SequentialModelBuilder

Java-friendly sequential model builder that mirrors the Kotlin DSL.

Example usage from Java:

Module model = new SequentialModelBuilder(ctx)
    .input(784)
    .dense(128)
    .relu()
    .dense(10)
    .build();

ctx

The execution context used for weight initialization.

dtype

The data type for the model. Defaults to FP32.

Constructors

constructor(ctx: ExecutionContext, dtype: DType = FP32)

Builds and returns the sequential model.

Adds a dense (fully connected) layer.

fun flatten(startDim: Int = 1, endDim: Int = -1): SequentialModelBuilder

Adds a Flatten layer.

Adds a GELU activation.

Sets the input size for the network. Must be called first.

Adds a ReLU activation.

Adds a Sigmoid activation.

Adds a SiLU (Swish) activation.

Adds a Softmax activation along the given dimension.