diff options
Diffstat (limited to 'src/runtime/CL/functions/CLLSTMLayer.cpp')
| -rw-r--r-- | src/runtime/CL/functions/CLLSTMLayer.cpp | 675 |
1 files changed, 440 insertions, 235 deletions
diff --git a/src/runtime/CL/functions/CLLSTMLayer.cpp b/src/runtime/CL/functions/CLLSTMLayer.cpp index 56f22e2fe0..3b50234c77 100644 --- a/src/runtime/CL/functions/CLLSTMLayer.cpp +++ b/src/runtime/CL/functions/CLLSTMLayer.cpp @@ -1,5 +1,5 @@ /* - * Copyright (c) 2018-2020 ARM Limited. + * Copyright (c) 2018-2021 Arm Limited. * * SPDX-License-Identifier: MIT * @@ -24,61 +24,172 @@ #include "arm_compute/runtime/CL/functions/CLLSTMLayer.h" #include "arm_compute/core/Utils.h" -#include "arm_compute/core/Validate.h" #include "arm_compute/core/utils/misc/InfoHelpers.h" #include "arm_compute/core/utils/misc/ShapeCalculator.h" #include "arm_compute/core/utils/quantization/AsymmHelpers.h" +#include "arm_compute/core/Validate.h" #include "arm_compute/runtime/CL/CLScheduler.h" +#include "src/common/utils/Log.h" +#include "src/core/CL/kernels/CLFillBorderKernel.h" +#include "src/gpu/cl/kernels/ClTransposeKernel.h" + namespace arm_compute { using namespace arm_compute::misc::shape_calculator; using namespace arm_compute::utils::info_helpers; CLLSTMLayer::CLLSTMLayer(std::shared_ptr<IMemoryManager> memory_manager) - : _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(), - _ones_memset_kernel(), _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) + : _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(std::make_unique<opencl::kernels::ClTransposeKernel>()), + _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(), + _ones_fill(), + _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) { } -void CLLSTMLayer::configure(const ICLTensor *input, - const ICLTensor *input_to_forget_weights, const ICLTensor *input_to_cell_weights, const ICLTensor *input_to_output_weights, - const ICLTensor *recurrent_to_forget_weights, const ICLTensor *recurrent_to_cell_weights, const ICLTensor *recurrent_to_output_weights, - const ICLTensor *forget_gate_bias, const ICLTensor *cell_bias, const ICLTensor *output_gate_bias, - const ICLTensor *output_state_in, const ICLTensor *cell_state_in, - ICLTensor *scratch_buffer, ICLTensor *output_state_out, ICLTensor *cell_state_out, ICLTensor *output, - const LSTMParams<ICLTensor> &lstm_params, const ActivationLayerInfo &activation_info, float cell_threshold, float projection_threshold) +CLLSTMLayer::~CLLSTMLayer() = default; + +void CLLSTMLayer::configure(const ICLTensor *input, + const ICLTensor *input_to_forget_weights, + const ICLTensor *input_to_cell_weights, + const ICLTensor *input_to_output_weights, + const ICLTensor *recurrent_to_forget_weights, + const ICLTensor *recurrent_to_cell_weights, + const ICLTensor *recurrent_to_output_weights, + const ICLTensor *forget_gate_bias, + const ICLTensor *cell_bias, + const ICLTensor *output_gate_bias, + const ICLTensor *output_state_in, + ICLTensor *cell_state_in, + ICLTensor *scratch_buffer, + ICLTensor *output_state_out, + ICLTensor *cell_state_out, + ICLTensor *output, + const LSTMParams<ICLTensor> &lstm_params, + const ActivationLayerInfo &activation_info, + float cell_threshold, + float projection_threshold) { - configure(CLKernelLibrary::get().get_compile_context(), input, input_to_forget_weights, input_to_cell_weights, input_to_output_weights, recurrent_to_forget_weights, recurrent_to_cell_weights, - recurrent_to_output_weights, forget_gate_bias, cell_bias, output_gate_bias, output_state_in, cell_state_in, scratch_buffer, output_state_out, cell_state_out, output, lstm_params, activation_info, + configure(CLKernelLibrary::get().get_compile_context(), input, input_to_forget_weights, input_to_cell_weights, + input_to_output_weights, recurrent_to_forget_weights, recurrent_to_cell_weights, + recurrent_to_output_weights, forget_gate_bias, cell_bias, output_gate_bias, output_state_in, + cell_state_in, scratch_buffer, output_state_out, cell_state_out, output, lstm_params, activation_info, cell_threshold, projection_threshold); } -void CLLSTMLayer::configure(const CLCompileContext &compile_context, const ICLTensor *input, - const ICLTensor *input_to_forget_weights, const ICLTensor *input_to_cell_weights, const ICLTensor *input_to_output_weights, - const ICLTensor *recurrent_to_forget_weights, const ICLTensor *recurrent_to_cell_weights, const ICLTensor *recurrent_to_output_weights, - const ICLTensor *forget_gate_bias, const ICLTensor *cell_bias, const ICLTensor *output_gate_bias, - const ICLTensor *output_state_in, const ICLTensor *cell_state_in, - ICLTensor *scratch_buffer, ICLTensor *output_state_out, ICLTensor *cell_state_out, ICLTensor *output, - const LSTMParams<ICLTensor> &lstm_params, const ActivationLayerInfo &activation_info, float cell_threshold, float projection_threshold) +void CLLSTMLayer::configure(const CLCompileContext &compile_context, + const ICLTensor *input, + const ICLTensor *input_to_forget_weights, + const ICLTensor *input_to_cell_weights, + const ICLTensor *input_to_output_weights, + const ICLTensor *recurrent_to_forget_weights, + const ICLTensor *recurrent_to_cell_weights, + const ICLTensor *recurrent_to_output_weights, + const ICLTensor *forget_gate_bias, + const ICLTensor *cell_bias, + const ICLTensor *output_gate_bias, + const ICLTensor *output_state_in, + ICLTensor *cell_state_in, + ICLTensor *scratch_buffer, + ICLTensor *output_state_out, + ICLTensor *cell_state_out, + ICLTensor *output, + const LSTMParams<ICLTensor> &lstm_params, + const ActivationLayerInfo &activation_info, + float cell_threshold, + float projection_threshold) { - ARM_COMPUTE_ERROR_ON_NULLPTR(input, - input_to_forget_weights, input_to_cell_weights, input_to_output_weights, + ARM_COMPUTE_ERROR_ON_NULLPTR(input, input_to_forget_weights, input_to_cell_weights, input_to_output_weights, recurrent_to_forget_weights, recurrent_to_cell_weights, recurrent_to_output_weights, - forget_gate_bias, cell_bias, output_gate_bias, - output_state_in, cell_state_in, + forget_gate_bias, cell_bias, output_gate_bias, output_state_in, cell_state_in, scratch_buffer, output_state_out, cell_state_out, output); + ARM_COMPUTE_LOG_PARAMS(input, input_to_forget_weights, input_to_cell_weights, input_to_output_weights, + recurrent_to_forget_weights, recurrent_to_cell_weights, recurrent_to_output_weights, + forget_gate_bias, cell_bias, output_gate_bias, output_state_in, cell_state_in, + scratch_buffer, output_state_out, cell_state_out, output, lstm_params, activation_info, + cell_threshold, projection_threshold); + _is_layer_norm_lstm = lstm_params.use_layer_norm(); // Set lstm parameters @@ -86,13 +197,12 @@ void CLLSTMLayer::configure(const CLCompileContext &compile_context, const ICLTe build_lstm_params_tensor_info(lstm_params, &lstm_params_info); // Validate - ARM_COMPUTE_ERROR_THROW_ON(CLLSTMLayer::validate(input->info(), input_to_forget_weights->info(), - input_to_cell_weights->info(), input_to_output_weights->info(), - recurrent_to_forget_weights->info(), recurrent_to_cell_weights->info(), recurrent_to_output_weights->info(), - forget_gate_bias->info(), cell_bias->info(), output_gate_bias->info(), - output_state_in->info(), cell_state_in->info(), - scratch_buffer->info(), output_state_out->info(), cell_state_out->info(), output->info(), - lstm_params_info, activation_info, cell_threshold, projection_threshold)); + ARM_COMPUTE_ERROR_THROW_ON(CLLSTMLayer::validate( + input->info(), input_to_forget_weights->info(), input_to_cell_weights->info(), input_to_output_weights->info(), + recurrent_to_forget_weights->info(), recurrent_to_cell_weights->info(), recurrent_to_output_weights->info(), + forget_gate_bias->info(), cell_bias->info(), output_gate_bias->info(), output_state_in->info(), + cell_state_in->info(), scratch_buffer->info(), output_state_out->info(), cell_state_out->info(), output->info(), + lstm_params_info, activation_info, cell_threshold, projection_threshold)); const TensorShape cell_state_shape = cell_state_in->info()->tensor_shape(); // Configure block that calculates the forget gate @@ -110,32 +220,37 @@ void CLLSTMLayer::configure(const CLCompileContext &compile_context, const ICLTe _forget_gate_out2.allocator()->init(TensorInfo(concat_shape, 1, input->info()->data_type())); _memory_group.manage(&_forget_gate_out2); - _concat_inputs_forget_gate.configure(compile_context, input, output_state_in, &_forget_gate_out2); + _concat_inputs_forget_gate.configure(compile_context, inputs_vector, &_forget_gate_out2, Window::DimX); std::vector<const ICLTensor *> weights_vector; weights_vector.emplace_back(input_to_forget_weights); weights_vector.emplace_back(recurrent_to_forget_weights); - const TensorShape weights_concat_shape = arm_compute::misc::shape_calculator::calculate_concatenate_shape(weights_vector, 0); + const TensorShape weights_concat_shape = + arm_compute::misc::shape_calculator::calculate_concatenate_shape(weights_vector, 0); _forget_gate_out6.allocator()->init(TensorInfo(weights_concat_shape, 1, input->info()->data_type())); - _concat_weights_forget_gate.configure(compile_context, input_to_forget_weights, recurrent_to_forget_weights, &_forget_gate_out6); + _concat_weights_forget_gate.configure(compile_context, weights_vector, &_forget_gate_out6, Window::DimX); _memory_group.manage(&_forget_gate_out5); - _fully_connected_forget_gate.configure(compile_context, &_forget_gate_out2, &_forget_gate_out6, (_is_layer_norm_lstm) ? nullptr : forget_gate_bias, &_forget_gate_out5); + _fully_connected_forget_gate.configure(compile_context, &_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(); CLTensor *forget_gate_out = &_forget_gate_out5; - if(lstm_params.has_peephole_opt()) + if (lstm_params.has_peephole_opt()) { _forget_gate_out4.allocator()->init(TensorInfo(cell_state_shape, 1, input->info()->data_type())); _run_peephole_opt = true; _memory_group.manage(&_forget_gate_out4); - _pixelwise_mul_forget_gate.configure(compile_context, cell_state_in, lstm_params.cell_to_forget_weights(), &_forget_gate_out4, 1, ConvertPolicy::SATURATE, RoundingPolicy::TO_NEAREST_EVEN); - _accum_forget_gate1.configure(compile_context, &_forget_gate_out5, &_forget_gate_out4, &_forget_gate_out3, ConvertPolicy::SATURATE); + _pixelwise_mul_forget_gate.configure(compile_context, cell_state_in, lstm_params.cell_to_forget_weights(), + &_forget_gate_out4, 1, ConvertPolicy::SATURATE, + RoundingPolicy::TO_NEAREST_EVEN); + _accum_forget_gate1.configure(compile_context, &_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; @@ -144,22 +259,25 @@ void CLLSTMLayer::configure(const CLCompileContext &compile_context, const ICLTe { _forget_gate_out3.allocator()->allocate(); } - if(_is_layer_norm_lstm) + 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(compile_context, forget_gate_out); - _pixelwise_mul_forget_gate_coeff.configure(compile_context, forget_gate_out, lstm_params.forget_layer_norm_weights(), &_forget_layer_norm_out1, 1, ConvertPolicy::SATURATE, - RoundingPolicy::TO_NEAREST_EVEN); + _pixelwise_mul_forget_gate_coeff.configure(compile_context, forget_gate_out, + lstm_params.forget_layer_norm_weights(), &_forget_layer_norm_out1, 1, + ConvertPolicy::SATURATE, RoundingPolicy::TO_NEAREST_EVEN); // 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(compile_context, ArithmeticOperation::ADD, &_forget_layer_norm_out1, forget_gate_bias, &_forget_layer_norm_out2, ConvertPolicy::SATURATE); + _accum_forget_gate_bias.configure(compile_context, &_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(compile_context, forget_gate_out, nullptr, ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::LOGISTIC)); + _activation_forget_gate.configure(compile_context, forget_gate_out, nullptr, + ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::LOGISTIC)); // Configure block that calculates the input gate // input_gate = Activation(input * input_to_input_weights + output_state * recurrent_to_input_weights + PixelWiseMul(cell_state, cell_to_input_weights) + input_gate_bias), without CIFG @@ -168,12 +286,13 @@ void CLLSTMLayer::configure(const CLCompileContext &compile_context, const ICLTe // input_gate = Activation((input,output_state) * (input_to_input_weights,recurrent_to_input_weights) + PixelWiseMul(cell_state, cell_to_input_weights) + input_gate_bias), without CIFG _input_gate_out1.allocator()->init(TensorInfo(cell_state_shape, 1, input->info()->data_type())); CLTensor *input_gate_out = &_input_gate_out1; - if(lstm_params.has_cifg_opt()) + if (lstm_params.has_cifg_opt()) { _memory_group.manage(&_input_gate_out1); _ones.allocator()->init(TensorInfo(cell_state_shape, 