From 0cbfda629dd8f684e625173341bab972f004222c Mon Sep 17 00:00:00 2001 From: Michele Di Giorgio Date: Thu, 13 Jun 2019 17:01:29 +0100 Subject: COMPMID-2343: Add layer normalization support in NELSTMLayer Change-Id: I1f620d70c6eaadfb9e3a1b345de350ac0253b65c Signed-off-by: Michele Di Giorgio Reviewed-on: https://review.mlplatform.org/c/1366 Tested-by: Arm Jenkins Reviewed-by: Manuel Bottini Comments-Addressed: Arm Jenkins Reviewed-by: Georgios Pinitas --- src/runtime/NEON/functions/NELSTMLayer.cpp | 193 +++++++++++++++++++++++++---- 1 file changed, 167 insertions(+), 26 deletions(-) (limited to 'src/runtime') diff --git a/src/runtime/NEON/functions/NELSTMLayer.cpp b/src/runtime/NEON/functions/NELSTMLayer.cpp index 42b805794b..ee2b2f4b28 100644 --- a/src/runtime/NEON/functions/NELSTMLayer.cpp +++ b/src/runtime/NEON/functions/NELSTMLayer.cpp @@ -38,15 +38,18 @@ using namespace arm_compute; using namespace arm_compute::misc::shape_calculator; NELSTMLayer::NELSTMLayer(std::shared_ptr memory_manager) - : _memory_group(std::move(memory_manager)), _fully_connected_input_gate(), _gemm_input_gate(), _transpose_input_gate(), _accum_input_gate1(), _accum_input_gate2(), _subtract_input_gate(), - _pixelwise_mul_input_gate(), _activation_input_gate(), _fully_connected_forget_gate(), _gemm_forget_gate(), _transpose_forget_gate(), _accum_forget_gate1(), _accum_forget_gate2(), - _pixelwise_mul_forget_gate(), _activation_forget_gate(), _fully_connected_cell_state(), _gemm_cell_state1(), _gemm_cell_state2(), _transpose_cell_state(), _accum_cell_state1(), _accum_cell_state2(), - _pixelwise_mul_cell_state1(), _activation_cell_state(), _cell_clip(), _pixelwise_mul_cell_state2(), _fully_connected_output(), _gemm_output(), _pixelwise_mul_output_state1(), _transpose_output(), - _accum_output1(), _accum_output2(), _activation_output(), _activation_output_state(), _pixelwise_mul_output_state2(), _fully_connected_output_state(), _gemm_output_state(), _accum_output_state(), - _projection_clip(), _copy_cell_state(), _copy_output(), _concat_scratch_buffer(), _concat_inputs_forget_gate(), _concat_weights_forget_gate(), _concat_weights_input_gate(), _concat_weights_output(), - _input_gate_out1(), _input_gate_out2(), _input_gate_out3(), _input_gate_out4(), _forget_gate_out1(), _forget_gate_out2(), _forget_gate_out3(), _forget_gate_out4(), _forget_gate_out5(), - _forget_gate_out6(), _cell_state_out1(), _cell_state_out2(), _cell_state_out3(), _cell_state_out4(), _cell_state_out5(), _output1(), _output2(), _output3(), _output4(), _cell_state_activation(), - _output_state1(), _ones(), _run_peephole_opt(false), _run_cifg_opt(false), _perform_cell_clipping(false), _has_projection_weights(false), _perform_projection_clipping(false), _is_prepared(false) + : _memory_group(std::move(memory_manager)), _fully_connected_input_gate(), _accum_input_gate1(), _subtract_input_gate(), _pixelwise_mul_input_gate(), _activation_input_gate(), + _fully_connected_forget_gate(), _accum_forget_gate1(), _pixelwise_mul_forget_gate(), _activation_forget_gate(), _fully_connected_cell_state(), _gemm_cell_state1(), _transpose_cell_state(), + _accum_cell_state1(), _accum_cell_state