<!--
{
  "availability" : [
    "iOS: 11.0.0 -",
    "iPadOS: 11.0.0 -",
    "macCatalyst: 13.0.0 -",
    "macOS: 10.13.0 -",
    "tvOS: 11.0.0 -",
    "visionOS: 1.0.0 -",
    "watchOS: 4.0.0 -"
  ],
  "documentType" : "symbol",
  "framework" : "CoreML",
  "identifier" : "/documentation/CoreML/MLParameterKey/biases",
  "metadataVersion" : "0.1.0",
  "role" : "Type Property",
  "symbol" : {
    "kind" : "Type Property",
    "modules" : [
      "Core ML",
      "CoreML"
    ],
    "preciseIdentifier" : "c:objc(cs)MLParameterKey(cpy)biases"
  },
  "title" : "biases"
}
-->

# biases

The key you use to access the biases of a layer in a neural network model.

```
class var biases: MLParameterKey { get }
```

## Discussion

The value type for the [`biases`](/documentation/CoreML/MLParameterKey/biases) key is an [`MLMultiArray`](/documentation/CoreML/MLMultiArray). You must scope this key with the name of the specific neural network layer whose biases you’d like to access. See [`scoped(to:)`](/documentation/CoreML/MLParameterKey/scoped(to:)).

> Note:
> You can only override the biases of a model’s *updatable* layers. Model developers mark these layers as updatable with the [Core ML Tools](https://coremltools.readme.io/).

---

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)