1, input->info()->data_type())); - _ones_memset_kernel.configure(compile_context, &_ones, PixelValue(1, _ones.info()->data_type())); - _subtract_input_gate.configure(compile_context, ArithmeticOperation::SUB, &_ones, forget_gate_out, &_input_gate_out1, ConvertPolicy::SATURATE); + _ones_fill.configure(compile_context, &_ones, PixelValue(1, _ones.info()->data_type())); + _subtract_input_gate.configure(compile_context, &_ones, forget_gate_out, &_input_gate_out1, + ConvertPolicy::SATURATE); _ones.allocator()->allocate(); _run_cifg_opt = true; } @@ -185,23 +304,29 @@ void CLLSTMLayer::configure(const CLCompileContext &compile_context, const ICLTe std::vector<const ICLTensor *> lstm_weights; lstm_weights.emplace_back(lstm_params.input_to_input_weights()); lstm_weights.emplace_back(lstm_params.recurrent_to_input_weights()); - TensorShape lstm_weights_concat_shape = arm_compute::misc::shape_calculator::calculate_concatenate_shape(lstm_weights, 0); + TensorShape lstm_weights_concat_shape = + arm_compute::misc::shape_calculator::calculate_concatenate_shape(lstm_weights, 0); _input_gate_out2.allocator()->init(TensorInfo(lstm_weights_concat_shape, 1, input->info()->data_type())); - _concat_weights_input_gate.configure(compile_context, lstm_params.input_to_input_weights(), lstm_params.recurrent_to_input_weights(), &_input_gate_out2); + _concat_weights_input_gate.configure(compile_context, lstm_weights, &_input_gate_out2, Window::DimX); _memory_group.manage(&_input_gate_out1); _memory_group.manage(&_input_gate_out3); - _fully_connected_input_gate.configure(compile_context, &_forget_gate_out2, &_input_gate_out2, (_is_layer_norm_lstm) ? nullptr : lstm_params.input_gate_bias(), &_input_gate_out3); + _fully_connected_input_gate.configure(compile_context, &_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; - if(_run_peephole_opt) + if (_run_peephole_opt) { _memory_group.manage(&_input_gate_out4); - _pixelwise_mul_input_gate.configure(compile_context, cell_state_in, lstm_params.cell_to_input_weights(), &_input_gate_out4, 1, ConvertPolicy::SATURATE, RoundingPolicy::TO_NEAREST_EVEN); - _accum_input_gate1.configure(compile_context, &_input_gate_out3, &_input_gate_out4, &_input_gate_out1, ConvertPolicy::SATURATE); + _pixelwise_mul_input_gate.configure(compile_context, cell_state_in, lstm_params.cell_to_input_weights(), + &_input_gate_out4, 1, ConvertPolicy::SATURATE, + RoundingPolicy::TO_NEAREST_EVEN); + _accum_input_gate1.configure(compile_context, &_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; @@ -211,22 +336,25 @@ void CLLSTMLayer::configure(const CLCompileContext &compile_context, const ICLTe _input_gate_out1.allocator()->allocate(); } - if(_is_layer_norm_lstm) + 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(compile_context, input_gate_out); - _pixelwise_mul_input_gate_coeff.configure(compile_context, input_gate_out, lstm_params.input_layer_norm_weights(), &_input_layer_norm_out1, 1, ConvertPolicy::SATURATE, - RoundingPolicy::TO_NEAREST_EVEN); + _pixelwise_mul_input_gate_coeff.configure(compile_context, input_gate_out, + lstm_params.input_layer_norm_weights(), &_input_layer_norm_out1, + 1, ConvertPolicy::SATURATE, RoundingPolicy::TO_NEAREST_EVEN); // 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(compile_context, ArithmeticOperation::ADD, &_input_layer_norm_out1, lstm_params.input_gate_bias(), &_input_layer_norm_out2, ConvertPolicy::SATURATE); + _accum_input_gate_bias.configure(compile_context, &_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(compile_context, input_gate_out, nullptr, ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::LOGISTIC)); + _activation_input_gate.configure(compile_context, input_gate_out, nullptr, + ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::LOGISTIC)); } // Configure block that calculates the cell state @@ -239,43 +367,54 @@ void CLLSTMLayer::configure(const CLCompileContext &compile_context, const ICLTe _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(compile_context, input, input_to_cell_weights, (_is_layer_norm_lstm) ? nullptr : cell_bias, &_cell_state_out1); + _fully_connected_cell_state.configure(compile_context, 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(compile_context, recurrent_to_cell_weights, &_cell_state_out2); + _transpose_cell_state->configure(compile_context, recurrent_to_cell_weights->info(), _cell_state_out2.info()); + _recurrent_to_cell_weights = recurrent_to_cell_weights; _memory_group.manage(&_cell_state_out3); - _gemm_cell_state1.configure(compile_context, output_state_in, &_cell_state_out2, nullptr, &_cell_state_out3, 1.f, 0.f); + _gemm_cell_state1.configure(compile_context, output_state_in, &_cell_state_out2, nullptr, &_cell_state_out3, 1.f, + 0.f); _cell_state_out2.allocator()->allocate(); _memory_group.manage(&_cell_state_out4); - _accum_cell_state1.configure(compile_context, ArithmeticOperation::ADD, &_cell_state_out1, &_cell_state_out3, &_cell_state_out4, ConvertPolicy::SATURATE); + _accum_cell_state1.configure(compile_context, &_cell_state_out1, &_cell_state_out3, &_cell_state_out4, + ConvertPolicy::SATURATE); CLTensor *cell_state_out_ptr = &_cell_state_out4; - if(_is_layer_norm_lstm) + 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(compile_context, cell_state_out_ptr); - _pixelwise_mul_cell_gate_coeff.configure(compile_context, cell_state_out_ptr, lstm_params.cell_layer_norm_weights(), &_cell_layer_norm_out1, 1, ConvertPolicy::SATURATE, - RoundingPolicy::TO_NEAREST_EVEN); + _pixelwise_mul_cell_gate_coeff.configure(compile_context, cell_state_out_ptr, + lstm_params.cell_layer_norm_weights(), &_cell_layer_norm_out1, 1, + ConvertPolicy::SATURATE, RoundingPolicy::TO_NEAREST_EVEN); // 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(compile_context, ArithmeticOperation::ADD, &_cell_layer_norm_out1, cell_bias, &_cell_layer_norm_out2, ConvertPolicy::SATURATE); + _accum_cell_gate_bias.configure(compile_context, &_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(compile_context, cell_state_out_ptr, nullptr, activation_info); _memory_group.manage(&_cell_state_out5); - _pixelwise_mul_cell_state1.configure(compile_context, cell_state_out_ptr, input_gate_out, &_cell_state_out5, 1, ConvertPolicy::SATURATE, RoundingPolicy::TO_NEAREST_EVEN); + _pixelwise_mul_cell_state1.configure(compile_context, cell_state_out_ptr, input_gate_out, &_cell_state_out5, 