arm_compute v20.05
diff --git a/src/runtime/CL/functions/CLGEMMDeconvolutionLayer.cpp b/src/runtime/CL/functions/CLGEMMDeconvolutionLayer.cpp
index 14bda11..1dcb341 100644
--- a/src/runtime/CL/functions/CLGEMMDeconvolutionLayer.cpp
+++ b/src/runtime/CL/functions/CLGEMMDeconvolutionLayer.cpp
@@ -28,7 +28,6 @@
#include "arm_compute/core/utils/misc/ShapeCalculator.h"
#include "arm_compute/core/utils/quantization/AsymmHelpers.h"
#include "arm_compute/runtime/CL/CLScheduler.h"
-#include "utils/TypePrinter.h"
#include <memory>
#include <tuple>
@@ -62,6 +61,33 @@
return { start, end };
}
+Status construct_gemmlowp_output_stage(const ITensorInfo *input, const ITensorInfo *weights, const ITensorInfo *output, GEMMLowpOutputStageInfo &output_stage_info)
+{
+ const auto data_type = input->data_type();
+
+ if(is_data_type_quantized_asymmetric(data_type))
+ {
+ const UniformQuantizationInfo iq_info = input->quantization_info().uniform();
+ const UniformQuantizationInfo wq_info = weights->quantization_info().uniform();
+ const UniformQuantizationInfo oq_info = output->quantization_info().uniform();
+
+ float multiplier = iq_info.scale * wq_info.scale / oq_info.scale;
+ int output_multiplier(0);
+ int output_shift(0);
+ ARM_COMPUTE_RETURN_ON_ERROR(quantization::calculate_quantized_multiplier(multiplier, &output_multiplier, &output_shift));
+
+ output_stage_info.type = GEMMLowpOutputStageType::QUANTIZE_DOWN_FIXEDPOINT;
+ output_stage_info.gemmlowp_multiplier = output_multiplier;
+ output_stage_info.gemmlowp_shift = output_shift;
+ output_stage_info.gemmlowp_offset = oq_info.offset;
+ const auto min_max_bound = get_min_max(data_type);
+ output_stage_info.gemmlowp_min_bound = (std::get<0>(min_max_bound)).get<int32_t>();
+ output_stage_info.gemmlowp_max_bound = (std::get<1>(min_max_bound)).get<int32_t>();
+ output_stage_info.output_data_type = data_type;
+ }
+ return Status{};
+}
+
} // namespace
CLGEMMDeconvolutionLayer::CLGEMMDeconvolutionLayer(std::shared_ptr<IMemoryManager> memory_manager) // NOLINT
@@ -93,7 +119,7 @@
Status CLGEMMDeconvolutionLayer::validate(const ITensorInfo *input, const ITensorInfo *weights, const ITensorInfo *bias, const ITensorInfo *output, const PadStrideInfo &deconv_info)
{
ARM_COMPUTE_RETURN_ERROR_ON_NULLPTR(input, weights, output);
- ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(input, 1, DataType::F32, DataType::F16, DataType::QASYMM8);
+ ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(input, 1, DataType::F32, DataType::F16, DataType::QASYMM8, DataType::QASYMM8_SIGNED);
ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES(input, weights);
ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_LAYOUT(input, weights);
@@ -141,10 +167,13 @@
TensorInfo gemm_output_info = reshaped_t_info.clone()->set_tensor_shape(gemm_output_shape).set_is_resizable(true);
GEMMInfo gemm_info(false, false, true, input->dimension(idx_h), true);
+ GEMMLowpOutputStageInfo output_stage_info;
+
if(is_quantized)
{
ARM_COMPUTE_RETURN_ON_ERROR(CLGEMMLowpMatrixMultiplyCore::validate(&input->clone()->set_tensor_shape(nhwc_input_shape), &reshaped_t_info, nullptr, &gemm_output_info.set_data_type(DataType::S32),
gemm_info));
+ ARM_COMPUTE_RETURN_ON_ERROR(construct_gemmlowp_output_stage(input, weights, output, output_stage_info));
