/aosp_15_r20/external/tensorflow/tensorflow/lite/kernels/internal/optimized/ |
H A D | neon_tensor_utils.h | 29 int n_batch, float* result) { in MatrixBatchVectorMultiplyAccumulate() 38 int n_batch, in MatrixBatchVectorMultiplyAccumulate() 48 int n_batch, int32_t* scratch, in MatrixBatchVectorMultiplyAccumulate() 58 int n_batch, float* __restrict__ result, const float* per_channel_scale, in MatrixBatchVectorMultiplyAccumulate() 69 const float* __restrict__ vector, int n_batch, float* __restrict__ result) { in SparseMatrixBatchVectorMultiplyAccumulate1x4() 76 int m_rows, int m_cols, const float* __restrict__ vector, int n_batch, in SparseMatrixBatchVectorMultiplyAccumulate() 86 int n_batch, const int32_t input_offset, const int32_t output_multiplier, in SparseMatrixBatchVectorMultiplyAccumulate1x16() 100 const float* scaling_factors, int n_batch, float* __restrict__ result) { in SparseMatrixBatchVectorMultiplyAccumulate() 108 int32_t n_batch, int32_t n_input, int32_t n_output, int32_t output_zp, in MatrixBatchVectorMultiplyAccumulate() 118 int32_t n_batch, int32_t n_input, int32_t n_output, int32_t output_zp, in MatrixBatchVectorMultiplyAccumulate() [all …]
|
H A D | sse_tensor_utils.h | 39 int n_batch, float* result) { in MatrixBatchVectorMultiplyAccumulate() 52 const float* __restrict__ scaling_factors, int n_batch, in MatrixBatchVectorMultiplyAccumulate() 61 int n_batch, float* __restrict__ result, const float* per_channel_scale, in MatrixBatchVectorMultiplyAccumulate() 72 const float* __restrict__ scaling_factors, int n_batch, in MatrixBatchVectorMultiplyAccumulate() 82 const float* __restrict__ vector, int n_batch, float* __restrict__ result) { in SparseMatrixBatchVectorMultiplyAccumulate1x4() 91 int n_batch, const int32_t input_offset, const int32_t output_multiplier, in SparseMatrixBatchVectorMultiplyAccumulate1x16() 104 int m_rows, int m_cols, const float* __restrict__ vector, int n_batch, in SparseMatrixBatchVectorMultiplyAccumulate() 113 const float* __restrict__ scaling_factors, int n_batch, in SparseMatrixBatchVectorMultiplyAccumulate() 122 int32_t n_batch, int32_t n_input, int32_t n_output, int32_t output_zp, in MatrixBatchVectorMultiplyAccumulate() 132 int32_t n_batch, int32_t n_input, int32_t n_output, int32_t output_zp, in MatrixBatchVectorMultiplyAccumulate() [all …]
|
H A D | neon_tensor_utils.cc | 227 int n_batch, float* result) { in NeonMatrixBatchVectorMultiplyAccumulate() 276 const int8_t* ShuffleVectors(const int8_t* vectors, const int n_batch, in ShuffleVectors() 331 const int8_t* vectors, const float* scaling_factors, int n_batch, in DotprodMatrixBatchFourVectorMultiplyAccumulate() 459 const int8_t* vectors, const float* scaling_factors, int n_batch, in DotprodMatrixBatchFourVectorMultiplyAccumulate() 589 const int8_t* vectors, const float* scaling_factors, int n_batch, in DotprodMatrixBatchFourVectorMultiplyAccumulate() 622 const int8_t* vectors, const float* scaling_factors, int n_batch, in DotprodMatrixBatchPaddedFourVectorMultiplyAccumulate() 694 const int8_t* vectors, const float* scaling_factors, int n_batch, in DotprodMatrixBatchPaddedFourVectorMultiplyAccumulate() 705 const float* scaling_factors, int n_batch, float* __restrict__ result) { in DotprodSparseMatrixBatchVectorMultiplyAccumulate() 769 int32_t n_batch, int32_t n_input, in NeonMatrixBatchVectorMultiplyImpl() 879 int32_t multiplier, int32_t shift, int32_t n_batch, int32_t n_output, in NeonMatrixBatchVectorAccumulateImpl() [all …]
|
