Home
last modified time | relevance | path

Searched refs:fc_accum_depth (Results 1 – 1 of 1) sorted by relevance

/aosp12/packages/modules/NeuralNetworks/common/operations/
H A DQuantizedLSTM.cpp88 const int fc_accum_depth = ArraySize(weights_dims, 0); in quantizedLstmStep() local
110 for (int d = 0; d < fc_accum_depth; ++d) { in quantizedLstmStep()
111 int16_t input_val = concat_temp_data_uint8[b * fc_accum_depth + d] - 128; in quantizedLstmStep()
113 weights_data_uint8[out_c * fc_accum_depth + d] - weights_zero_point; in quantizedLstmStep()