<!--
{
  "availability" : [
    "iOS: 14.0.0 -",
    "iPadOS: 14.0.0 -",
    "macCatalyst: 14.2.0 - 27.0.0",
    "macOS: 11.0.0 - 27.0.0",
    "tvOS: 14.0.0 -"
  ],
  "documentType" : "symbol",
  "framework" : "MLCompute",
  "identifier" : "/documentation/MLCompute/MLCRMSPropOptimizer",
  "metadataVersion" : "0.1.0",
  "role" : "Class",
  "symbol" : {
    "kind" : "Class",
    "modules" : [
      "ML Compute"
    ],
    "preciseIdentifier" : "c:objc(cs)MLCRMSPropOptimizer"
  },
  "title" : "MLCRMSPropOptimizer"
}
-->

# MLCRMSPropOptimizer

An optimizer that represents the root mean square propagation algorithm.

```
class MLCRMSPropOptimizer
```

## Topics

### Creating an RMSProp Optimizer

[`init(descriptor:)`](/documentation/MLCompute/MLCRMSPropOptimizer/init(descriptor:))

Creates an RMSProp optimizer with the descriptor you specify.

[`init(descriptor:momentumScale:alpha:epsilon:isCentered:)`](/documentation/MLCompute/MLCRMSPropOptimizer/init(descriptor:momentumScale:alpha:epsilon:isCentered:))

Creates an RMSProp optimizer with the descriptor, momentum scale, smoothing, epsilon, and option to compute the centered RMSProp that you specify.

### Inspecting an RMSProp Optimizer

[`momentumScale`](/documentation/MLCompute/MLCRMSPropOptimizer/momentumScale)

A hyper-parameter that specifies the momentum factor.

[`alpha`](/documentation/MLCompute/MLCRMSPropOptimizer/alpha)

The constant for smoothing.

[`epsilon`](/documentation/MLCompute/MLCRMSPropOptimizer/epsilon)

The epsilon value you use to improve numerical stability.

[`isCentered`](/documentation/MLCompute/MLCRMSPropOptimizer/isCentered)

A Boolean that indicates whether you compute the centered RMSProp.


---

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)