}
else
{
@@ -160,9 +189,8 @@
{
const auto start_end = compute_start_end_slice_coordinates(col2im_output_info, deconv_info, is_nchw);
ARM_COMPUTE_RETURN_ON_ERROR(CLDeconvolutionReshapeOutputKernel::validate(&gemm_output_info, bias, &col2im_output_info, input, weights, deconv_info));
- ARM_COMPUTE_RETURN_ON_ERROR(CLGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPoint::validate(&col2im_output_info, nullptr,
- &col2im_output_info.clone()->set_is_resizable(true).set_data_type(DataType::QASYMM8)));
- ARM_COMPUTE_RETURN_ON_ERROR(CLSlice::validate(&col2im_output_info.clone()->set_is_resizable(true).set_data_type(DataType::QASYMM8), output, start_end.first, start_end.second));
+ ARM_COMPUTE_RETURN_ON_ERROR(CLGEMMLowpOutputStage::validate(&col2im_output_info, nullptr, &col2im_output_info.clone()->set_is_resizable(true).set_data_type(input->data_type()), output_stage_info));
+ ARM_COMPUTE_RETURN_ON_ERROR(CLSlice::validate(&col2im_output_info.clone()->set_is_resizable(true).set_data_type(input->data_type()), output, start_end.first, start_end.second));
}
else if(padded_input)
{
@@ -173,16 +201,7 @@
else if(is_quantized)
{
ARM_COMPUTE_RETURN_ON_ERROR(CLDeconvolutionReshapeOutputKernel::validate(&gemm_output_info, bias, &col2im_output_info, input, weights, deconv_info));
-
- const UniformQuantizationInfo iq_info = input->quantization_info().uniform();
- const UniformQuantizationInfo wq_info = weights->quantization_info().uniform();
- const UniformQuantizationInfo oq_info = output->quantization_info().uniform();
-
- float multiplier = iq_info.scale * wq_info.scale / oq_info.scale;
- int output_multiplier(0);
- int output_shift(0);
- ARM_COMPUTE_RETURN_ON_ERROR(quantization::calculate_quantized_multiplier(multiplier, &output_multiplier, &output_shift));
- ARM_COMPUTE_RETURN_ON_ERROR(CLGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPoint::validate(&col2im_output_info, nullptr, output));
+ ARM_COMPUTE_RETURN_ON_ERROR(CLGEMMLowpOutputStage::validate(&col2im_output_info, nullptr, output, output_stage_info));
}
else
{
@@ -194,6 +213,12 @@
void CLGEMMDeconvolutionLayer::configure(const ICLTensor *input, const ICLTensor *weights, const ICLTensor *bias, ICLTensor *output, const PadStrideInfo &deconv_info)
{
+ configure(CLKernelLibrary::get().get_compile_context(), input, weights, bias, output, deconv_info);
+}
+
+void CLGEMMDeconvolutionLayer::configure(const CLCompileContext &compile_context, const ICLTensor *input, const ICLTensor *weights, const ICLTensor *bias, ICLTensor *output,
+ const PadStrideInfo &deconv_info)
+{
ARM_COMPUTE_ERROR_ON_NULLPTR(input, weights, output);
ARM_COMPUTE_ERROR_THROW_ON(CLGEMMDeconvolutionLayer::validate(input->info(),
weights->info(),
@@ -216,9 +241,9 @@
if(_is_nchw)
{
_memory_group.manage(&_permuted_input);
- _permute_input_to_nhwc.configure(input, &_permuted_input, PermutationVector(2U, 0U, 1U));
+ _permute_input_to_nhwc.configure(compile_context, input, &_permuted_input, PermutationVector(2U, 0U, 1U));
- _permute_weights_to_nhwc.configure(weights, &_permuted_weights, PermutationVector(2U, 0U, 1U));
+ _permute_weights_to_nhwc.configure(compile_context, weights, &_permuted_weights, PermutationVector(2U, 0U, 1U));