H A D | sse_tensor_utils.cc | 161 const float* __restrict__ vector, int n_batch, float* __restrict__ result) { in Avx2MatrixBatchVectorMultiplyAccumulateImpl() 202 const float* __restrict__ scaling_factors, int n_batch, in Avx2MatrixBatchVectorMultiplyAccumulateImpl() 294 const float* __restrict__ scaling_factors, int n_batch, in SseMatrixBatchVectorMultiplyAccumulateImpl() 379 const int8_t* input_to_gate_weights, int32_t n_batch, in SseCpuBackendGemm() 413 const float* __restrict__ scaling_factors, int n_batch, in SseMatrixBatchVectorMultiplyAccumulate() 424 const float* __restrict__ scaling_factors, int n_batch, int32_t* scratch, in SseMatrixBatchVectorMultiplyAccumulate() 475 const float* __restrict__ scaling_factors, int n_batch, in SseMatrixBatchVectorMultiplyAccumulate() 597 const float* __restrict__ scaling_factors, int n_batch, in SseSparseMatrixBatchVectorMultiplyAccumulate()
|
/aosp_15_r20/external/tensorflow/tensorflow/lite/kernels/internal/reference/ |
H A D | portable_tensor_utils.h | 60 int n_batch, float* result) { in MatrixBatchVectorMultiplyAccumulate() 69 int n_batch, in MatrixBatchVectorMultiplyAccumulate() 78 int n_batch, float* __restrict__ result, const float* per_channel_scale, in MatrixBatchVectorMultiplyAccumulate() 91 int n_batch, int32_t* scratch, in MatrixBatchVectorMultiplyAccumulate() 101 const float* __restrict__ vector, int n_batch, float* __restrict__ result) { in SparseMatrixBatchVectorMultiplyAccumulate1x4() 108 int m_rows, int m_cols, const float* __restrict__ vector, int n_batch, in SparseMatrixBatchVectorMultiplyAccumulate() 118 int n_batch, const int32_t input_offset, const int32_t output_multiplier, in SparseMatrixBatchVectorMultiplyAccumulate1x16() 132 const float* scaling_factors, int n_batch, float* __restrict__ result) { in SparseMatrixBatchVectorMultiplyAccumulate() 141 int32_t n_batch, int32_t n_input, int32_t n_output, int32_t output_zp, in MatrixBatchVectorMultiplyAccumulate() 151 int32_t n_batch, int32_t n_input, int32_t n_output, int32_t output_zp, in MatrixBatchVectorMultiplyAccumulate() [all …]
|
H A D | portable_tensor_utils.cc | 122 int n_batch, float* result) { in PortableMatrixBatchVectorMultiplyAccumulate() 141 int n_batch, float* __restrict__ result) { in PortableMatrixBatchVectorMultiplyAccumulate() 166 int n_batch, float* __restrict__ result, const float* per_channel_scale, in PortableMatrixBatchVectorMultiplyAccumulate() 209 const float* __restrict__ vector, int n_batch, float* __restrict__ result) { in PortableSparseMatrixBatchVectorMultiplyAccumulate1x4() 234 int n_batch, const int32_t input_offset, const int32_t output_multiplier, in PortableSparseMatrixBatchVectorMultiplyAccumulate1x16() 267 int m_rows, int m_cols, const float* __restrict__ vector, int n_batch, in PortableSparseMatrixBatchVectorMultiplyAccumulate() 297 const float* scaling_factors, int n_batch, float* __restrict__ result) { in PortableSparseMatrixBatchVectorMultiplyAccumulate() 331 int32_t n_batch, int32_t n_input, int32_t n_output, int32_t output_zp, in PortableMatrixBatchVectorMultiplyAccumulateImpl() 360 int32_t n_batch, int32_t n_input, int32_t n_output, int32_t output_zp, in PortableMatrixBatchVectorMultiplyAccumulate() 370 int32_t n_batch, int32_t n_input, int32_t n_output, int32_t output_zp, in PortableMatrixBatchVectorMultiplyAccumulate() [all …]
|
/aosp_15_r20/external/tensorflow/tensorflow/lite/kernels/ |
H A D | unidirectional_sequence_lstm_test.cc | 36 int n_batch, int n_input, int n_cell, int n_output, int sequence_length, in HybridUnidirectionalLSTMOpModel() 248 const int n_batch = 1; in TEST_F() local 310 const int n_batch = 1; in TEST_F() local 377 const int n_batch = 1; in TEST_P() local 440 const int n_batch = 1; in TEST_P() local 551 const int n_batch = 1; in TEST_F() local 613 const int n_batch = 1; in TEST_P() local 675 const int n_batch = 1; in TEST_P() local 1338 const int n_batch = 2; in TEST_F() local 1405 const int n_batch = 2; in TEST_P() local [all …]
|
