Module activation 

Activation functions with forward and backward passes for tensor computations.

Provides element-wise activation functions (ReLU, LeakyReLU, Sigmoid, Tanh, Softmax, GELU, Swish) with both forward evaluation and gradient (backward) computation.

Structs

ActivationConfig

Configuration for an activation function.

ActivationStats

Runtime statistics for a TensorActivation instance.

TensorActivation

Activation layer with forward/backward support and call statistics.

Enums

ActivationType

Type of activation function.