Loss (Deep Java Library 0.3.0 API specification)

java.lang.Object
- ai.djl.training.evaluator.Evaluator
- - ai.djl.training.loss.Loss

Direct Known Subclasses:

AbstractCompositeLoss, HingeLoss, L1Loss, L2Loss, SigmoidBinaryCrossEntropyLoss, SoftmaxCrossEntropyLoss
```
public abstract class Loss
extends Evaluator
```
Loss functions (or Cost functions) are used to evaluate the model predictions against true labels for optimization.
Although all evaluators can be used to measure the performance of a model, not all of them are suited to being used by an optimizer. Loss functions are usually non-negative where a larger loss represents worse performance. They are also real-valued to accurately compare models.
When creating a loss function, you should avoid having the loss depend on the batch size. For example, if you have a loss per item in a batch and sum those losses, your loss would be numItemsInBatch*avgLoss. Instead, you should take the mean of those losses to reduce out the batchSize factor. Otherwise, it can make it difficult to tune the learning rate since any change in the batch size would throw it off. If you have a variable batch size, it would be even more difficult.
For more details about the class internals, see Evaluator.

Field Summary
- Fields inherited from class ai.djl.training.evaluator.Evaluator
  totalInstances

Constructor Summary

Constructors
Constructor and Description

Loss(java.lang.String name)
Base class for metric with abstract update methods.

Constructors
Constructor and Description
`Loss(java.lang.String name)` Base class for metric with abstract update methods.

Method Summary

All Methods Static Methods Instance Methods Concrete Methods
Modifier and Type	Method and Description
`void`	`addAccumulator(java.lang.String key)` Adds an accumulator for the results of the evaluation with the given key.
`float`	`getAccumulator(java.lang.String key)` Returns the accumulated evaluator value.
`static HingeLoss`	`hingeLoss()` Returns a new instance of `HingeLoss` with default arguments.
`static HingeLoss`	`hingeLoss(java.lang.String name)` Returns a new instance of `HingeLoss` with default arguments.
`static HingeLoss`	`hingeLoss(java.lang.String name, int margin, float weight)` Returns a new instance of `HingeLoss` with the given arguments.
`static L1Loss`	`l1Loss()` Returns a new instance of `L1Loss` with default weight and batch axis.
`static L1Loss`	`l1Loss(java.lang.String name)` Returns a new instance of `L1Loss` with default weight and batch axis.
`static L1Loss`	`l1Loss(java.lang.String name, float weight)` Returns a new instance of `L1Loss` with given weight and batch axis.
`static L2Loss`	`l2Loss()` Returns a new instance of `L2Loss` with default weight and batch axis.
`static L2Loss`	`l2Loss(java.lang.String name)` Returns a new instance of `L2Loss` with default weight and batch axis.
`static L2Loss`	`l2Loss(java.lang.String name, float weight)` Returns a new instance of `L2Loss` with given weight and batch axis.
`void`	`resetAccumulator(java.lang.String key)` Resets the evaluator value with the given key.
`static SigmoidBinaryCrossEntropyLoss`	`sigmoidBinaryCrossEntropyLoss()` Returns a new instance of `SigmoidBinaryCrossEntropyLoss` with default arguments.
`static SigmoidBinaryCrossEntropyLoss`	`sigmoidBinaryCrossEntropyLoss(java.lang.String name)` Returns a new instance of `SigmoidBinaryCrossEntropyLoss` with default arguments.
`static SigmoidBinaryCrossEntropyLoss`	`sigmoidBinaryCrossEntropyLoss(java.lang.String name, float weight, boolean fromSigmoid)` Returns a new instance of `SigmoidBinaryCrossEntropyLoss` with the given arguments.
`static SoftmaxCrossEntropyLoss`	`softmaxCrossEntropyLoss()` Returns a new instance of `SoftmaxCrossEntropyLoss` with default arguments.
`static SoftmaxCrossEntropyLoss`	`softmaxCrossEntropyLoss(java.lang.String name)` Returns a new instance of `SoftmaxCrossEntropyLoss` with default arguments.
`static SoftmaxCrossEntropyLoss`	`softmaxCrossEntropyLoss(java.lang.String name, float weight, int classAxis, boolean sparseLabel, boolean fromLogit)` Returns a new instance of `SoftmaxCrossEntropyLoss` with the given arguments.
`void`	`updateAccumulator(java.lang.String key, NDList labels, NDList predictions)` Updates the evaluator with the given key based on a `NDList` of labels and predictions.