H A D | bidirectional_sequence_lstm_test.cc | 32 BidirectionalLSTMOpModel(int n_batch, int n_input, int n_cell, int n_output, in BidirectionalLSTMOpModel() 430 const int n_batch = 1; in TEST_P() local 596 const int n_batch = 2; in TEST_P() local 762 const int n_batch = 1; in TEST() local 923 const int n_batch = 1; in TEST() local 1074 const int n_batch = 1; in TEST() local 1226 const int n_batch = 2; in TEST() local 1929 const int n_batch = 2; in TEST() local 2630 const int n_batch = 1; in TEST_P() local 2804 const int n_batch = 1; in TEST_P() local
|
H A D | lstm_eval.cc | 47 float* output, int m_rows, int m_cols, int n_batch, in MatrixBatchVectorMultiplyAccumulate() 187 const int n_batch, const int n_input, const int n_aux_input, in CalculateLstmGateFloat() 255 void UpdateLstmCellFloat(int n_batch, int n_cell, float* cell_state, in UpdateLstmCellFloat() 302 void CalculateLstmOutputFloat(int n_batch, int n_cell, int n_output, in CalculateLstmOutputFloat() 363 const int n_batch, const int n_input, const int n_aux_input, in CalculateLstmGateHybrid() 475 int n_batch, int n_cell, int n_output, const float* cell_state, in CalculateLstmOutputHybrid() 548 const int n_batch, const int n_input, const int n_output, const int n_cell, in CalculateLstmGateInteger8x8_16() 613 void UpdateLstmCellInteger(int n_batch, int n_cell, int16_t* cell_state, in UpdateLstmCellInteger() 659 int n_batch, int n_cell, int n_output, const int16_t* cell_state, in CalculateLstmOutputInteger8x8_16() 713 const int n_batch, const int n_input, const int n_output, const int n_cell, in CalculateLstmGateInteger8x8_8() [all …]
|
H A D | lstm_test.cc | 39 LSTMOpModel(int n_batch, int n_input, int n_cell, int n_output, bool use_cifg, in LSTMOpModel() 413 const int n_batch = 1; in TEST_P() local 490 const int n_batch = 1; in TEST_P() local 561 const int n_batch = 2; in TEST_P() local 1144 const int n_batch = 2; in TEST_P() local 1239 const int n_batch = 2; in TEST_P() local 1314 LSTMIntegerOpModel(int n_batch, int n_input, int n_cell, int n_output, in LSTMIntegerOpModel() 1598 const int n_batch = 2; in TEST() local 1760 const int n_batch = 2; in TEST() local 1933 const int n_batch = 2; in TEST() local [all …]
|
H A D | unidirectional_sequence_gru_test.cc | 34 explicit GRUOpModel(int n_batch, int n_input, int n_output, in GRUOpModel() 101 const int n_batch = 2; in TEST() local
|
H A D | optional_tensor_test.cc | 31 LSTMOpModel(int n_batch, int n_input, int n_cell, int n_output, bool use_cifg, in LSTMOpModel() 229 const int n_batch = 1; in TEST() local
|
H A D | unidirectional_sequence_gru.cc | 40 const int n_batch = input->dims->data[1]; in GruImpl() local 142 const int n_batch = input->dims->data[1]; in Prepare() local
|
/aosp_15_r20/packages/modules/NeuralNetworks/common/ |
D | QuantUtils.cpp | 12 int n_batch, int n_input, int16_t* output) { in ApplyLayerNorm() 86 void ApplySigmoid(const int16_t* input, int32_t n_batch, int32_t n_input, int16_t* output) { in ApplySigmoid() 99 void CwiseMul(const int16_t* input_1, const int16_t* input_2, int n_batch, int n_input, int shift, in CwiseMul() 113 int32_t n_batch, int32_t n_input, int32_t output_zp, int8_t* output) { in CwiseMul() 138 void CwiseAdd(const int16_t* input_1, const int16_t* input_2, int n_batch, int n_input, in CwiseAdd() 150 void CwiseClipping(int16_t* input, const int16_t clipping_value, int32_t n_batch, int32_t n_input) { in CwiseClipping() 164 void CwiseClipping(int8_t* input, const int8_t clipping_value, int32_t n_batch, int32_t n_input) { in CwiseClipping() 179 const int16_t* batch_vector, int n_batch, in VectorBatchVectorCwiseProductAccumulate()
|
D | QuantUtils.h | 36 int32_t shift, int32_t n_batch, int32_t n_input, in MatrixBatchVectorMultiplyAccumulate() 154 void ApplyTanh(const int16_t* input, int32_t n_batch, int32_t n_input, int16_t* output) { in ApplyTanh() 167 inline void ApplyTanh(int32_t integer_bits, const int16_t* input, int32_t n_batch, int32_t n_input, in ApplyTanh()
