Package io.github.kirstenali.deepj.transformer

Class GPTTransformerBuilder

java.lang.Object

io.github.kirstenali.deepj.transformer.GPTTransformerBuilder

public final class GPTTransformerBuilder
extends Object

Builder for GPT-style transformer stacks.

Assembles a TransformerStack of GPTTransformerBlocks (LayerNorm + multi-head self-attention + configurable FFN).

See Also:

• LlamaTransformerBuilder
• DeepSeekTransformerBuilder

Constructor Summary

Constructors

Constructor	Description
GPTTransformerBuilder()

Method Summary

Modifier and Type	Method	Description
TransformerStack	build()
GPTTransformerBuilder	dFF(int dFF)
GPTTransformerBuilder	dModel(int dModel)
GPTTransformerBuilder	ffnActivation(Supplier<ActivationFunction> factory)	Activation inside the FFN.
GPTTransformerBuilder	nHeads(int nHeads)
GPTTransformerBuilder	nLayers(int nLayers)
GPTTransformerBuilder	random(Random rnd)
GPTTransformerBuilder	seed(long seed)

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Details

GPTTransformerBuilder

public GPTTransformerBuilder()

Method Details

dModel

public GPTTransformerBuilder dModel(int dModel)

nHeads

public GPTTransformerBuilder nHeads(int nHeads)

dFF

public GPTTransformerBuilder dFF(int dFF)

nLayers

public GPTTransformerBuilder nLayers(int nLayers)

ffnActivation

public GPTTransformerBuilder ffnActivation(Supplier<ActivationFunction> factory)

Activation inside the FFN. Default: GELU.

seed

public GPTTransformerBuilder seed(long seed)

random

public GPTTransformerBuilder random(Random rnd)

build

public TransformerStack build()