GlorotUniformInitializationStrategy<T>
Models & Types · AiDotNet.Initialization
Glorot and Bengio’s uniform initialization (TensorFlow’s and Keras’s default): weights from U(−√(6 / (fan_in + fan_out)), √(6 / (fan_in + fan_out))) and zero biases.
Inheritance: InitializationStrategyBase<T> → GlorotUniformInitializationStrategy<T>
How It Works
For Beginners: Random starting weights sized so a signal keeps roughly the same spread as it passes through each layer, forward and backward. It is what the Soft Actor-Critic authors’ TensorFlow code starts every network with.
Constructors
| Constructor | Summary |
|---|---|
GlorotUniformInitializationStrategy(Random random) |
Creates the initializer. |
Properties
| Property | Summary |
|---|---|
bool IsLazy { get; } |
|
bool LoadFromExternal { get; } |
Methods
| Method | Summary |
|---|---|
void InitializeBiases(Tensor<T> biases) |
|
void InitializeWeights(Tensor<T> weights, int inputSize, int outputSize) |