2(), _pixelwise_mul_cell_state1(), _activation_cell_state(), _cell_clip(), _pixelwise_mul_cell_state2(), _fully_connected_output(), + _pixelwise_mul_output_state1(), _accum_output1(), _activation_output(), _activation_output_state(), _pixelwise_mul_output_state2(), _fully_connected_output_state(), _projection_clip(), + _copy_cell_state(), _copy_output(), _concat_scratch_buffer(), _concat_inputs_forget_gate(), _concat_weights_forget_gate(), _concat_weights_input_gate(), _concat_weights_output(), + _mean_std_norm_input_gate(), _pixelwise_mul_input_gate_coeff(), _accum_input_gate_bias(), _mean_std_norm_forget_gate(), _pixelwise_mul_forget_gate_coeff(), _accum_forget_gate_bias(), + _mean_std_norm_cell_gate(), _pixelwise_mul_cell_gate_coeff(), _accum_cell_gate_bias(), _mean_std_norm_output_gate(), _pixelwise_mul_output_gate_coeff(), _accum_output_gate_bias(), _input_gate_out1(), + _input_gate_out2(), _input_gate_out3(), _input_gate_out4(), _forget_gate_out1(), _forget_gate_out2(), _forget_gate_out3(), _forget_gate_out4(), _forget_gate_out5(), _forget_gate_out6(), + _cell_state_out1(), _cell_state_out2(), _cell_state_out3(), _cell_state_out4(), _cell_state_out5(), _output1(), _output2(), _output3(), _output4(), _cell_state_activation(), _output_state1(), _ones(), + _input_layer_norm_out1(), _input_layer_norm_out2(), _forget_layer_norm_out1(), _forget_layer_norm_out2(), _cell_layer_norm_out1(), _cell_layer_norm_out2(), _output_layer_norm_out1(), + _output_layer_norm_out2(), _run_peephole_opt(false), _run_cifg_opt(false), _perform_cell_clipping(false), _has_projection_weights(false), _perform_projection_clipping(false), _is_prepared(false), + _is_layer_norm_lstm(false) { } @@ -65,6 +68,8 @@ void NELSTMLayer::configure(const ITensor *input, output_state_in, cell_state_in, scratch_buffer, output_state_out, cell_state_out, output); + _is_layer_norm_lstm = lstm_params.use_layer_norm(); + // Set lstm parameters LSTMParams lstm_params_info; if(lstm_params.has_peephole_opt()) @@ -117,7 +122,7 @@ void NELSTMLayer::configure(const ITensor *input, _concat_weights_forget_gate.configure(weights_vector, &_forget_gate_out6, Window::DimX); _memory_group.manage(&_forget_gate_out5); - _fully_connected_forget_gate.configure(&_forget_gate_out2, &_forget_gate_out6, forget_gate_bias, &_forget_gate_out5); + _fully_connected_forget_gate.configure(&_forget_gate_out2, &_forget_gate_out6, (_is_layer_norm_lstm) ? nullptr : forget_gate_bias, &_forget_gate_out5); _memory_group.manage(&_forget_gate_out1); _memory_group.manage(&_forget_gate_out3); _forget_gate_out6.allocator()->allocate(); @@ -130,7 +135,7 @@ void NELSTMLayer::configure(const ITensor *input, _run_peephole_opt = true; _memory_group.manage(&_forget_gate_out4); _pixelwise_mul_forget_gate.configure(cell_state_in, lstm_params.cell_to_forget_weights(), &_forget_gate_out4, 1, ConvertPolicy::SATURATE, RoundingPolicy::TO_ZERO); - _accum_forget_gate2.configure(&_forget_gate_out5, &_forget_gate_out4, &_forget_gate_out3, ConvertPolicy::SATURATE); + _accum_forget_gate1.configure(&_forget_gate_out5, &_forget_gate_out4, &_forget_gate_out3, ConvertPolicy::SATURATE); _forget_gate_out4.allocator()->allocate(); _forget_gate_out5.allocator()->allocate(); forget_gate_out = &_forget_gate_out3; @@ -139,6 +144,20 @@ void NELSTMLayer::configure(const ITensor *input, { _forget_gate_out3.allocator()->allocate(); } + if(_is_layer_norm_lstm) + { + _forget_layer_norm_out1.allocator()->init(TensorInfo(cell_state_shape, 1, input->info()->data_type())); + _forget_layer_norm_out2.allocator()->init(TensorInfo(cell_state_shape, 1, input->info()->data_type())); + _memory_group.manage(&_forget_layer_norm_out1); + _memory_group.manage(&_forget_layer_norm_out2); + _mean_std_norm_forget_gate.configure(forget_gate_out); + _pixelwise_mul_forget_gate_coeff.configure(forget_gate_out, lstm_params.forget_layer_norm_weights(), &_forget_layer_norm_out1, 1, ConvertPolicy::SATURATE, RoundingPolicy::TO_ZERO); + // forget_gate_out is going to be reassigned, so allocate the tensor that it was assigned to before + forget_gate_out->allocator()->allocate(); + _accum_forget_gate_bias.configure(&_forget_layer_norm_out1, forget_gate_bias, &_forget_layer_norm_out2, ConvertPolicy::SATURATE); + _forget_layer_norm_out1.allocator()->allocate(); + forget_gate_out = &_forget_layer_norm_out2; + } _activation_forget_gate.configure(forget_gate_out, nullptr, ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::LOGISTIC)); // Configure block that calculates the input gate @@ -170,7 +189,7 @@ void NELSTMLayer::configure(const ITensor *input, _memory_group.manage(&_input_gate_out1); _memory_group.manage(&_input_gate_out4); - _fully_connected_input_gate.configure(&_forget_gate_out2, &_input_gate_out2, lstm_params.input_gate_bias(), &_input_gate_out3); + _fully_connected_input_gate.configure(&_forget_gate_out2, &_input_gate_out2, (_is_layer_norm_lstm) ? nullptr : lstm_params.input_gate_bias(), &_input_gate_out3); _input_gate_out2.allocator()->allocate(); input_gate_out = &_input_gate_out3; @@ -178,7 +197,7 @@ void NELSTMLayer::configure(const ITensor *input, { _memory_group.manage(&_input_gate_out4); _pixelwise_mul_input_gate.configure(cell_state_in, lstm_params.cell_to_input_weights(), &_input_gate_out4, 1, ConvertPolicy::SATURATE, RoundingPolicy::TO_ZERO); - _accum_input_gate2.configure(&_input_gate_out3, &_input_gate_out4, &_input_gate_out1, ConvertPolicy::SATURATE); + _accum_input_gate1.configure(&_input_gate_out3, &_input_gate_out4, &_input_gate_out1, ConvertPolicy::SATURATE); _input_gate_out3.allocator()->allocate(); _input_gate_out4.allocator()->allocate(); input_gate_out = &_input_gate_out1; @@ -187,6 +206,21 @@ void NELSTMLayer::configure(const ITensor *input, { _input_gate_out1.allocator()->allocate(); } + + if(_is_layer_norm_lstm) + { + _input_layer_norm_out1.allocator()->init(TensorInfo(cell_state_shape, 1, input->info()->data_type())); + _input_layer_norm_out2.allocator()->init(TensorInfo(cell_state_shape, 1, input->info()->data_type())); + _memory_group.manage(&_input_layer_norm_out1); + _memory_group.manage(&_input_layer_norm_out2); + _mean_std_norm_input_gate.configure(input_gate_out); + _pixelwise_mul_input_gate_coeff.configure(input_gate_out, lstm_params.input_layer_norm_weights(), &_input_layer_norm_out1, 