Methods inherited from class ai.djl.training.evaluator.Evaluator
checkLabelShapes, checkLabelShapes, evaluate, getName

Methods inherited from class java.lang.Object
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

- Constructor Detail
  - Loss
```
public Loss(java.lang.String name)
```
    Base class for metric with abstract update methods.
    
    Parameters:
    
    name - The display name of the Loss
- Method Detail
  - l1Loss
```
public static L1Loss l1Loss()
```
    Returns a new instance of L1Loss with default weight and batch axis.
    
    Returns:
    
    a new instance of L1Loss
  - l1Loss
```
public static L1Loss l1Loss(java.lang.String name)
```
    Returns a new instance of L1Loss with default weight and batch axis.
    
    Parameters:
    
    name - the name of the loss
    
    Returns:
    
    a new instance of L1Loss
  - l1Loss
```
public static L1Loss l1Loss(java.lang.String name,
                            float weight)
```
    Returns a new instance of L1Loss with given weight and batch axis.
    
    Parameters:
    
    name - the name of the loss
    
    weight - the weight to apply on loss value, default 1
    
    Returns:
    
    a new instance of L1Loss
  - l2Loss
```
public static L2Loss l2Loss()
```
    Returns a new instance of L2Loss with default weight and batch axis.
    
    Returns:
    
    a new instance of L2Loss
  - l2Loss
```
public static L2Loss l2Loss(java.lang.String name)
```
    Returns a new instance of L2Loss with default weight and batch axis.
    
    Parameters:
    
    name - the name of the loss
    
    Returns:
    
    a new instance of L2Loss
  - l2Loss
```
public static L2Loss l2Loss(java.lang.String name,
                            float weight)
```
    Returns a new instance of L2Loss with given weight and batch axis.
    
    Parameters:
    
    name - the name of the loss
    
    weight - the weight to apply on loss value, default 1
    
    Returns:
    
    a new instance of L2Loss
  - sigmoidBinaryCrossEntropyLoss
```
public static SigmoidBinaryCrossEntropyLoss sigmoidBinaryCrossEntropyLoss()
```
    Returns a new instance of SigmoidBinaryCrossEntropyLoss with default arguments.
    
    Returns:
    
    a new instance of SigmoidBinaryCrossEntropyLoss
  - sigmoidBinaryCrossEntropyLoss
```
public static SigmoidBinaryCrossEntropyLoss sigmoidBinaryCrossEntropyLoss(java.lang.String name)
```
    Returns a new instance of SigmoidBinaryCrossEntropyLoss with default arguments.
    
    Parameters:
    
    name - the name of the loss
    
    Returns:
    
    a new instance of SigmoidBinaryCrossEntropyLoss
  - sigmoidBinaryCrossEntropyLoss
```
public static SigmoidBinaryCrossEntropyLoss sigmoidBinaryCrossEntropyLoss(java.lang.String name,
                                                                          float weight,
                                                                          boolean fromSigmoid)
```
    Returns a new instance of SigmoidBinaryCrossEntropyLoss with the given arguments.
    
    Parameters:
    
    name - the name of the loss
    
    weight - the weight to apply on the loss value, default 1
    
    fromSigmoid - whether the input is from the output of sigmoid, default false
    
    Returns:
    
    a new instance of SigmoidBinaryCrossEntropyLoss
  - softmaxCrossEntropyLoss
```
public static SoftmaxCrossEntropyLoss softmaxCrossEntropyLoss()
```
    Returns a new instance of SoftmaxCrossEntropyLoss with default arguments.
    
