EncoderDecoderTransformer class - encoder_decoder library

Constructors

EncoderDecoderTransformer({required int sourceVocabSize, required int targetVocabSize, required int embedDim, required int numLayers, required int numHeads, required int maxSourceLen, required int maxTargetLen, int? ffnDim, double dropoutP = 0.0, Device device = Device.CPU, int seed = 0})

decoder → TransformerDecoder: final
embedDim → int: final
encoder → TransformerEncoder: final
hashCode → int: The hash code for this object.
no setterinherited
head → Linear: final
maxSourceLen → int: final
maxTargetLen → int: final
numHeads → int: final
numLayers → int: final
runtimeType → Type: A representation of the runtime type of the object.
no setterinherited
sourceEmb → Embedding: final
sourceVocabSize → int: final
srcPosEnc → SinusoidalPositionalEncoding: final
targetEmb → Embedding: final
targetVocabSize → int: final
tgtPosEnc → SinusoidalPositionalEncoding: final
training ↔ bool: Whether this module is in training mode. Layers that behave differently between training and inference (e.g. Dropout) read this flag in their call method. Defaults to training mode.
getter/setter pairinherited

call(Tensor srcTokens, Tensor tgtTokens) → Tensor: Full forward: encode source, decode target against memory, return logits [St, targetVocabSize] or [B, St, V].
decode(Tensor tgtTokens, Tensor memory) → Tensor: Decoder-side forward given a precomputed memory (encoder output). Applies causal masking to the decoder self-attention.
encode(Tensor srcTokens) → Tensor: Encoder-only pass returning "memory" for the decoder.
eval() → void: Put this module (and any registered submodules) into evaluation mode.
inherited
noSuchMethod(Invocation invocation) → dynamic: Invoked when a nonexistent method or property is accessed.
inherited
parameters() → List<Tensor>: Trainable tensors owned by this module (and its submodules).
override
submodules() → List<Module>: Submodules owned by this module. Subclasses that compose other modules should override this so train() / eval() propagate. Default: empty.
override
toString() → String: A string representation of this object.
inherited
train() → void: Put this module (and any registered submodules) into training mode.
inherited
zeroGrad() → void: Zero every parameter's gradient. Safe to call before each backward.
inherited