Searched refs:QuantizationType (Results 1 – 6 of 6) sorted by relevance
31 enum class QuantizationType { NONE = 0, UINT8 }; enum54 QuantizationType quant_type;119 matrix.quant_type = QuantizationType::NONE; in GetHiddenLayerMatrix()136 matrix.quant_type = QuantizationType::NONE; in GetHiddenLayerBias()161 matrix.quant_type = QuantizationType::NONE; in GetSoftmaxMatrix()178 matrix.quant_type = QuantizationType::NONE; in GetSoftmaxBias()237 virtual QuantizationType embeddings_quant_type(int i) const { in embeddings_quant_type()238 return QuantizationType::NONE; in embeddings_quant_type()
67 QuantizationType quant_type() const { return quant_type_; } in quant_type()81 if (quant_type_ == QuantizationType::NONE) { in get_embedding()89 static int GetRowSizeInBytes(int cols, QuantizationType quant_type) { in GetRowSizeInBytes()91 case QuantizationType::NONE: in GetRowSizeInBytes()93 case QuantizationType::UINT8: in GetRowSizeInBytes()108 const QuantizationType quant_type_;
134 QuantizationType embeddings_quant_type(int i) const override { in embeddings_quant_type()136 return proto_->embeddings(i).is_quantized() ? QuantizationType::UINT8 in embeddings_quant_type()137 : QuantizationType::NONE; in embeddings_quant_type()
32 if (matrix.quant_type != QuantizationType::NONE) { in CheckNoQuantization()176 embedding_matrix->quant_type() != QuantizationType::NONE; in GetEmbeddingInternal()318 if (matrix.quant_type != QuantizationType::UINT8) { in EmbeddingNetwork()
91 QuantizationType embeddings_quant_type(int i) const override { in embeddings_quant_type()94 return QuantizationType::UINT8; in embeddings_quant_type()96 return QuantizationType::NONE; in embeddings_quant_type()
54 nlp_core::QuantizationType embeddings_quant_type(int i) const override { in embeddings_quant_type()114 nlp_core::QuantizationType embeddings_quant_type(int i) const override { in embeddings_quant_type()