Type Members

1.  case class BertEncoder(tokenEmbedding: Embedding, segmentEmbedding: Embedding, positionalEmbedding: Constant, blocks: Seq[TransformerEncoderBlock]) extends GenericModule[(Variable, Variable), Variable] with Product with Serializable

   BertEncoder module

   BertEncoder module

   Input is (tokens, segments) where tokens and segments are both (batch,num tokens) long tensor.

   Output is (batch, num tokens, out dimension)

2.  case class BertLoss(pretrain: BertPretrainModule, mlmLoss: LossFunction, wholeSentenceLoss: LossFunction) extends GenericModule[BertLossInput, Variable] with Product with Serializable
3.  case class BertLossInput(input: BertPretrainInput, maskedLanguageModelTarget: STen, wholeSentenceTarget: STen) extends Product with Serializable
4.  case class BertPretrainInput(tokens: Variable, segments: Variable, positions: STen) extends Product with Serializable
5.  case class BertPretrainModule(encoder: BertEncoder, mlm: MaskedLanguageModelModule, wholeSentenceBinaryClassifier: MLP) extends GenericModule[BertPretrainInput, BertPretrainOutput] with Product with Serializable
6.  case class BertPretrainOutput(encoded: Variable, languageModelScores: Variable, wholeSentenceBinaryClassifierScore: Variable) extends Product with Serializable
7.  case class MaskedLanguageModelModule(mlp: MLP) extends GenericModule[(Variable, STen), Variable] with Product with Serializable

   Masked Language Model Input of (embedding, positions) Embedding of size (batch, num tokens, embedding dim) Positions of size (batch, max num tokens) long tensor indicating which positions to make predictions on Output (batch, len(Positions), vocabulary size)