1, ConvertPolicy::SATURATE, RoundingPolicy::TO_ZERO); + // input_gate_out is going to be reassigned, so allocate the tensor that it was assigned to before + input_gate_out->allocator()->allocate(); + _accum_input_gate_bias.configure(&_input_layer_norm_out1, lstm_params.input_gate_bias(), &_input_layer_norm_out2, ConvertPolicy::SATURATE); + _input_layer_norm_out1.allocator()->allocate(); + input_gate_out = &_input_layer_norm_out2; + } _activation_input_gate.configure(input_gate_out, nullptr, ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::LOGISTIC)); } @@ -200,7 +234,7 @@ void NELSTMLayer::configure(const ITensor *input, _cell_state_out5.allocator()->init(TensorInfo(cell_state_shape, 1, input->info()->data_type())); _memory_group.manage(&_cell_state_out1); - _fully_connected_cell_state.configure(input, input_to_cell_weights, cell_bias, &_cell_state_out1); + _fully_connected_cell_state.configure(input, input_to_cell_weights, (_is_layer_norm_lstm) ? nullptr : cell_bias, &_cell_state_out1); _memory_group.manage(&_cell_state_out2); _transpose_cell_state.configure(recurrent_to_cell_weights, &_cell_state_out2); _memory_group.manage(&_cell_state_out3); @@ -208,10 +242,25 @@ void NELSTMLayer::configure(const ITensor *input, _cell_state_out2.allocator()->allocate(); _memory_group.manage(&_cell_state_out4); _accum_cell_state1.configure(&_cell_state_out1, &_cell_state_out3, &_cell_state_out4, ConvertPolicy::SATURATE); - _activation_cell_state.configure(&_cell_state_out4, nullptr, activation_info); + Tensor *cell_state_out_ptr = &_cell_state_out4; + if(_is_layer_norm_lstm) + { + _cell_layer_norm_out1.allocator()->init(TensorInfo(cell_state_shape, 1, input->info()->data_type())); + _cell_layer_norm_out2.allocator()->init(TensorInfo(cell_state_shape, 1, input->info()->data_type())); + _memory_group.manage(&_cell_layer_norm_out1); + _memory_group.manage(&_cell_layer_norm_out2); + _mean_std_norm_cell_gate.configure(cell_state_out_ptr); + _pixelwise_mul_cell_gate_coeff.configure(cell_state_out_ptr, lstm_params.cell_layer_norm_weights(), &_cell_layer_norm_out1, 1, ConvertPolicy::SATURATE, RoundingPolicy::TO_ZERO); + // cell_state_out_ptr is going to be reassigned, so allocate the tensor that it was assigned to before + cell_state_out_ptr->allocator()->allocate(); + _accum_cell_gate_bias.configure(&_cell_layer_norm_out1, cell_bias, &_cell_layer_norm_out2, ConvertPolicy::SATURATE); + _cell_layer_norm_out1.allocator()->allocate(); + cell_state_out_ptr = &_cell_layer_norm_out2; + } + _activation_cell_state.configure(cell_state_out_ptr, nullptr, activation_info); _memory_group.manage(&_cell_state_out5); - _pixelwise_mul_cell_state1.configure(&_cell_state_out4, input_gate_out, &_cell_state_out5, 1, ConvertPolicy::SATURATE, RoundingPolicy::TO_ZERO); - _cell_state_out4.allocator()->allocate(); + _pixelwise_mul_cell_state1.configure(cell_state_out_ptr, input_gate_out, &_cell_state_out5, 1, ConvertPolicy::SATURATE, RoundingPolicy::TO_ZERO); + cell_state_out_ptr->allocator()->allocate(); _pixelwise_mul_cell_state2.configure(forget_gate_out, cell_state_in, &_cell_state_out3, 