1, + ConvertPolicy::SATURATE, RoundingPolicy::TO_NEAREST_EVEN); cell_state_out_ptr->allocator()->allocate(); - _pixelwise_mul_cell_state2.configure(compile_context, forget_gate_out, cell_state_in, &_cell_state_out3, 1, ConvertPolicy::SATURATE, RoundingPolicy::TO_NEAREST_EVEN); - _accum_cell_state2.configure(compile_context, ArithmeticOperation::ADD, &_cell_state_out5, &_cell_state_out3, &_cell_state_out1, ConvertPolicy::SATURATE); + _pixelwise_mul_cell_state2.configure(compile_context, forget_gate_out, cell_state_in, &_cell_state_out3, 1, + ConvertPolicy::SATURATE, RoundingPolicy::TO_NEAREST_EVEN); + _accum_cell_state2.configure(compile_context, &_cell_state_out5, &_cell_state_out3, &_cell_state_out1, + ConvertPolicy::SATURATE); _cell_state_out3.allocator()->allocate(); _cell_state_out5.allocator()->allocate(); // Perform clipping - if(cell_threshold != 0.f) + if (cell_threshold != 0.f) { _perform_cell_clipping = true; - _cell_clip.configure(compile_context, &_cell_state_out1, nullptr, ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::LU_BOUNDED_RELU, -cell_threshold, cell_threshold)); + _cell_clip.configure(compile_context, &_cell_state_out1, nullptr, + ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::LU_BOUNDED_RELU, + cell_threshold, -cell_threshold)); } // Configure block that calculates the output @@ -287,26 +426,29 @@ void CLLSTMLayer::configure(const CLCompileContext &compile_context, const ICLTe std::vector<const ICLTensor *> in_out_weights; in_out_weights.emplace_back(input_to_output_weights); in_out_weights.emplace_back(recurrent_to_output_weights); - TensorShape in_out_weights_concat_shape = arm_compute::misc::shape_calculator::calculate_concatenate_shape(in_out_weights, 0); + TensorShape in_out_weights_concat_shape = + arm_compute::misc::shape_calculator::calculate_concatenate_shape(in_out_weights, 0); _output2.allocator()->init(TensorInfo(in_out_weights_concat_shape, 1, input->info()->data_type())); - _concat_weights_output.configure(compile_context, input_to_output_weights, recurrent_to_output_weights, &_output2); + _concat_weights_output.configure(compile_context, in_out_weights, &_output2, Window::DimX); _memory_group.manage(&_output1); _memory_group.manage(&_output4); - _fully_connected_output.configure(compile_context, &_forget_gate_out2, &_output2, (_is_layer_norm_lstm) ? nullptr : output_gate_bias, &_output4); + _fully_connected_output.configure(compile_context, &_forget_gate_out2, &_output2, + (_is_layer_norm_lstm) ? nullptr : output_gate_bias, &_output4); _output2.allocator()->allocate(); _forget_gate_out2.allocator()->allocate(); CLTensor *output_gate_out = &_output4; - if(lstm_params.has_peephole_opt()) + if (lstm_params.has_peephole_opt()) { _output3.allocator()->init(TensorInfo(_cell_state_out1.info()->tensor_shape(), 1, input->info()->data_type())); _memory_group.manage(&_output3); - _pixelwise_mul_output_state1.configure(compile_context, &_cell_state_out1, lstm_params.cell_to_output_weights(), &_output3, 1, ConvertPolicy::SATURATE, RoundingPolicy::TO_NEAREST_EVEN); + _pixelwise_mul_output_state1.configure(compile_context, &_cell_state_out1, lstm_params.cell_to_output_weights(), + &_output3, 1, ConvertPolicy::SATURATE, RoundingPolicy::TO_NEAREST_EVEN); _accum_output1.configure(compile_context, &_output4, &_output3, &_output1, ConvertPolicy::SATURATE); _output4.allocator()->allocate(); output_gate_out = &_output1; @@ -318,22 +460,25 @@ void CLLSTMLayer::configure(const CLCompileContext &compile_context, const ICLTe { _output1.allocator()->allocate(); } - if(_is_layer_norm_lstm) + 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(compile_context, output_gate_out); - _pixelwise_mul_output_gate_coeff.configure(compile_context, output_gate_out, lstm_params.output_layer_norm_weights(), &_output_layer_norm_out1, 1, ConvertPolicy::SATURATE, - RoundingPolicy::TO_NEAREST_EVEN); + _pixelwise_mul_output_gate_coeff.configure(compile_context, output_gate_out, + lstm_params.output_layer_norm_weights(), &_output_layer_norm_out1, 1, + ConvertPolicy::SATURATE, RoundingPolicy::TO_NEAREST_EVEN); // 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(compile_context, ArithmeticOperation::ADD, &_output_layer_norm_out1, output_gate_bias, &_output_layer_norm_out2, ConvertPolicy::SATURATE); + _accum_output_gate_bias.configure(compile_context, &_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(compile_context, output_gate_out, nullptr, ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::LOGISTIC)); + _activation_output.configure(compile_context, output_gate_out, nullptr, + ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::LOGISTIC)); // Configure block that calculates the output state /** lstm_res = PixelwiseMul(output, Activation(cell_state)) @@ -350,19 +495,24 @@ void CLLSTMLayer::configure(const CLCompileContext &compile_context, const ICLTe _memory_group.manage(&_cell_state_activation); _activation_output_state.configure(compile_context, &_cell_state_out1, &_cell_state_activation, activation_info); - _pixelwise_mul_output_state2.configure(compile_context, &_cell_state_activation, output_gate_out, output_state_out_tmp, 1, ConvertPolicy::SATURATE, RoundingPolicy::TO_NEAREST_EVEN); + _pixelwise_mul_output_state2.configure(compile_context, &_cell_state_activation, output_gate_out, + output_state_out_tmp, 1, ConvertPolicy::SATURATE, + RoundingPolicy::TO_NEAREST_EVEN); _cell_state_activation.allocator()->allocate(); - if(lstm_params.has_projection()) + if (lstm_params.has_projection()) { _has_projection_weights = true; - _fully_connected_output_state.configure(compile_context, output_state_out_tmp, lstm_params.projection_weights(), lstm_params.projection_bias(), output_state_out); + _fully_connected_output_state.configure(compile_context, output_state_out_tmp, lstm_params.projection_weights(), + lstm_params.projection_bias(), output_state_out); _output_state1.allocator()->allocate(); // Perform clipping - if(projection_threshold != 0.f) + if (projection_threshold != 0.f) { _perform_projection_clipping = true; - _projection_clip.configure(compile_context, output_state_out, nullptr, ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::LU_BOUNDED_RELU, -projection_threshold, projection_threshold)); + _projection_clip.configure(compile_context, output_state_out, nullptr, + ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::LU_BOUNDED_RELU, + -projection_threshold, projection_threshold)); } } @@ -371,8 +521,8 @@ void CLLSTMLayer::configure(const CLCompileContext &compile_context, const ICLTe _copy_output.configure(compile_context, output_state_out, output); // Vector for holding the tensors to store in scratch buffer - std::vector<ICLTensor *> scratch_inputs; - if(!lstm_params.has_cifg_opt()) + std::vector<const ICLTensor *> scratch_inputs; + if (!lstm_params.has_cifg_opt()) { scratch_inputs.emplace_back(input_gate_out); } @@ -386,29 +536,38 @@ void CLLSTMLayer::configure(const CLCompileContext &compile_context, const ICLTe output_gate_out->allocator()->allocate(); } -Status CLLSTMLayer::validate(const ITensorInfo *input, - const ITensorInfo *input_to_forget_weights, const ITensorInfo *input_to_cell_weights, const ITensorInfo *input_to_output_weights, - const ITensorInfo *recurrent_to_forget_weights, const ITensorInfo *recurrent_to_cell_weights, const ITensorInfo *recurrent_to_output_weights, - const ITensorInfo *forget_gate_bias, const ITensorInfo *cell_bias, const ITensorInfo *output_gate_bias, - const ITensorInfo *output_state_in, const ITensorInfo *cell_state_in, - const ITensorInfo *scratch_buffer, const ITensorInfo *output_state_out, const ITensorInfo *cell_state_out, const ITensorInfo *output, - const LSTMParams<ITensorInfo> &lstm_params, const ActivationLayerInfo &activation_info, float cell_threshold, float projection_threshold) +Status CLLSTMLayer::validate(const ITensorInfo *input, + const ITensorInfo *input_to_forget_weights, + const ITensorInfo *input_to_cell_weights, + const ITensorInfo *input_to_output_weights, + const ITensorInfo *recurrent_to_forget_weights, + const ITensorInfo *recurrent_to_cell_weights, + const ITensorInfo *recurrent_to_output_weights, + const ITensorInfo *forget_gate_bias, + const ITensorInfo *cell_bias, + const ITensorInfo *output_gate_bias, + const ITensorInfo *output_state_in, + const ITensorInfo *cell_state_in, + const ITensorInfo *scratch_buffer, + const ITensorInfo *output_state_out, + const ITensorInfo *cell_state_out, + const ITensorInfo *output, + const LSTMParams<ITensorInfo> &lstm_params, + const ActivationLayerInfo &activation_info, + float cell_threshold, + float projection_threshold) { - ARM_COMPUTE_RETURN_ERROR_ON_NULLPTR(input, - input_to_forget_weights, input_to_cell_weights, input_to_output_weights, - recurrent_to_forget_weights, recurrent_to_cell_weights, recurrent_to_output_weights, - forget_gate_bias, cell_bias, output_gate_bias, - output_state_in, cell_state_in, - scratch_buffer, output_state_out, cell_state_out, output); + ARM_COMPUTE_RETURN_ERROR_ON_NULLPTR( + input, input_to_forget_weights, input_to_cell_weights, input_to_output_weights, recurrent_to_forget_weights, + recurrent_to_cell_weights, recurrent_to_output_weights, forget_gate_bias, cell_bias, output_gate_bias, + output_state_in, cell_state_in, scratch_buffer, output_state_out, cell_state_out, output); // Check data types ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(input, 1, DataType::F16, DataType::F32); - ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES(input, - input_to_forget_weights, input_to_cell_weights, input_to_output_weights, - recurrent_to_forget_weights, recurrent_to_cell_weights, recurrent_to_output_weights, - forget_gate_bias, cell_bias, output_gate_bias, - output_state_in, cell_state_in, - scratch_buffer, output_state_out, cell_state_out, output); + ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES( + input, input_to_forget_weights, input_to_cell_weights, input_to_output_weights, recurrent_to_forget_weights, + recurrent_to_cell_weights, recurrent_to_output_weights, forget_gate_bias, cell_bias, output_gate_bias, + output_state_in, cell_state_in, scratch_buffer, output_state_out, cell_state_out, output); // Check dimensions ARM_COMPUTE_RETURN_ERROR_ON(input->num_dimensions() > 2); @@ -427,16 +586,16 @@ Status CLLSTMLayer::validate(const ITensorInfo *input, ARM_COMPUTE_RETURN_ERROR_ON(output_state_out->num_dimensions() > 2); ARM_COMPUTE_RETURN_ERROR_ON(cell_state_out->num_dimensions() > 2); ARM_COMPUTE_RETURN_ERROR_ON(output->num_dimensions() > 2); - ARM_COMPUTE_RETURN_ERROR_ON(cell_bias->dimension(0) * 4 != scratch_buffer->dimension(0) - && cell_bias->dimension(0) * 3 != scratch_buffer->dimension(0)); + ARM_COMPUTE_RETURN_ERROR_ON(cell_bias->dimension(0) * 4 != scratch_buffer->dimension(0) && + cell_bias->dimension(0) * 3 != scratch_buffer->dimension(0)); 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 (lstm_params.use_layer_norm()) { // If CIFG is used, input layer normalization weights tensor is omitted - if(lstm_params.has_cifg_opt()) + if (lstm_params.has_cifg_opt()) { ARM_COMPUTE_RETURN_ERROR_ON(lstm_params.input_layer_norm_weights() != nullptr); } @@ -448,8 +607,12 @@ Status CLLSTMLayer::validate(const ITensorInfo *input, 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_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); @@ -459,7 +622,7 @@ Status CLLSTMLayer::validate(const ITensorInfo *input, } // Check peephole optimization - if(lstm_params.has_peephole_opt()) + if (lstm_params.has_peephole_opt()) { ARM_COMPUTE_RETURN_ERROR_ON_NULLPTR(lstm_params.cell_to_output_weights(), lstm_params.cell_to_forget_weights()); ARM_COMPUTE_RETURN_ERROR_ON(lstm_params.cell_to_forget_weights()->num_dimensions() > 1); @@ -477,36 +640,42 @@ Status CLLSTMLayer::validate(const ITensorInfo *input, TensorInfo cell_state_tmp = TensorInfo(TensorShape(num_cells, num_batches), 1, input->data_type()); // Validate forget gate - ARM_COMPUTE_RETURN_ON_ERROR(CLFullyConnectedLayer::validate(input, input_to_forget_weights, (lstm_params.use_layer_norm()) ? nullptr : forget_gate_bias, &forget_gate)); + ARM_COMPUTE_RETURN_ON_ERROR(CLFullyConnectedLayer::validate( + input, input_to_forget_weights, (lstm_params.use_layer_norm()) ? nullptr : forget_gate_bias, &forget_gate)); std::vector<const ITensorInfo *> inputs_vector; inputs_vector.emplace_back(input); inputs_vector.emplace_back(output_state_in); - const TensorShape concat_shape = arm_compute::misc::shape_calculator::calculate_concatenate_shape(inputs_vector, 0); + const TensorShape concat_shape = arm_compute::misc::shape_calculator::calculate_concatenate_shape(inputs_vector, 0); TensorInfo forget_gate_concat = TensorInfo(concat_shape, 