47 if(output->total_size() != 0)
59 : _input(nullptr), _output(nullptr), _axis(nullptr)
93 const int axis_i = *(
reinterpret_cast<const int *
>(axis->
buffer()) + i);
94 axis_bit |= 1 << axis_i;
99 const int window_start_x = window.
x().
start();
100 const int window_end_x = window.
x().
end();
108 int x = window_start_x;
109 for(; x <= (window_end_x - window_step_x); x += window_step_x)
120 const int offset_x = (axis_bit & 0x1) ? output->
info()->
dimension(0) - x - window_step_x : x;
121 const int offset_y = (axis_bit & 0x2) ? output->
info()->
dimension(1) -
id.y() - 1 :
id.y();
122 const int offset_z = (axis_bit & 0x4) ? output->
info()->
dimension(2) -
id.z() - 1 :
id.z();
123 const int offset_w = (axis_bit & 0x8) ? output->
info()->
dimension(3) -
id[3] - 1 :
id[3];
130 for(; x < window_end_x; ++x)
132 const auto in = *(
reinterpret_cast<T *
>(input_it.
ptr()) + x);
134 const int offset_x = (axis_bit & 0x1) ? output->
info()->
dimension(0) - x - 1 : x;
135 const int offset_y = (axis_bit & 0x2) ? output->
info()->
dimension(1) -
id.y() - 1 :
id.y();
136 const int offset_z = (axis_bit & 0x4) ? output->
info()->
dimension(2) -
id.z() - 1 :
id.z();
137 const int offset_w = (axis_bit & 0x8) ? output->
info()->
dimension(3) -
id[3] - 1 :
id[3];
154 run_reverse<uint32_t>(
window, _input, _axis, _output);
157 run_reverse<uint16_t>(
window, _input, _axis, _output);
160 run_reverse<uint8_t>(
window, _input, _axis, _output);
void run(const Window &window, const ThreadInfo &info) override
Execute the kernel on the passed window.
Window calculate_max_window(const ValidRegion &valid_region, const Steps &steps, bool skip_border, BorderSize border_size)
const Window & window() const
The maximum window the kernel can be executed on.
uint8_t * ptr_to_element(const Coordinates &id) const
Return a pointer to the element at the passed coordinates.
virtual size_t dimension(size_t index) const =0
Return the size of the requested dimension.
static Status validate(const ITensorInfo *input, const ITensorInfo *output, const ITensorInfo *axis)
Static function to check if given info will lead to a valid configuration of NEReverseKernel.
#define ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_QUANTIZATION_INFO(...)
#define ARM_COMPUTE_ERROR(msg)
Print the given message then throw an std::runtime_error.
uint8x16_t vloadq(const uint8_t *ptr)
#define ARM_COMPUTE_RETURN_ON_ERROR(status)
Checks if a status contains an error and returns it.
Store the tensor's metadata.
#define ARM_COMPUTE_ERROR_THROW_ON(status)
Describe one of the image's dimensions with a start, end and step.
#define ARM_COMPUTE_RETURN_ERROR_ON(cond)
If the condition is true, an error is returned.
Interface for Neon tensor.
NEReverseKernel()
Default constructor.
void configure(const ITensor *input, ITensor *output, const ITensor *axis)
Initialise the kernel's inputs and output.
Copyright (c) 2017-2021 Arm Limited.
#define ARM_COMPUTE_RETURN_ERROR_ON_NULLPTR(...)
static constexpr size_t DimX
Alias for dimension 0 also known as X dimension.
#define ARM_COMPUTE_UNUSED(...)
To avoid unused variables warnings.
1 channel, 1 U32 per channel
virtual uint8_t * buffer() const =0
Interface to be implemented by the child class to return a pointer to CPU memory. ...
bool auto_init_if_empty(ITensorInfo &info, const TensorShape &shape, int num_channels, DataType data_type, QuantizationInfo quantization_info=QuantizationInfo())
Auto initialize the tensor info (shape, number of channels and data type) if the current assignment i...
virtual std::unique_ptr< T > clone() const =0
Provide a clone of the current object of class T.
virtual ITensorInfo * info() const =0
Interface to be implemented by the child class to return the tensor's metadata.
constexpr uint8_t * ptr() const
Return a pointer to the current pixel.
uint8x8_t vgetlow(const uint8x16_t val)
virtual size_t element_size() const =0
Element size in bytes calculated as data_size() * num_channels()
void set(size_t dimension, const Dimension &dim)
Set the values of a given dimension.
uint8x16_t vcombine(const uint8x8_t &a, const uint8x8_t &b)
#define ARM_COMPUTE_ERROR_ON_UNCONFIGURED_KERNEL(k)
uint8x8_t vgethigh(const uint8x16_t val)
ScaleKernelInfo info(interpolation_policy, default_border_mode, PixelValue(), sampling_policy, false)
Information about executing thread and CPU.
#define ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_SHAPES(...)
#define ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES(...)
#define ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(t, c,...)
Status validate_arguments(const ITensorInfo *input, const ITensorInfo *bias, const ITensorInfo *output, const GEMMLowpOutputStageInfo *output_stage)
uint8x8_t vrev64(const uint8x8_t &a)
void vstore(uint8_t *ptr, uint8x8_t val)
#define ARM_COMPUTE_RETURN_ERROR_ON_MSG(cond, msg)
If the condition is true, an error is returned.
#define ARM_COMPUTE_ERROR_ON_NULLPTR(...)
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...
Includes all wrapper headers at once.
void run_reverse(const Window &window, const ITensor *input, const ITensor *axis, ITensor *output)
constexpr int end() const
Return the end of the dimension.
Iterator updated by execute_window_loop for each window element.
constexpr int start() const
Return the start of the dimension.
Describe a multidimensional execution window.
#define ARM_COMPUTE_ERROR_ON_INVALID_SUBWINDOW(f, s)
constexpr const Dimension & x() const
Alias to access the first dimension of the window.