SGD

Stochastic gradient descent optimizer.

NOTE: It's not an equivalent for keras.sgd, it is a pure SGD with simple 'variable' update by subtracting 'alpha' * 'delta' from it.

Constructors

SGD

fun SGD(learningRate: Float = 0.2f, clipGradient: ClipGradientAction = NoClipGradient())

SGD

fun SGD(clipGradient: ClipGradientAction = NoClipGradient())

clipGradient

Strategy of gradient clipping as subclass of ClipGradientAction.

optimizerName

open override val optimizerName: String

Returns optimizer name.