1, input->data_type()); - ARM_COMPUTE_RETURN_ON_ERROR(CLWidthConcatenate2TensorsKernel::validate(input, output_state_in, &forget_gate_concat)); + ARM_COMPUTE_RETURN_ON_ERROR(CLConcatenateLayer::validate(inputs_vector, &forget_gate_concat, Window::DimX)); - if(lstm_params.has_peephole_opt()) + if (lstm_params.has_peephole_opt()) { - ARM_COMPUTE_RETURN_ON_ERROR(CLPixelWiseMultiplicationKernel::validate(cell_state_in, lstm_params.cell_to_forget_weights(), &forget_gate, 1, ConvertPolicy::SATURATE, RoundingPolicy::TO_NEAREST_EVEN)); - ARM_COMPUTE_RETURN_ON_ERROR(CLArithmeticAddition::validate(&forget_gate, &forget_gate, &forget_gate, ConvertPolicy::SATURATE)); + ARM_COMPUTE_RETURN_ON_ERROR( + CLPixelWiseMultiplication::validate(cell_state_in, lstm_params.cell_to_forget_weights(), &forget_gate, 1, + ConvertPolicy::SATURATE, RoundingPolicy::TO_NEAREST_EVEN)); + ARM_COMPUTE_RETURN_ON_ERROR( + CLArithmeticAddition::validate(&forget_gate, &forget_gate, &forget_gate, ConvertPolicy::SATURATE)); } - if(lstm_params.use_layer_norm()) + if (lstm_params.use_layer_norm()) { ARM_COMPUTE_RETURN_ON_ERROR(CLMeanStdDevNormalizationLayer::validate(&forget_gate)); - ARM_COMPUTE_RETURN_ON_ERROR(CLPixelWiseMultiplicationKernel::validate(&forget_gate, lstm_params.forget_layer_norm_weights(), &forget_gate, 1, ConvertPolicy::SATURATE, - RoundingPolicy::TO_NEAREST_EVEN)); - ARM_COMPUTE_RETURN_ON_ERROR(CLArithmeticAddition::validate(&forget_gate, forget_gate_bias, &forget_gate, ConvertPolicy::SATURATE)); + ARM_COMPUTE_RETURN_ON_ERROR( + CLPixelWiseMultiplication::validate(&forget_gate, lstm_params.forget_layer_norm_weights(), &forget_gate, 1, + ConvertPolicy::SATURATE, RoundingPolicy::TO_NEAREST_EVEN)); + ARM_COMPUTE_RETURN_ON_ERROR( + CLArithmeticAddition::validate(&forget_gate, forget_gate_bias, &forget_gate, ConvertPolicy::SATURATE)); } - ARM_COMPUTE_RETURN_ON_ERROR(CLActivationLayerKernel::validate(&forget_gate, &forget_gate, ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::LOGISTIC))); + ARM_COMPUTE_RETURN_ON_ERROR(CLActivationLayer::validate( + &forget_gate, &forget_gate, ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::LOGISTIC))); // Validate input gate - if(!lstm_params.has_cifg_opt()) + if (!lstm_params.has_cifg_opt()) { ARM_COMPUTE_RETURN_ERROR_ON_NULLPTR(lstm_params.input_to_input_weights(), - lstm_params.recurrent_to_input_weights(), - lstm_params.input_gate_bias()); + lstm_params.recurrent_to_input_weights(), lstm_params.input_gate_bias()); ARM_COMPUTE_RETURN_ERROR_ON(lstm_params.input_to_input_weights()->num_dimensions() > 2); ARM_COMPUTE_RETURN_ERROR_ON(lstm_params.recurrent_to_input_weights()->num_dimensions() > 2); ARM_COMPUTE_RETURN_ERROR_ON(lstm_params.input_gate_bias()->num_dimensions() > 1); @@ -514,98 +683,131 @@ Status CLLSTMLayer::validate(const ITensorInfo *input, std::vector<const ITensorInfo *> lstm_weights; lstm_weights.emplace_back(lstm_params.input_to_input_weights()); lstm_weights.emplace_back(lstm_params.recurrent_to_input_weights()); - 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(CLWidthConcatenate2TensorsKernel::validate(lstm_params.input_to_input_weights(), lstm_params.recurrent_to_input_weights(), &lstm_gate_concat)); + 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(CLConcatenateLayer::validate(lstm_weights, &lstm_gate_concat, Window::DimX)); - ARM_COMPUTE_RETURN_ON_ERROR(CLFullyConnectedLayer::validate(input, lstm_params.input_to_input_weights(), (lstm_params.use_layer_norm()) ? nullptr : lstm_params.input_gate_bias(), &input_gate)); + ARM_COMPUTE_RETURN_ON_ERROR(CLFullyConnectedLayer::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()) + if (lstm_params.has_peephole_opt()) { ARM_COMPUTE_RETURN_ERROR_ON_NULLPTR(lstm_params.cell_to_input_weights()); ARM_COMPUTE_RETURN_ERROR_ON(lstm_params.cell_to_input_weights()->num_dimensions() > 1); - ARM_COMPUTE_RETURN_ON_ERROR(CLPixelWiseMultiplicationKernel::validate(cell_state_in, lstm_params.cell_to_input_weights(), &input_gate, 1, ConvertPolicy::SATURATE, RoundingPolicy::TO_NEAREST_EVEN)); - ARM_COMPUTE_RETURN_ON_ERROR(CLArithmeticAddition::validate(&input_gate, &input_gate, &input_gate, ConvertPolicy::SATURATE)); + ARM_COMPUTE_RETURN_ON_ERROR( + CLPixelWiseMultiplication::validate(cell_state_in, lstm_params.cell_to_input_weights(), &input_gate, 1, + ConvertPolicy::SATURATE, RoundingPolicy::TO_NEAREST_EVEN)); + ARM_COMPUTE_RETURN_ON_ERROR( + CLArithmeticAddition::validate(&input_gate, &input_gate, &input_gate, ConvertPolicy::SATURATE)); } - if(lstm_params.use_layer_norm()) + if (lstm_params.use_layer_norm()) { ARM_COMPUTE_RETURN_ON_ERROR(CLMeanStdDevNormalizationLayer::validate(&input_gate)); - ARM_COMPUTE_RETURN_ON_ERROR(CLPixelWiseMultiplicationKernel::validate(&input_gate, lstm_params.input_layer_norm_weights(), &input_gate, 1, ConvertPolicy::SATURATE, RoundingPolicy::TO_NEAREST_EVEN)); - ARM_COMPUTE_RETURN_ON_ERROR(CLArithmeticAddition::validate(&input_gate, lstm_params.input_gate_bias(), &input_gate, ConvertPolicy::SATURATE)); + ARM_COMPUTE_RETURN_ON_ERROR( + CLPixelWiseMultiplication::validate(&input_gate, lstm_params.input_layer_norm_weights(), &input_gate, 1, + ConvertPolicy::SATURATE, RoundingPolicy::TO_NEAREST_EVEN)); + ARM_COMPUTE_RETURN_ON_ERROR(CLArithmeticAddition::validate(&input_gate, lstm_params.input_gate_bias(), + &input_gate, ConvertPolicy::SATURATE)); } - ARM_COMPUTE_RETURN_ON_ERROR(CLActivationLayerKernel::validate(&input_gate, nullptr, ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::LOGISTIC))); + ARM_COMPUTE_RETURN_ON_ERROR(CLActivationLayer::validate( + &input_gate, nullptr, ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::LOGISTIC))); } else { - ARM_COMPUTE_RETURN_ON_ERROR(CLSaturatedArithmeticOperationKernel::validate(ArithmeticOperation::SUB, &forget_gate, &forget_gate, &forget_gate, ConvertPolicy::SATURATE)); + ARM_COMPUTE_RETURN_ON_ERROR( + CLArithmeticSubtraction::validate(&forget_gate, &forget_gate, &forget_gate, ConvertPolicy::SATURATE)); } // Validate cell state - ARM_COMPUTE_RETURN_ON_ERROR(CLFullyConnectedLayer::validate(input, input_to_cell_weights, (lstm_params.use_layer_norm()) ? nullptr : cell_bias, &cell_state_tmp)); - ARM_COMPUTE_RETURN_ON_ERROR(CLGEMM::validate(output_state_in, &units_out_transposed_info, nullptr, &cell_state_tmp, 1.f, 