---
title: "BNNS.LossFunction.huber(huberDelta:)"
framework: accelerate
role: symbol
role_heading: Case
path: "accelerate/bnns/lossfunction/huber(huberdelta:)"
---

# BNNS.LossFunction.huber(huberDelta:)

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

## Declaration

```swift
case huber(huberDelta: Float)
```

## See Also

### Loss Functions

- [BNNS.LossFunction.categoricalCrossEntropy](accelerate/bnns/lossfunction/categoricalcrossentropy.md)
- [BNNS.LossFunction.cosineDistance](accelerate/bnns/lossfunction/cosinedistance.md)
- [BNNS.LossFunction.hinge](accelerate/bnns/lossfunction/hinge.md)
- [BNNS.LossFunction.log](accelerate/bnns/lossfunction/log.md)
- [BNNS.LossFunction.meanAbsoluteError](accelerate/bnns/lossfunction/meanabsoluteerror.md)
- [BNNS.LossFunction.meanSquareError](accelerate/bnns/lossfunction/meansquareerror.md)
- [BNNS.LossFunction.sigmoidCrossEntropy(labelSmoothing:)](accelerate/bnns/lossfunction/sigmoidcrossentropy(labelsmoothing:).md)
- [BNNS.LossFunction.softmaxCrossEntropy(labelSmoothing:)](accelerate/bnns/lossfunction/softmaxcrossentropy(labelsmoothing:).md)
- [BNNS.LossFunction.yolo(parameters:)](accelerate/bnns/lossfunction/yolo(parameters:).md)
- [BNNS.LossFunction.YoloParameters](accelerate/bnns/lossfunction/yoloparameters.md)