SentenceEncoder class - sentence_encoder library

Constructors

SentenceEncoder({required int vocabSize, required int maxSeqLen, required int embedDim, int numLayers = 4, int numHeads = 4, double dropoutP = 0.0, PoolingMode pooling = PoolingMode.mean, bool normalize = true, double eps = 1e-12, Device device = Device.CPU, int seed = 0})
SentenceEncoder.wrap(TokenEncoder backbone, {PoolingMode pooling = PoolingMode.mean, bool normalize = true, double eps = 1e-12}): Wrap an already-constructed backbone. Useful for loading a pretrained transformer and swapping the pooling head.

backbone → TokenEncoder: final
embedDim → int: no setter
eps → double: final
hashCode → int: The hash code for this object.
no setterinherited
normalize → bool: final
pooling → PoolingMode: final
runtimeType → Type: A representation of the runtime type of the object.
no setterinherited
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 tokens) → Tensor: Encode a single sequence of token indices [seqLen] to a [1, embedDim] sentence embedding.
encodeBatch(List<Tensor> tokenSequences) → Tensor: Encode a batch of token sequences and stack the results into a single [N, embedDim] tensor. Each sequence still runs through the backbone independently (the transformer here does not batch over a leading dimension), so this is a convenience wrapper — still differentiable when called under a live grad tape.
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