input_to_use = &_permuted_input;
weights_to_use = &_permuted_weights;
@@ -230,8 +255,8 @@
1,
input->info()->data_type(), weights->info()->quantization_info()));
- _reshape_weights.configure(weights_to_use, &_reshaped_weights);
- _transpose_weights.configure(&_reshaped_weights, &_reshaped_weights_t);
+ _reshape_weights.configure(compile_context, weights_to_use, &_reshaped_weights);
+ _transpose_weights.configure(compile_context, &_reshaped_weights, &_reshaped_weights_t);
const size_t idx_h = get_data_layout_dimension_index(input->info()->data_layout(), DataLayoutDimension::HEIGHT);
GEMMInfo gemm_info(false, false, true, input->info()->dimension(idx_h), true);
@@ -247,14 +272,14 @@
input_to_use->info()->set_quantization_info(QuantizationInfo(iq_info.uniform().scale, -iq_info.uniform().offset));
_reshaped_weights_t.info()->set_quantization_info(QuantizationInfo(wq_info.uniform().scale, -wq_info.uniform().offset));
- _mm_gemmlowp.configure(input_to_use, &_reshaped_weights_t, nullptr, &_gemm_output, gemm_info);
+ _mm_gemmlowp.configure(compile_context, input_to_use, &_reshaped_weights_t, nullptr, &_gemm_output, gemm_info);
input_to_use->info()->set_quantization_info(iq_info);
_reshaped_weights_t.info()->set_quantization_info(wq_info);
}
else
{
- _mm_gemm.configure(input_to_use, &_reshaped_weights_t, nullptr, &_gemm_output, 1.f, 0.0f, gemm_info);
+ _mm_gemm.configure(compile_context, input_to_use, &_reshaped_weights_t, nullptr, &_gemm_output, 1.f, 0.0f, gemm_info);
}
if(_is_nchw)
@@ -292,20 +317,14 @@
}
// Configure a Col2Im call to reshape the output of GEMM
- _deconv_reshape.configure(&_gemm_output, bias, deconv_reshape_output, input->info(), weights->info(), deconv_info);
+ _deconv_reshape.configure(compile_context, &_gemm_output, bias, deconv_reshape_output, input->info(), weights->info(), deconv_info);
_gemm_output.allocator()->allocate();
if(_is_quantized)
{
- const UniformQuantizationInfo iq_info = input->info()->quantization_info().uniform();
- const UniformQuantizationInfo wq_info = weights->info()->quantization_info().uniform();
- const UniformQuantizationInfo oq_info = output->info()->quantization_info().uniform();
-
- float multiplier = iq_info.scale * wq_info.scale / oq_info.scale;
- int output_multiplier(0);
- int output_shift(0);
- quantization::calculate_quantized_multiplier(multiplier, &output_multiplier, &output_shift);
- _gemmlowp_output_stage.configure(&_gemmlowp_final, nullptr, output_stage_output, output_multiplier, output_shift, oq_info.offset);
+ GEMMLowpOutputStageInfo output_stage_info;
+ construct_gemmlowp_output_stage(input->info(), weights->info(), output->info(), output_stage_info);
+ _gemmlowp_output_stage.configure(compile_context, &_gemmlowp_final, nullptr, output_stage_output, output_stage_info);
_gemmlowp_final.allocator()->allocate();
}
@@ -313,7 +332,7 @@
if(_padded_input)
{
const auto start_end = compute_start_end_slice_coordinates(*deconv_reshape_output->info(), deconv_info, _is_nchw);
- _slice_gemm.configure(&_slice_gemm_input, slice_output, start_end.first, start_end.second);
+ _slice_gemm.configure(compile_context, &_slice_gemm_input, slice_output, start_end.first, start_end.second);
_slice_gemm_input.allocator()->allocate();
}
}