0.f, GEMMInfo())); - ARM_COMPUTE_RETURN_ON_ERROR(CLArithmeticAddition::validate(&cell_state_tmp, &cell_state_tmp, &cell_state_tmp, ConvertPolicy::SATURATE)); - if(lstm_params.use_layer_norm()) + ARM_COMPUTE_RETURN_ON_ERROR(CLFullyConnectedLayer::validate( + input, input_to_cell_weights, (lstm_params.use_layer_norm()) ? nullptr : cell_bias, &cell_state_tmp)); + ARM_COMPUTE_RETURN_ON_ERROR( + CLGEMM::validate(output_state_in, &units_out_transposed_info, nullptr, &cell_state_tmp, 1.f, 0.f, GEMMInfo())); + ARM_COMPUTE_RETURN_ON_ERROR( + CLArithmeticAddition::validate(&cell_state_tmp, &cell_state_tmp, &cell_state_tmp, ConvertPolicy::SATURATE)); + if (lstm_params.use_layer_norm()) { ARM_COMPUTE_RETURN_ON_ERROR(CLMeanStdDevNormalizationLayer::validate(&cell_state_tmp)); - ARM_COMPUTE_RETURN_ON_ERROR(CLPixelWiseMultiplicationKernel::validate(&cell_state_tmp, lstm_params.cell_layer_norm_weights(), &cell_state_tmp, 1, ConvertPolicy::SATURATE, - RoundingPolicy::TO_NEAREST_EVEN)); - ARM_COMPUTE_RETURN_ON_ERROR(CLArithmeticAddition::validate(&cell_state_tmp, cell_bias, &cell_state_tmp, ConvertPolicy::SATURATE)); - } - ARM_COMPUTE_RETURN_ON_ERROR(CLActivationLayerKernel::validate(&cell_state_tmp, nullptr, activation_info)); - ARM_COMPUTE_RETURN_ON_ERROR(CLPixelWiseMultiplicationKernel::validate(&cell_state_tmp, &input_gate, &cell_state_tmp, 1, ConvertPolicy::SATURATE, RoundingPolicy::TO_NEAREST_EVEN)); - ARM_COMPUTE_RETURN_ON_ERROR(CLPixelWiseMultiplicationKernel::validate(&cell_state_tmp, &forget_gate, &cell_state_tmp, 1, ConvertPolicy::SATURATE, RoundingPolicy::TO_NEAREST_EVEN)); - ARM_COMPUTE_RETURN_ON_ERROR(CLArithmeticAddition::validate(&cell_state_tmp, &cell_state_tmp, &cell_state_tmp, ConvertPolicy::SATURATE)); - if(cell_threshold != 0.f) - { - ARM_COMPUTE_RETURN_ON_ERROR(CLActivationLayerKernel::validate(&cell_state_tmp, nullptr, ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::LU_BOUNDED_RELU, -cell_threshold, - cell_threshold))); + ARM_COMPUTE_RETURN_ON_ERROR( + CLPixelWiseMultiplication::validate(&cell_state_tmp, lstm_params.cell_layer_norm_weights(), &cell_state_tmp, + 1, ConvertPolicy::SATURATE, RoundingPolicy::TO_NEAREST_EVEN)); + ARM_COMPUTE_RETURN_ON_ERROR( + CLArithmeticAddition::validate(&cell_state_tmp, cell_bias, &cell_state_tmp, ConvertPolicy::SATURATE)); + } + ARM_COMPUTE_RETURN_ON_ERROR(CLActivationLayer::validate(&cell_state_tmp, nullptr, activation_info)); + ARM_COMPUTE_RETURN_ON_ERROR(CLPixelWiseMultiplication::validate( + &cell_state_tmp, &input_gate, &cell_state_tmp, 1, ConvertPolicy::SATURATE, RoundingPolicy::TO_NEAREST_EVEN)); + ARM_COMPUTE_RETURN_ON_ERROR(CLPixelWiseMultiplication::validate( + &cell_state_tmp, &forget_gate, &cell_state_tmp, 1, ConvertPolicy::SATURATE, RoundingPolicy::TO_NEAREST_EVEN)); + ARM_COMPUTE_RETURN_ON_ERROR( + CLArithmeticAddition::validate(&cell_state_tmp, &cell_state_tmp, &cell_state_tmp, ConvertPolicy::SATURATE)); + if (cell_threshold != 0.f) + { + ARM_COMPUTE_RETURN_ON_ERROR( + CLActivationLayer::validate(&cell_state_tmp, nullptr, + ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::LU_BOUNDED_RELU, + cell_threshold, -cell_threshold))); } std::vector<const ITensorInfo *> in_out_weights; in_out_weights.emplace_back(input_to_output_weights); in_out_weights.emplace_back(recurrent_to_output_weights); - TensorShape in_out_weights_concat_shape = arm_compute::misc::shape_calculator::calculate_concatenate_shape(in_out_weights, 0); - TensorInfo in_out_gate_concat = TensorInfo(in_out_weights_concat_shape, 1, input->data_type()); - ARM_COMPUTE_RETURN_ON_ERROR(CLWidthConcatenate2TensorsKernel::validate(input_to_output_weights, recurrent_to_output_weights, &in_out_gate_concat)); + TensorShape in_out_weights_concat_shape = + arm_compute::misc::shape_calculator::calculate_concatenate_shape(in_out_weights, 0); + TensorInfo in_out_gate_concat = TensorInfo(in_out_weights_concat_shape, 1, input->data_type()); + ARM_COMPUTE_RETURN_ON_ERROR(CLConcatenateLayer::validate(in_out_weights, &in_out_gate_concat, Window::DimX)); // Validate output gate tmp - ARM_COMPUTE_RETURN_ON_ERROR(CLFullyConnectedLayer::validate(input, input_to_output_weights, (lstm_params.use_layer_norm()) ? nullptr : output_gate_bias, &output_gate_tmp)); + ARM_COMPUTE_RETURN_ON_ERROR(CLFullyConnectedLayer::validate( + input, input_to_output_weights, (lstm_params.use_layer_norm()) ? nullptr : output_gate_bias, &output_gate_tmp)); - if(lstm_params.has_peephole_opt()) + if (lstm_params.has_peephole_opt()) { - ARM_COMPUTE_RETURN_ON_ERROR(CLPixelWiseMultiplicationKernel::validate(&cell_state_tmp, lstm_params.cell_to_output_weights(), &output_gate_tmp, 1, ConvertPolicy::SATURATE, - RoundingPolicy::TO_NEAREST_EVEN)); - ARM_COMPUTE_RETURN_ON_ERROR(CLArithmeticAddition::validate(&output_gate_tmp, &output_gate_tmp, &output_gate_tmp, ConvertPolicy::SATURATE)); + ARM_COMPUTE_RETURN_ON_ERROR( + CLPixelWiseMultiplication::validate(&cell_state_tmp, lstm_params.cell_to_output_weights(), &output_gate_tmp, + 1, ConvertPolicy::SATURATE, RoundingPolicy::TO_NEAREST_EVEN)); + ARM_COMPUTE_RETURN_ON_ERROR(CLArithmeticAddition::validate(&output_gate_tmp, &output_gate_tmp, &output_gate_tmp, + ConvertPolicy::SATURATE)); } - if(lstm_params.use_layer_norm()) + if (lstm_params.use_layer_norm()) { ARM_COMPUTE_RETURN_ON_ERROR(CLMeanStdDevNormalizationLayer::validate(&output_gate_tmp)); - ARM_COMPUTE_RETURN_ON_ERROR(CLPixelWiseMultiplicationKernel::validate(&output_gate_tmp, lstm_params.output_layer_norm_weights(), &output_gate_tmp, 1, ConvertPolicy::SATURATE, - RoundingPolicy::TO_NEAREST_EVEN)); - ARM_COMPUTE_RETURN_ON_ERROR(CLArithmeticAddition::validate(&output_gate_tmp, output_gate_bias, &output_gate_tmp, ConvertPolicy::SATURATE)); + ARM_COMPUTE_RETURN_ON_ERROR(CLPixelWiseMultiplication::validate( + &output_gate_tmp, lstm_params.output_layer_norm_weights(), &output_gate_tmp, 1, ConvertPolicy::SATURATE, + RoundingPolicy::TO_NEAREST_EVEN)); + ARM_COMPUTE_RETURN_ON_ERROR(CLArithmeticAddition::validate(&output_gate_tmp, output_gate_bias, &output_gate_tmp, + ConvertPolicy::SATURATE)); } - ARM_COMPUTE_RETURN_ON_ERROR(CLActivationLayerKernel::validate(&output_gate_tmp, nullptr, ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::LOGISTIC))); + ARM_COMPUTE_RETURN_ON_ERROR(CLActivationLayer::validate( + &output_gate_tmp, nullptr, ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::LOGISTIC))); // Validate output state - ARM_COMPUTE_RETURN_ON_ERROR(CLActivationLayerKernel::validate(&cell_state_tmp, &cell_state_tmp, activation_info)); - ARM_COMPUTE_RETURN_ON_ERROR(CLPixelWiseMultiplicationKernel::validate(&cell_state_tmp, &output_gate_tmp, &output_gate_tmp, 1, ConvertPolicy::SATURATE, RoundingPolicy::TO_NEAREST_EVEN)); - if(lstm_params.has_projection()) - { - ARM_COMPUTE_RETURN_ON_ERROR(CLFullyConnectedLayer::validate(&output_gate_tmp, lstm_params.projection_weights(), lstm_params.projection_bias(), output_state_out)); - if(projection_threshold != 0.f) + ARM_COMPUTE_RETURN_ON_ERROR(CLActivationLayer::validate(&cell_state_tmp, &cell_state_tmp, activation_info)); + ARM_COMPUTE_RETURN_ON_ERROR(CLPixelWiseMultiplication::validate(&cell_state_tmp, &output_gate_tmp, &output_gate_tmp, + 1, ConvertPolicy::SATURATE, + RoundingPolicy::TO_NEAREST_EVEN)); + if (lstm_params.has_projection()) + { + ARM_COMPUTE_RETURN_ON_ERROR(CLFullyConnectedLayer::validate(&output_gate_tmp, lstm_params.projection_weights(), + lstm_params.projection_bias(), output_state_out)); + if (projection_threshold != 0.f) { - ARM_COMPUTE_RETURN_ON_ERROR(CLActivationLayerKernel::validate(output_state_out, output_state_out, - ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::LU_BOUNDED_RELU, -projection_threshold, projection_threshold))); + ARM_COMPUTE_RETURN_ON_ERROR(CLActivationLayer::validate( + output_state_out, output_state_out, + ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::LU_BOUNDED_RELU, -projection_threshold, + projection_threshold))); } } // Validate copy kernel - ARM_COMPUTE_RETURN_ON_ERROR(CLCopyKernel::validate(&cell_state_tmp, cell_state_out)); - ARM_COMPUTE_RETURN_ON_ERROR(CLCopyKernel::validate(output_state_out, output)); + ARM_COMPUTE_RETURN_ON_ERROR(CLCopy::validate(&cell_state_tmp, cell_state_out)); + ARM_COMPUTE_RETURN_ON_ERROR(CLCopy::validate(output_state_out, output)); // Validate scratch concatenation - std::vector<ITensorInfo *> inputs_vector_info_raw; - if(!lstm_params.has_cifg_opt()) + std::vector<const ITensorInfo *> inputs_vector_info_raw; + if (!lstm_params.has_cifg_opt()) { inputs_vector_info_raw.push_back(&input_gate); } @@ -623,110 +825,113 @@ void CLLSTMLayer::run() MemoryGroupResourceScope scope_mg(_memory_group); - CLScheduler::get().enqueue(_concat_inputs_forget_gate); + _concat_inputs_forget_gate.run(); _fully_connected_forget_gate.run(); - if(_run_peephole_opt) + if (_run_peephole_opt) { - CLScheduler::get().enqueue(_pixelwise_mul_forget_gate); + _pixelwise_mul_forget_gate.run(); _accum_forget_gate1.run(); } - if(_is_layer_norm_lstm) + if (_is_layer_norm_lstm) { _mean_std_norm_forget_gate.run(); - CLScheduler::get().enqueue(_pixelwise_mul_forget_gate_coeff); - CLScheduler::get().enqueue(_accum_forget_gate_bias); + _pixelwise_mul_forget_gate_coeff.run(); + _accum_forget_gate_bias.run(); } - CLScheduler::get().enqueue(_activation_forget_gate); + _activation_forget_gate.run(); - if(_run_cifg_opt) + if (_run_cifg_opt) { - CLScheduler::get().enqueue(_ones_memset_kernel); - CLScheduler::get().enqueue(_subtract_input_gate); + _ones_fill.run(); + _subtract_input_gate.run(); } else { _fully_connected_input_gate.run(); - if(_run_peephole_opt) + if (_run_peephole_opt) { - CLScheduler::get().enqueue(_pixelwise_mul_input_gate); + _pixelwise_mul_input_gate.run(); _accum_input_gate1.run(); } - if(_is_layer_norm_lstm) + if (_is_layer_norm_lstm) { _mean_std_norm_input_gate.run(); - CLScheduler::get().enqueue(_pixelwise_mul_input_gate_coeff); - CLScheduler::get().enqueue(_accum_input_gate_bias); + _pixelwise_mul_input_gate_coeff.run(); + _accum_input_gate_bias.run(); } - CLScheduler::get().enqueue(_activation_input_gate); + _activation_input_gate.run(); } _fully_connected_cell_state.run(); - CLScheduler::get().enqueue(_transpose_cell_state); + ITensorPack pack; + pack.add_tensor(TensorType::ACL_SRC, _recurrent_to_cell_weights); + pack.add_tensor(TensorType::ACL_DST, &_cell_state_out2); + CLScheduler::get().enqueue_op(*_transpose_cell_state, pack, false); _gemm_cell_state1.run(); - CLScheduler::get().enqueue(_accum_cell_state1); - if(_is_layer_norm_lstm) + _accum_cell_state1.run(); + if (_is_layer_norm_lstm) { _mean_std_norm_cell_gate.run(); - CLScheduler::get().enqueue(_pixelwise_mul_cell_gate_coeff); - CLScheduler::get().enqueue(_accum_cell_gate_bias); + _pixelwise_mul_cell_gate_coeff.run(); + _accum_cell_gate_bias.run(); } - CLScheduler::get().enqueue(_activation_cell_state); - CLScheduler::get().enqueue(_pixelwise_mul_cell_state1); - CLScheduler::get().enqueue(_pixelwise_mul_cell_state2); - CLScheduler::get().enqueue(_accum_cell_state2); + _activation_cell_state.run(); + _pixelwise_mul_cell_state1.run(); + _pixelwise_mul_cell_state2.run(); + _accum_cell_state2.run(); - if(_perform_cell_clipping) + if (_perform_cell_clipping) { - CLScheduler::get().enqueue(_cell_clip); + _cell_clip.run(); } _fully_connected_output.run(); - if(_run_peephole_opt) + if (_run_peephole_opt) { - CLScheduler::get().enqueue(_pixelwise_mul_output_state1); + _pixelwise_mul_output_state1.run(); _accum_output1.run(); } - if(_is_layer_norm_lstm) + if (_is_layer_norm_lstm) { _mean_std_norm_output_gate.run(); - CLScheduler::get().enqueue(_pixelwise_mul_output_gate_coeff); - CLScheduler::get().enqueue(_accum_output_gate_bias); + _pixelwise_mul_output_gate_coeff.run(); + _accum_output_gate_bias.run(); } - CLScheduler::get().enqueue(_activation_output); + _activation_output.run(); - CLScheduler::get().enqueue(_activation_output_state); - CLScheduler::get().enqueue(_pixelwise_mul_output_state2); + _activation_output_state.run(); + _pixelwise_mul_output_state2.run(); - if(_has_projection_weights) + if (_has_projection_weights) { _fully_connected_output_state.run(); - if(_perform_projection_clipping) + if (_perform_projection_clipping) { - CLScheduler::get().enqueue(_projection_clip); + _projection_clip.run(); } } - CLScheduler::get().enqueue(_copy_cell_state); - CLScheduler::get().enqueue(_copy_output); + _copy_cell_state.run(); + _copy_output.run(); _concat_scratch_buffer.run(); } void CLLSTMLayer::prepare() { - if(!_is_prepared) + if (!_is_prepared) { - CLScheduler::get().enqueue(_concat_weights_forget_gate); - if(!_run_cifg_opt) + _concat_weights_forget_gate.run(); + if (!_run_cifg_opt) { - CLScheduler::get().enqueue(_concat_weights_input_gate); + _concat_weights_input_gate.run(); } - CLScheduler::get().enqueue(_concat_weights_output); + _concat_weights_output.run(); _is_prepared = true; } } |