24.02.1
|
Go to the documentation of this file.
72 _height_offset = height_offset;
76 ICpuKernel::configure(win);
95 uint8_t *dst_ptr =
dst->buffer() +
dst->info()->offset_first_element_in_bytes() +
98 const auto window_start_x =
static_cast<int>(
window.
x().
start());
99 const auto window_end_x =
static_cast<int>(
window.
x().
end()) *
static_cast<int>(
dst->info()->element_size());
100 const int window_step_x = 16;
119 int x = window_start_x;
120 for (; x <= (window_end_x - window_step_x); x += window_step_x)
122 vst1q_u8(dst_ptr + dst_it.
offset() + x,
127 for (; x < window_end_x; ++x)
129 *(dst_ptr + dst_it.
offset() + x) =
141 int x = window_start_x;
142 for (; x <= (window_end_x - window_step_x); x += window_step_x)
145 reinterpret_cast<int8_t *
>(dst_ptr + dst_it.
offset() + x),
151 for (; x < window_end_x; ++x)
153 *(dst_ptr + dst_it.
offset() + x) =
165 const auto in_ptr = src_it.
ptr();
166 const auto out_ptr = dst_ptr + dst_it.
offset();
168 int x = window_start_x;
169 for (; x <= (window_end_x - window_step_x); x += window_step_x)
175 for (; x < window_end_x; ++x)
177 *(out_ptr + x) = *(in_ptr + x);
186 return "CpuConcatenateHeightKernel";
Class to describe a number of elements in each dimension.
float32x4x2_t vdequantize(const uint8x8_t &qv, const UniformQuantizationInfo &qi)
Dequantize a neon vector holding 8 quantized values.
constexpr int start() const
Return the start of the dimension.
float dequantize_qasymm8(uint8_t value, const INFO_TYPE &qinfo)
Dequantize a value given an unsigned 8-bit asymmetric quantization scheme.
SimpleTensor< float > src
Window calculate_max_window(const ValidRegion &valid_region, const Steps &steps, bool skip_border, BorderSize border_size)
@ QASYMM8
quantized, asymmetric fixed-point 8-bit number unsigned
Status validate_arguments(const ITensorInfo *src, const ITensorInfo *weights, const ITensorInfo *dst, const PadStrideInfo &conv_info)
#define ARM_COMPUTE_ERROR_ON_UNCONFIGURED_KERNEL(k)
static constexpr size_t DimX
Alias for dimension 0 also known as X dimension.
uint8_t quantize_qasymm8(float value, const INFO_TYPE &qinfo, RoundingPolicy rounding_policy=RoundingPolicy::TO_NEAREST_UP)
Quantize a value given an unsigned 8-bit asymmetric quantization scheme.
void run_op(ITensorPack &tensors, const Window &window, const ThreadInfo &info) override
Execute the kernel on the passed window.
ITensor * get_tensor(int id)
Get tensor of a given id from the pac.
uint8x16_t vloadq(const uint8_t *ptr)
Includes all wrapper headers at once.
#define ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES(...)
#define ARM_COMPUTE_RETURN_ON_ERROR(status)
Checks if a status contains an error and returns it.
#define ARM_COMPUTE_ERROR_ON_NULLPTR(...)
constexpr uint8_t * ptr() const
Return a pointer to the current pixel.
const ITensor * get_const_tensor(int id) const
Get constant tensor of a given id.
const char * name() const override
Name of the kernel.
#define ARM_COMPUTE_ERROR_THROW_ON(status)
void execute_window_loop(const Window &w, L &&lambda_function, Ts &&...iterators)
Iterate through the passed window, automatically adjusting the iterators and calling the lambda_funct...
Iterator updated by execute_window_loop for each window element.
int8_t quantize_qasymm8_signed(float value, const INFO_TYPE &qinfo, RoundingPolicy rounding_policy=RoundingPolicy::TO_NEAREST_UP)
Quantize a value given a signed 8-bit asymmetric quantization scheme.
#define ARM_COMPUTE_RETURN_ERROR_ON(cond)
If the condition is true, an error is returned.
@ QASYMM8_SIGNED
quantized, asymmetric fixed-point 8-bit number signed
#define ARM_COMPUTE_ERROR_ON_INVALID_SUBWINDOW(f, s)
#define ARM_COMPUTE_UNUSED(...)
To avoid unused variables warnings.
Describe one of the image's dimensions with a start, end and step.
void set(size_t dimension, const Dimension &dim)
Set the values of a given dimension.
static constexpr size_t DimY
Alias for dimension 1 also known as Y dimension.
const Window & window() const
The maximum window the kernel can be executed on.
Information about executing thread and CPU.
void vstore(uint8_t *ptr, uint8x8_t val)
Describe a multidimensional execution window.
float dequantize_qasymm8_signed(int8_t value, const INFO_TYPE &qinfo)
Dequantize a value given a signed 8-bit asymmetric quantization scheme.
constexpr size_t offset() const
Return the offset in bytes from the first element to the current position of the iterator.
uint8x8_t vquantize(const float32x4x2_t &qv, const UniformQuantizationInfo &qi)
Quantize a neon vector holding 8 floating point values.
Copyright (c) 2017-2024 Arm Limited.
static Status validate(const ITensorInfo *src, unsigned int height_offset, const ITensorInfo *dst)
Static function to check if given info will lead to a valid configuration.
void configure(const ITensorInfo *src, unsigned int height_offset, ITensorInfo *dst)
Configure kernel for a given list of arguments.
#define ARM_COMPUTE_RETURN_ERROR_ON_NULLPTR(...)
Store the tensor's metadata.
ScaleKernelInfo info(interpolation_policy, default_border_mode, PixelValue(), sampling_policy, false)
@ UNKNOWN
Unknown data type.
constexpr int end() const
Return the end of the dimension.
int8x8_t vquantize_signed(const float32x4x2_t &qv, const UniformQuantizationInfo &qi)
Quantize a neon vector holding 8 floating point values.
constexpr const Dimension & x() const
Alias to access the first dimension of the window.
DataType
Available data types.
static constexpr size_t num_max_dimensions
Number of dimensions the tensor has.