skainet-lang-core/sk.ainet.lang.nn.dsl/TrainingConfig

TrainingConfig

class TrainingConfig<T : DType, V>(source)

Experimental DSL for configuring and running training. This is kept internal/experimental to avoid early API lock-in.

Constructors

constructor()

Functions

fun loss(block: () -> Loss)

Define the loss function.

fun model(block: () -> Module<T, V>)

Define the model to be trained.

fun optimizer(block: () -> Optimizer)

Define the optimizer.