set[Key[variables.Variable]]

final def wait(): Unit

final def wait(arg0: Long, arg1: Int): Unit

final def wait(arg0: Long): Unit

def weights: Option[tensors.Tensor[types.FLOAT32]]

Inherited from Metric[(Output, Output), Output]

Inherited from AnyRef

Inherited from Any

Ungrouped