|
/aosp_15_r20/external/tensorflow/tensorflow/lite/kernels/internal/ |
H A D | portable_tensor_utils.h | 51 void VectorBatchVectorAdd(const T* vector, int v_size, int n_batch, in VectorBatchVectorAdd() 73 const T* batch_vector, int n_batch, in VectorBatchVectorCwiseProduct() 101 int n_batch, T* result) { in VectorBatchVectorCwiseProductAccumulate() 112 void VectorBatchVectorAssign(const T* vector, int v_size, int n_batch, in VectorBatchVectorAssign() 153 inline void BatchQuantizeFloats(const float* float_data_ptr, int n_batch, in BatchQuantizeFloats() 419 int v_size, int n_batch, in BatchVectorBatchVectorDotProduct()
|
/aosp_15_r20/external/tensorflow/tensorflow/lite/delegates/gpu/cl/kernels/ |
H A D | lstm_full_test.cc | 35 LSTMOpModel(int n_batch, int n_input, int n_cell, int n_output, bool use_cifg, in LSTMOpModel() 265 const int n_batch = 1; in TEST_P() local 339 const int n_batch = 1; in TEST_P() local 407 const int n_batch = 1; in TEST_P() local 970 const int n_batch = 1; in TEST_F() local 1056 const int n_batch = 1; in TEST_F() local 1120 const int n_batch = 1; in TEST_F() local
|
/aosp_15_r20/packages/modules/NeuralNetworks/common/cpu_operations/ |
D | LSTMTest.cpp | 78 LSTMOpModel(uint32_t n_batch, uint32_t n_input, uint32_t n_cell, uint32_t n_output, in LSTMOpModel() 273 const int n_batch = 1; in TEST() local 377 const int n_batch = 1; in TEST() local 474 const int n_batch = 2; in TEST() local
|
D | LayerNormLSTMTest.cpp | 87 LayerNormLSTMOpModel(uint32_t n_batch, uint32_t n_input, uint32_t n_cell, uint32_t n_output, in LayerNormLSTMOpModel() 287 const int n_batch = 2; in TEST() local
|
/aosp_15_r20/external/tensorflow/tensorflow/lite/tools/optimize/calibration/builtin_logging_ops/ |
H A D | lstm.cc | 46 const int n_batch, const int n_input, const int n_aux_input, in CalculateLstmGateFloat() 100 void UpdateLstmCellFloat(int n_batch, int n_cell, float* cell_state, in UpdateLstmCellFloat() 124 int n_batch, int n_cell, int n_output, const float* cell_state, in CalculateLstmOutputCalibration() 181 const TfLiteLSTMParams* params, int n_batch, int n_cell, int n_input, in LstmStepCalibration() 294 int max_time, n_batch; in EvalCalibration() local
|
/aosp_15_r20/packages/modules/NeuralNetworks/runtime/test/specs/V1_3/ |
D | bidirectional_sequence_lstm.mod.py | 22 n_batch, argument 455 n_batch = 1 variable 503 n_batch = 1 variable
|
/aosp_15_r20/packages/modules/NeuralNetworks/runtime/test/specs/V1_2/ |
D | quantized_lstm.mod.py | 20 n_batch = 2 variable 107 n_batch = 1 variable
|
D | layer_norm_lstm.mod.py | 22 n_batch = 2 variable 194 n_batch = 2 variable
|
/aosp_15_r20/external/pytorch/aten/src/ATen/native/cpu/ |
H A D | batch_norm_kernel.cpp | 80 int64_t n_batch = input.size(0); in batch_norm_cpu_contiguous_impl() local 132 int64_t n_batch = input.size(0); in batch_norm_cpu_channels_last_impl() local 183 int64_t n_batch = input.size(0); in batch_norm_cpu_collect_stats_contiguous_impl() local 414 int64_t n_batch = input.size(0); in batch_norm_cpu_backward_contiguous_impl() local 716 int64_t n_batch = input.size(0); in batch_norm_cpu_contiguous_impl() local 780 int64_t n_batch = input.size(0); in batch_norm_cpu_channels_last_impl() local 835 int64_t n_batch = input.size(0); in batch_norm_cpu_collect_stats_contiguous_internal() local 1004 int64_t n_batch = input.size(0); in batch_norm_cpu_backward_contiguous_internal() local
|
/aosp_15_r20/external/pytorch/aten/src/ATen/ |
H A D | SparseCsrTensorUtils.h | 335 int64_t n_batch = numBatchDimensions(self); in getBlockSize() local 341 int64_t n_batch = numBatchDimensions(self); in getSymIntBlockSize() local
|