12                     int n_batch, int n_input, int16_t* output) {  in ApplyLayerNorm()  argument
48             output[index] = static_cast<int16_t>(val5);  in ApplyLayerNorm()
54                                     int32_t n_col, int32_t* output) {  in MatrixScalarMultiplyAccumulate()  argument
60         output[i] += row_sum * scalar;  in MatrixScalarMultiplyAccumulate()
66                                             std::unique_ptr<int32_t[]>* output) {  in PrecomputeZeroPointTimesWeightWithBias()  argument
74     *output = std::make_unique<int32_t[]>(row);  in PrecomputeZeroPointTimesWeightWithBias()
76         memset(output->get(), 0, row * sizeof(int32_t));  in PrecomputeZeroPointTimesWeightWithBias()
78         memcpy(output->get(), bias_tensor, row * sizeof(int32_t));  in PrecomputeZeroPointTimesWeightWithBias()
81         MatrixScalarMultiplyAccumulate(weight_tensor, zero_point, row, col, output->get());  in PrecomputeZeroPointTimesWeightWithBias()
86 void ApplySigmoid(const int16_t* input, int32_t n_batch, int32_t n_input, int16_t* output) {  in ApplySigmoid()  argument
94             output[index] = sigmoid_output.raw();  in ApplySigmoid()
100               int16_t* output) {  in CwiseMul()  argument
107             output[index] = static_cast<int16_t>(gemmlowp::RoundingDivideByPOT(value, shift));  in CwiseMul()
113               int32_t n_batch, int32_t n_input, int32_t output_zp, int8_t* output) {  in CwiseMul()  argument
124             output[index] = static_cast<int8_t>(value);  in CwiseMul()
139               int16_t* output) {  in CwiseAdd()  argument
145             output[index] = static_cast<int16_t>(sum_clamped);  in CwiseAdd()
185             int32_t output = prod + *result;  in VectorBatchVectorCwiseProductAccumulate()  local
186             output = std::max(std::min(32767, output), -32768);  in VectorBatchVectorCwiseProductAccumulate()
187             *result++ = output;  in VectorBatchVectorCwiseProductAccumulate()