# BNNS.SGDMomentumOptimizer An optimizer that uses the stochastic gradient descent (SGD) with the momentum optimization method. ``` struct SGDMomentumOptimizer ``` ## Topics ### Creating an SGD with Momentum Optimizer [`init(learningRate: Float, momentum: Float, gradientScale: Float, regularizationScale: Float, clipsGradientsTo: ClosedRange?, usesNestrovMomentum: Bool, regularizationFunction: BNNSOptimizerRegularizationFunction, sgdMomentumVariant: BNNSOptimizerSGDMomentumVariant)`](/documentation/Accelerate/BNNS/SGDMomentumOptimizer/init(learningRate:momentum:gradientScale:regularizationScale:clipsGradientsTo:usesNestrovMomentum:regularizationFunction:sgdMomentumVariant:)) Returns a new stochastic gradient descent (SGD) with momentum optimizer object. ### Inspecting the Properties of an SGD with Momentum Optimizer [`var learningRate: Float`](/documentation/Accelerate/BNNS/SGDMomentumOptimizer/learningRate) A value that specifies the learning rate. [`var momentum: Float`](/documentation/Accelerate/BNNS/SGDMomentumOptimizer/momentum) The rate of momentum decay. [`var gradientScale: Float`](/documentation/Accelerate/BNNS/SGDMomentumOptimizer/gradientScale) A value that specifies the gradient scaling factor. [`var regularizationScale: Float`](/documentation/Accelerate/BNNS/SGDMomentumOptimizer/regularizationScale) A value that specifies the regularization scaling factor. [`var gradientBounds: ClosedRange?`](/documentation/Accelerate/BNNS/SGDMomentumOptimizer/gradientBounds) The values for the minimum and maximum gradients. [`var gradientClipping: BNNS.GradientClipping`](/documentation/Accelerate/BNNS/SGDMomentumOptimizer/gradientClipping) The gradient clipping. [`enum GradientClipping`](/documentation/Accelerate/BNNS/GradientClipping) Constants that describe clipping functions. [`var usesNestrovMomentum: Bool`](/documentation/Accelerate/BNNS/SGDMomentumOptimizer/usesNestrovMomentum) A Boolean value that specifies whether to use Nesterov momentum update. [`var regularizationFunction: BNNSOptimizerRegularizationFunction`](/documentation/Accelerate/BNNS/SGDMomentumOptimizer/regularizationFunction) The variable that specifies the regularization function. [`var sgdMomentumVariant: BNNSOptimizerSGDMomentumVariant`](/documentation/Accelerate/BNNS/SGDMomentumOptimizer/sgdMomentumVariant) The variable that specifies the momentum variant. [`var accumulatorCountMultiplier: Int`](/documentation/Accelerate/BNNS/SGDMomentumOptimizer/accumulatorCountMultiplier) The number of accumulators required for each parameter. ### Applying an Optimization Step ### Initializers [`init(learningRate: Float, momentum: Float, gradientScale: Float, regularizationScale: Float, clipsGradientsTo: ClosedRange?, usesNesterovMomentum: Bool, regularizationFunction: BNNSOptimizerRegularizationFunction, sgdMomentumVariant: BNNSOptimizerSGDMomentumVariant)`](/documentation/Accelerate/BNNS/SGDMomentumOptimizer/init(learningRate:momentum:gradientScale:regularizationScale:clipsGradientsTo:usesNesterovMomentum:regularizationFunction:sgdMomentumVariant:)) [`init(learningRate: Float, momentum: Float, gradientScale: Float, regularizationScale: Float, gradientClipping: BNNS.GradientClipping, usesNesterovMomentum: Bool, regularizationFunction: BNNSOptimizerRegularizationFunction, sgdMomentumVariant: BNNSOptimizerSGDMomentumVariant)`](/documentation/Accelerate/BNNS/SGDMomentumOptimizer/init(learningRate:momentum:gradientScale:regularizationScale:gradientClipping:usesNesterovMomentum:regularizationFunction:sgdMomentumVariant:)) ### Instance Properties [`var usesNesterovMomentum: Bool`](/documentation/Accelerate/BNNS/SGDMomentumOptimizer/usesNesterovMomentum) ## Relationships ### Conforms To [`BNNSOptimizer`](/documentation/Accelerate/BNNSOptimizer) --- Copyright © 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)