<!--
{
  "availability" : [
    "iOS: -",
    "iPadOS: -",
    "macCatalyst: -",
    "macOS: -",
    "tvOS: -",
    "visionOS: -",
    "watchOS: -"
  ],
  "documentType" : "symbol",
  "framework" : "Accelerate",
  "identifier" : "/documentation/Accelerate/BNNSLossFunction",
  "metadataVersion" : "0.1.0",
  "role" : "Structure",
  "symbol" : {
    "kind" : "Structure",
    "modules" : [
      "Accelerate"
    ],
    "preciseIdentifier" : "c:@EA@BNNSLossFunction"
  },
  "title" : "BNNSLossFunction"
}
-->

# BNNSLossFunction

Constants that describe loss functions.

```
struct BNNSLossFunction
```

## Topics

### Raw Values

### Loss Functions

[`init(_:)`](/documentation/Accelerate/BNNSLossFunction/init(_:))

[`init(rawValue:)`](/documentation/Accelerate/BNNSLossFunction/init(rawValue:))

[`rawValue`](/documentation/Accelerate/BNNSLossFunction/rawValue)

[`BNNSLossFunctionCategoricalCrossEntropy`](/documentation/Accelerate/BNNSLossFunctionCategoricalCrossEntropy)

Performs categorical cross entropy computation between input prediction and labels.

[`BNNSLossFunctionCosineDistance`](/documentation/Accelerate/BNNSLossFunctionCosineDistance)

Performs cosine distance loss computation between input predictions and labels.

[`BNNSLossFunctionHinge`](/documentation/Accelerate/BNNSLossFunctionHinge)

Performs Hinge loss computation between labels and unbounded zero-centered binary predictions.

[`BNNSLossFunctionHuber`](/documentation/Accelerate/BNNSLossFunctionHuber)

Huber loss computation between input logits and one-hot encoded labels.

[`BNNSLossFunctionLog`](/documentation/Accelerate/BNNSLossFunctionLog)

Log loss computation between labels and predictions.

[`BNNSLossFunctionMeanAbsoluteError`](/documentation/Accelerate/BNNSLossFunctionMeanAbsoluteError)

Mean absolute error (MAE) computation between input prediction and labels.

[`BNNSLossFunctionMeanSquareError`](/documentation/Accelerate/BNNSLossFunctionMeanSquareError)

Mean square error (MSE) computation between input logits and one-hot encoded labels.

[`BNNSLossFunctionSigmoidCrossEntropy`](/documentation/Accelerate/BNNSLossFunctionSigmoidCrossEntropy)

Sigmoid activation on input logits, and independent computation of cross-entropy loss for each class.

[`BNNSLossFunctionSoftmaxCrossEntropy`](/documentation/Accelerate/BNNSLossFunctionSoftmaxCrossEntropy)

Softmax activation on input logits, and computation of cross-entropy loss with one-hot encoded labels.

[`BNNSLossFunctionYolo`](/documentation/Accelerate/BNNSLossFunctionYolo)

You Only Look Once (YOLO) loss computation between prediction and ground truth labels.


---

Copyright &copy; 2026 Apple Inc. All rights reserved. | [Terms of Use](https://www.apple.com/legal/internet-services/terms/site.html) | [Privacy Policy](https://www.apple.com/privacy/privacy-policy)