1, ConvertPolicy::SATURATE, RoundingPolicy::TO_ZERO); _accum_cell_state2.configure(&_cell_state_out5, &_cell_state_out3, &_cell_state_out1, ConvertPolicy::SATURATE); _cell_state_out3.allocator()->allocate(); @@ -238,7 +287,7 @@ void NELSTMLayer::configure(const ITensor *input, _memory_group.manage(&_output1); _memory_group.manage(&_output4); - _fully_connected_output.configure(&_forget_gate_out2, &_output2, output_gate_bias, &_output4); + _fully_connected_output.configure(&_forget_gate_out2, &_output2, (_is_layer_norm_lstm) ? nullptr : output_gate_bias, &_output4); _output2.allocator()->allocate(); _forget_gate_out2.allocator()->allocate(); @@ -250,7 +299,7 @@ void NELSTMLayer::configure(const ITensor *input, _memory_group.manage(&_output3); _pixelwise_mul_output_state1.configure(&_cell_state_out1, lstm_params.cell_to_output_weights(), &_output3, 1, ConvertPolicy::SATURATE, RoundingPolicy::TO_ZERO); - _accum_output2.configure(&_output4, &_output3, &_output1, ConvertPolicy::SATURATE); + _accum_output1.configure(&_output4, &_output3, &_output1, ConvertPolicy::SATURATE); _output4.allocator()->allocate(); output_gate_out = &_output1; @@ -261,6 +310,20 @@ void NELSTMLayer::configure(const ITensor *input, { _output1.allocator()->allocate(); } + if(_is_layer_norm_lstm) + { + _output_layer_norm_out1.allocator()->init(TensorInfo(cell_state_shape, 1, input->info()->data_type())); + _output_layer_norm_out2.allocator()->init(TensorInfo(cell_state_shape, 1, input->info()->data_type())); + _memory_group.manage(&_output_layer_norm_out1); + _memory_group.manage(&_output_layer_norm_out2); + _mean_std_norm_output_gate.configure(output_gate_out); + _pixelwise_mul_output_gate_coeff.configure(output_gate_out, lstm_params.output_layer_norm_weights(), &_output_layer_norm_out1, 1, ConvertPolicy::SATURATE, RoundingPolicy::TO_ZERO); + // output_gate_out is going to be reassigned, so allocate the tensor that it was assigned to before + output_gate_out->allocator()->allocate(); + _accum_output_gate_bias.configure(&_output_layer_norm_out1, output_gate_bias, &_output_layer_norm_out2, ConvertPolicy::SATURATE); + _output_layer_norm_out1.allocator()->allocate(); + output_gate_out = &_output_layer_norm_out2; + } _activation_output.configure(output_gate_out, nullptr, ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::LOGISTIC)); // Configure block that calculates the output state @@ -362,6 +425,31 @@ Status NELSTMLayer::validate(const ITensorInfo *input, const unsigned int num_batches = input->dimension(1); const unsigned int num_cells = input_to_output_weights->dimension(1); + if(lstm_params.use_layer_norm()) + { + // If CIFG is used, input layer normalization weights tensor is omitted + if(lstm_params.has_cifg_opt()) + { + ARM_COMPUTE_RETURN_ERROR_ON(lstm_params.input_layer_norm_weights() != nullptr); + } + else + { + ARM_COMPUTE_RETURN_ERROR_ON_NULLPTR(lstm_params.input_layer_norm_weights()); + ARM_COMPUTE_RETURN_ERROR_ON(lstm_params.input_layer_norm_weights()->num_dimensions() > 1); + ARM_COMPUTE_RETURN_ERROR_ON(lstm_params.input_layer_norm_weights()->dimension(0) != num_batches); + ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES(input, lstm_params.input_layer_norm_weights()); + } + + ARM_COMPUTE_RETURN_ERROR_ON_NULLPTR(lstm_params.forget_layer_norm_weights(), lstm_params.cell_layer_norm_weights(), lstm_params.output_layer_norm_weights()); + ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES(input, lstm_params.forget_layer_norm_weights(), lstm_params.cell_layer_norm_weights(), lstm_params.output_layer_norm_weights()); + ARM_COMPUTE_RETURN_ERROR_ON(lstm_params.forget_layer_norm_weights()->num_dimensions() > 1); + ARM_COMPUTE_RETURN_ERROR_ON(lstm_params.cell_layer_norm_weights()->num_dimensions() > 1); + ARM_COMPUTE_RETURN_ERROR_ON(lstm_params.output_layer_norm_weights()->num_dimensions() > 1); + ARM_COMPUTE_RETURN_ERROR_ON(lstm_params.forget_layer_norm_weights()->dimension(0) != num_batches); + ARM_COMPUTE_RETURN_ERROR_ON(lstm_params.cell_layer_norm_weights()->dimension(0) != num_batches); + ARM_COMPUTE_RETURN_ERROR_ON(lstm_params.output_layer_norm_weights()->dimension(0) != num_batches); + } + // Check peephole optimization if(lstm_params.has_peephole_opt()) { @@ -388,13 +476,20 @@ Status NELSTMLayer::validate(const ITensorInfo *input, ARM_COMPUTE_RETURN_ON_ERROR(NEConcatenateLayer::validate(inputs_vector, &forget_gate_concat, Window::DimX)); // Validate forget gate - ARM_COMPUTE_RETURN_ON_ERROR(NEFullyConnectedLayer::validate(input, input_to_forget_weights, forget_gate_bias, &forget_gate)); + ARM_COMPUTE_RETURN_ON_ERROR(NEFullyConnectedLayer::validate(input, input_to_forget_weights, (lstm_params.use_layer_norm()) ? nullptr : forget_gate_bias, &forget_gate)); if(lstm_params.has_peephole_opt()) { ARM_COMPUTE_RETURN_ON_ERROR(NEPixelWiseMultiplicationKernel::validate(cell_state_in, lstm_params.cell_to_forget_weights(), &forget_gate, 1, ConvertPolicy::SATURATE, RoundingPolicy::TO_ZERO)); ARM_COMPUTE_RETURN_ON_ERROR(NEArithmeticAddition::validate(&forget_gate, &forget_gate, &forget_gate, ConvertPolicy::SATURATE)); } + if(lstm_params.use_layer_norm()) + { + ARM_COMPUTE_RETURN_ON_ERROR(NEMeanStdDevNormalizationLayer::validate(&forget_gate)); + ARM_COMPUTE_RETURN_ON_ERROR(NEPixelWiseMultiplicationKernel::validate(&forget_gate, lstm_params.forget_layer_norm_weights(), &forget_gate, 1, ConvertPolicy::SATURATE, + RoundingPolicy::TO_ZERO)); + ARM_COMPUTE_RETURN_ON_ERROR(NEArithmeticAddition::validate(&forget_gate, forget_gate_bias, &forget_gate, ConvertPolicy::SATURATE)); + } ARM_COMPUTE_RETURN_ON_ERROR(NEActivationLayerKernel::validate(&forget_gate, &forget_gate, ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::LOGISTIC))); // Validate input gate @@ -413,7 +508,7 @@ Status NELSTMLayer::validate(const ITensorInfo *input, TensorShape lstm_weights_concat_shape = arm_compute::misc::shape_calculator::calculate_concatenate_shape(lstm_weights, 0); TensorInfo lstm_gate_concat = TensorInfo(lstm_weights_concat_shape, 1, input->data_type()); ARM_COMPUTE_RETURN_ON_ERROR(NEConcatenateLayer::validate(lstm_weights, &lstm_gate_concat, Window::DimX)); - ARM_COMPUTE_RETURN_ON_ERROR(NEFullyConnectedLayer::validate(input, lstm_params.input_to_input_weights(), lstm_params.input_gate_bias(), &input_gate)); + ARM_COMPUTE_RETURN_ON_ERROR(NEFullyConnectedLayer::validate(input, lstm_params.input_to_input_weights(), (lstm_params.use_layer_norm()) ? nullptr : lstm_params.input_gate_bias(), &input_gate)); if(lstm_params.has_peephole_opt()) { @@ -422,6 +517,13 @@ Status NELSTMLayer::validate(const ITensorInfo *input, ARM_COMPUTE_RETURN_ON_ERROR(NEPixelWiseMultiplicationKernel::validate(cell_state_in, lstm_params.cell_to_input_weights(), &input_gate, 1, ConvertPolicy::SATURATE, RoundingPolicy::TO_ZERO)); ARM_COMPUTE_RETURN_ON_ERROR(NEArithmeticAddition::validate(&input_gate, &input_gate, &input_gate, ConvertPolicy::SATURATE)); } + + if(lstm_params.use_layer_norm()) + { + ARM_COMPUTE_RETURN_ON_ERROR(NEMeanStdDevNormalizationLayer::validate(&input_gate)); + ARM_COMPUTE_RETURN_ON_ERROR(NEPixelWiseMultiplicationKernel::validate(&input_gate, lstm_params.input_layer_norm_weights(), &input_gate, 1, ConvertPolicy::SATURATE, RoundingPolicy::TO_ZERO)); + ARM_COMPUTE_RETURN_ON_ERROR(NEArithmeticAddition::validate(&input_gate, lstm_params.input_gate_bias(), &input_gate, ConvertPolicy::SATURATE)); + } ARM_COMPUTE_RETURN_ON_ERROR(NEActivationLayerKernel::validate(&input_gate, nullptr, ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::LOGISTIC))); } else @@ -430,9 +532,16 @@ Status NELSTMLayer::validate(const ITensorInfo *input, } // Validate cell state - ARM_COMPUTE_RETURN_ON_ERROR(NEFullyConnectedLayer::validate(input, input_to_cell_weights, cell_bias, &cell_state_tmp)); + ARM_COMPUTE_RETURN_ON_ERROR(NEFullyConnectedLayer::validate(input, input_to_cell_weights, (lstm_params.use_layer_norm()) ? nullptr : cell_bias, &cell_state_tmp)); ARM_COMPUTE_RETURN_ON_ERROR(NEGEMM::validate(output_state_in, &units_out_transposed_info, nullptr, &cell_state_tmp, 1.f, 0.f, GEMMInfo())); ARM_COMPUTE_RETURN_ON_ERROR(NEArithmeticAddition::validate(&cell_state_tmp, &cell_state_tmp, &cell_state_tmp, ConvertPolicy::SATURATE)); + if(lstm_params.use_layer_norm()) + { + ARM_COMPUTE_RETURN_ON_ERROR(NEMeanStdDevNormalizationLayer::validate(&cell_state_tmp)); + ARM_COMPUTE_RETURN_ON_ERROR(NEPixelWiseMultiplicationKernel::validate(&cell_state_tmp, lstm_params.cell_layer_norm_weights(), &cell_state_tmp, 1, ConvertPolicy::SATURATE, + RoundingPolicy::TO_ZERO)); + ARM_COMPUTE_RETURN_ON_ERROR(NEArithmeticAddition::validate(&cell_state_tmp, cell_bias, &cell_state_tmp, ConvertPolicy::SATURATE)); + } ARM_COMPUTE_RETURN_ON_ERROR(NEActivationLayerKernel::validate(&cell_state_tmp, nullptr, activation_info)); ARM_COMPUTE_RETURN_ON_ERROR(NEPixelWiseMultiplicationKernel::validate(&cell_state_tmp, &input_gate, &cell_state_tmp, 1, ConvertPolicy::SATURATE, RoundingPolicy::TO_ZERO)); ARM_COMPUTE_RETURN_ON_ERROR(NEPixelWiseMultiplicationKernel::validate(&cell_state_tmp, &forget_gate, &cell_state_tmp, 1, ConvertPolicy::SATURATE, RoundingPolicy::TO_ZERO)); @@ -451,7 +560,7 @@ Status NELSTMLayer::validate(const ITensorInfo *input, TensorInfo in_out_gate_concat = TensorInfo(in_out_weights_concat_shape, 1, input->data_type()); ARM_COMPUTE_RETURN_ON_ERROR(NEConcatenateLayer::validate(in_out_weights, &in_out_gate_concat, Window::DimX)); - ARM_COMPUTE_RETURN_ON_ERROR(NEFullyConnectedLayer::validate(input, input_to_output_weights, output_gate_bias, &output_gate_tmp)); + ARM_COMPUTE_RETURN_ON_ERROR(NEFullyConnectedLayer::validate(input, input_to_output_weights, (lstm_params.use_layer_norm()) ? nullptr : output_gate_bias, &output_gate_tmp)); if(lstm_params.has_peephole_opt()) { @@ -459,6 +568,13 @@ Status NELSTMLayer::validate(const ITensorInfo *input, RoundingPolicy::TO_ZERO)); ARM_COMPUTE_RETURN_ON_ERROR(NEArithmeticAddition::validate(&output_gate_tmp, &output_gate_tmp, &output_gate_tmp, ConvertPolicy::SATURATE)); } + if(lstm_params.use_layer_norm()) + { + ARM_COMPUTE_RETURN_ON_ERROR(NEMeanStdDevNormalizationLayer::validate(&output_gate_tmp)); + ARM_COMPUTE_RETURN_ON_ERROR(NEPixelWiseMultiplicationKernel::validate(&output_gate_tmp, lstm_params.output_layer_norm_weights(), &output_gate_tmp, 1, ConvertPolicy::SATURATE, + RoundingPolicy::TO_ZERO)); + ARM_COMPUTE_RETURN_ON_ERROR(NEArithmeticAddition::validate(&output_gate_tmp, output_gate_bias, &output_gate_tmp, ConvertPolicy::SATURATE)); + } ARM_COMPUTE_RETURN_ON_ERROR(NEActivationLayerKernel::validate(&output_gate_tmp, nullptr, ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::LOGISTIC))); // Validate output state @@ -504,7 +620,13 @@ void NELSTMLayer::run() if(_run_peephole_opt) { NEScheduler::get().schedule(&_pixelwise_mul_forget_gate, Window::DimY); - _accum_forget_gate2.run(); + _accum_forget_gate1.run(); + } + if(_is_layer_norm_lstm) + { + _mean_std_norm_forget_gate.run(); + NEScheduler::get().schedule(&_pixelwise_mul_forget_gate_coeff, Window::DimY); + NEScheduler::get().schedule(&_accum_forget_gate_bias, Window::DimY); } NEScheduler::get().schedule(&_activation_forget_gate, Window::DimY); @@ -527,7 +649,14 @@ void NELSTMLayer::run() if(_run_peephole_opt) { NEScheduler::get().schedule(&_pixelwise_mul_input_gate, Window::DimY); - _accum_input_gate2.run(); + _accum_input_gate1.run(); + } + + if(_is_layer_norm_lstm) + { + _mean_std_norm_input_gate.run(); + NEScheduler::get().schedule(&_pixelwise_mul_input_gate_coeff, Window::DimY); + NEScheduler::get().schedule(&_accum_input_gate_bias, Window::DimY); } NEScheduler::get().schedule(&_activation_input_gate, Window::DimY); } @@ -536,6 +665,12 @@ void NELSTMLayer::run() NEScheduler::get().schedule(&_transpose_cell_state, Window::DimY); _gemm_cell_state1.run(); NEScheduler::get().schedule(&_accum_cell_state1, Window::DimY); + if(_is_layer_norm_lstm) + { + _mean_std_norm_cell_gate.run(); + NEScheduler::get().schedule(&_pixelwise_mul_cell_gate_coeff, Window::DimY); + NEScheduler::get().schedule(&_accum_cell_gate_bias, Window::DimY); + } NEScheduler::get().schedule(&_activation_cell_state, Window::DimY); NEScheduler::get().schedule(&_pixelwise_mul_cell_state1, Window::DimY); NEScheduler::get().schedule(&_pixelwise_mul_cell_state2, Window::DimY); @@ -550,7 +685,13 @@ void NELSTMLayer::run() if(_run_peephole_opt) { NEScheduler::get().schedule(&_pixelwise_mul_output_state1, Window::DimY); - _accum_output2.run(); + _accum_output1.run(); + } + if(_is_layer_norm_lstm) + { + _mean_std_norm_output_gate.run(); + NEScheduler::get().schedule(&_pixelwise_mul_output_gate_coeff, Window::DimY); + NEScheduler::get().schedule(&_accum_output_gate_bias, Window::DimY); } NEScheduler::get().schedule(&_activation_output, Window::DimY); -- cgit v1.2.1