    Returns:
    
    a new instance of SoftmaxCrossEntropyLoss
  - softmaxCrossEntropyLoss
```
public static SoftmaxCrossEntropyLoss softmaxCrossEntropyLoss(java.lang.String name)
```
    Returns a new instance of SoftmaxCrossEntropyLoss with default arguments.
    
    Parameters:
    
    name - the name of the loss
    
    Returns:
    
    a new instance of SoftmaxCrossEntropyLoss
  - softmaxCrossEntropyLoss
```
public static SoftmaxCrossEntropyLoss softmaxCrossEntropyLoss(java.lang.String name,
                                                              float weight,
                                                              int classAxis,
                                                              boolean sparseLabel,
                                                              boolean fromLogit)
```
    Returns a new instance of SoftmaxCrossEntropyLoss with the given arguments.
    
    Parameters:
    
    name - the name of the loss
    
    weight - the weight to apply on the loss value, default 1
    
    classAxis - the axis that represents the class probabilities, default -1
    
    sparseLabel - whether labels are integer array or probabilities, default true
    
    fromLogit - whether labels are log probabilities or un-normalized numbers
    
    Returns:
    
    a new instance of SoftmaxCrossEntropyLoss
  - hingeLoss
```
public static HingeLoss hingeLoss()
```
    Returns a new instance of HingeLoss with default arguments.
    
    Returns:
    
    a new instance of HingeLoss
  - hingeLoss
```
public static HingeLoss hingeLoss(java.lang.String name)
```
    Returns a new instance of HingeLoss with default arguments.
    
    Parameters:
    
    name - the name of the loss
    
    Returns:
    
    a new instance of HingeLoss
  - hingeLoss
```
public static HingeLoss hingeLoss(java.lang.String name,
                                  int margin,
                                  float weight)
```
    Returns a new instance of HingeLoss with the given arguments.
    
    Parameters:
    
    name - the name of the loss
    
    margin - the margin in hinge loss. Defaults to 1.0
    
    weight - the weight to apply on loss value, default 1
    
    Returns:
    
    a new instance of HingeLoss
  - addAccumulator
```
public void addAccumulator(java.lang.String key)
```
    Adds an accumulator for the results of the evaluation with the given key.
    
    Specified by:
    
    addAccumulator in class Evaluator
    
    Parameters:
    
    key - the key for the new accumulator
  - updateAccumulator
```
public void updateAccumulator(java.lang.String key,
                              NDList labels,
                              NDList predictions)
```
    Updates the evaluator with the given key based on a NDList of labels and predictions.
    This is a synchronized operation. You should only call it at the end of a batch or epoch.
    
    Specified by:
    
    updateAccumulator in class Evaluator
    
    Parameters:
    
    key - the key of the accumulator to update
    
    labels - a NDList of labels
    
    predictions - a NDList of predictions
  - resetAccumulator
```
public void resetAccumulator(java.lang.String key)
```
    Resets the evaluator value with the given key.
    
    Specified by:
    
    resetAccumulator in class Evaluator
    
    Parameters:
    
    key - the key of the accumulator to reset
  - getAccumulator
```
public float getAccumulator(java.lang.String key)
```
    Returns the accumulated evaluator value.
    
    Specified by:
    
    getAccumulator in class Evaluator
    
    Parameters:
    
    key - the key of the accumulator to get
    
    Returns:
    
    the accumulated value

Class Loss

Field Summary

Fields inherited from class ai.djl.training.evaluator.Evaluator

Constructor Summary

Method Summary

Methods inherited from class ai.djl.training.evaluator.Evaluator

Methods inherited from class java.lang.Object

Constructor Detail

Loss

Method Detail

l1Loss

l1Loss

l1Loss

l2Loss

l2Loss

l2Loss

sigmoidBinaryCrossEntropyLoss

sigmoidBinaryCrossEntropyLoss

sigmoidBinaryCrossEntropyLoss

softmaxCrossEntropyLoss

softmaxCrossEntropyLoss

softmaxCrossEntropyLoss

hingeLoss

hingeLoss

hingeLoss

addAccumulator

updateAccumulator

resetAccumulator

getAccumulator