Compute Library
 21.11
ClQuantizeKernel.cpp
Go to the documentation of this file.
1 /*
2  * Copyright (c) 2017-2021 Arm Limited.
3  *
4  * SPDX-License-Identifier: MIT
5  *
6  * Permission is hereby granted, free of charge, to any person obtaining a copy
7  * of this software and associated documentation files (the "Software"), to
8  * deal in the Software without restriction, including without limitation the
9  * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
10  * sell copies of the Software, and to permit persons to whom the Software is
11  * furnished to do so, subject to the following conditions:
12  *
13  * The above copyright notice and this permission notice shall be included in all
14  * copies or substantial portions of the Software.
15  *
16  * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
17  * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
18  * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
19  * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
20  * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
21  * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
22  * SOFTWARE.
23  */
25 
29 #include "arm_compute/core/Error.h"
31 #include "arm_compute/core/Utils.h"
34 
35 #include "src/core/CL/CLValidate.h"
37 
38 #include "support/Cast.h"
39 #include "support/StringSupport.h"
40 
41 namespace arm_compute
42 {
43 namespace opencl
44 {
45 namespace kernels
46 {
47 namespace
48 {
49 Status validate_arguments(const ITensorInfo *src, const ITensorInfo *dst)
50 {
54 
55  // Output must always be initialized
56  ARM_COMPUTE_RETURN_ERROR_ON(dst->tensor_shape().total_size() == 0);
59 
60  return Status{};
61 }
62 } // namespace
63 
65 {
67 }
68 
69 void ClQuantizeKernel::configure(const CLCompileContext &compile_context, const ITensorInfo *src, ITensorInfo *dst)
70 {
72 
73  auto padding_info = get_padding_info({ src, dst });
74 
75  ARM_COMPUTE_ERROR_THROW_ON(validate_arguments(src, dst));
76 
77  const int vec_size_x = 16 / src->element_size();
78  const int input_width_x = src->tensor_shape().x();
79  const bool multi_access_x = (input_width_x / vec_size_x > 0);
80 
82  const DataType output_data_type = dst->data_type();
83 
84  float scale_to_apply = qinfo.scale;
85  int32_t offset_to_apply = qinfo.offset;
87  {
88  /*
89  * In case of requantization of a quantized input tensor to an output tensor with another quantization
90  * instead of of apply dequantization and then a quantization functions, we just compute new scale and
91  * offset to apply.
92  *
93  * Assuming:
94  * - q_i as input quantized value
95  * - q_o as output quantized value
96  * - z_i as input quantization offset value
97  * - z_o as output quantization offset value
98  * - s_i as input quantization scale value
99  * - s_o as output quantization scale value
100  * - z_n as new quantization offset value
101  * - s_n as new quantization scale value
102  *
103  * q_o = ( q_i - z_i ) * s_i / s_o + z_o
104  *
105  * We can rewrite the formula as:
106  *
107  * q_o = ( q_i * s_i / s_o ) - z_i * s_i / s_o + z_o
108  *
109  * q_o = q_i / s_n + z_n
110  *
111  * Where:
112  *
113  * s_n = s_o / s_i
114  *
115  * z_n = - z_i * s_i / s_o + z_o
116  *
117  */
118  const UniformQuantizationInfo qinfo_in = src->quantization_info().uniform();
119  scale_to_apply /= qinfo_in.scale;
120  // In order to minimize flooring we convert the offset to a float,
121  // then compute the new offset in the float domain,
122  // finally we convert it back as int32_t
123  offset_to_apply -= static_cast<int32_t>(static_cast<float>(qinfo_in.offset) * qinfo_in.scale / qinfo.scale);
124  }
125 
126  // Create kernel
127  CLBuildOptions build_opts;
128  build_opts.add_option_if(is_data_type_float(src->data_type()), "-DIS_FLOAT");
129  build_opts.add_option("-DSCALE=" + float_to_string_with_full_precision(scale_to_apply));
130  build_opts.add_option("-DOFFSET=" + support::cpp11::to_string(offset_to_apply));
131  build_opts.add_option("-DVEC_SIZE=" + support::cpp11::to_string(vec_size_x));
132  build_opts.add_option("-DDATA_TYPE_IN=" + get_cl_type_from_data_type(src->data_type()));
133  build_opts.add_option("-DDATA_TYPE_OUT=" + get_cl_type_from_data_type(output_data_type));
134  build_opts.add_option_if(multi_access_x, "-DLAST_ACCESSED_X=" + support::cpp11::to_string(std::max<int>(input_width_x - vec_size_x, 0)));
135  std::pair<int, int> min_max_quant_values = quantization::get_min_max_values_from_quantized_data_type(output_data_type);
136  build_opts.add_option("-DMIN_QUANT_VAL=" + support::cpp11::to_string(min_max_quant_values.first));
137  build_opts.add_option("-DMAX_QUANT_VAL=" + support::cpp11::to_string(min_max_quant_values.second));
138 
139  _kernel = create_kernel(compile_context, "quantization_layer", build_opts.options());
140 
141  // Configure kernel window
142  Window win = calculate_max_window(*src, Steps());
143  if(multi_access_x)
144  {
145  win.set(Window::DimX, Window::Dimension(win.x().start(), ceil_to_multiple(win.x().end(), vec_size_x), vec_size_x));
146  }
147  ICLKernel::configure_internal(win);
148 
150 }
151 
153 {
154  ARM_COMPUTE_RETURN_ON_ERROR(validate_arguments(src, dst));
155  return Status{};
156 }
157 
158 void ClQuantizeKernel::run_op(ITensorPack &tensors, const Window &window, cl::CommandQueue &queue)
159 {
162 
163  auto src = utils::cast::polymorphic_downcast<const ICLTensor *>(tensors.get_const_tensor(TensorType::ACL_SRC));
164  auto dst = utils::cast::polymorphic_downcast<ICLTensor *>(tensors.get_tensor(TensorType::ACL_DST));
165 
166  Window window_collapsed = window.collapse_if_possible(ICLKernel::window(), 3);
167  Window slice = window_collapsed.first_slice_window_3D();
168 
169  do
170  {
171  unsigned int idx = 0;
172  add_3D_tensor_argument(idx, src, slice);
173  add_3D_tensor_argument(idx, dst, slice);
174  enqueue(queue, *this, slice, lws_hint());
175  }
176  while(window_collapsed.slide_window_slice_3D(slice));
177 }
178 } // namespace kernels
179 } // namespace opencl
180 } // namespace arm_compute
Window calculate_max_window(const ValidRegion &valid_region, const Steps &steps, bool skip_border, BorderSize border_size)
#define ARM_COMPUTE_RETURN_ERROR_ON_F16_UNSUPPORTED(tensor)
Definition: CLValidate.h:35
const Window & window() const
The maximum window the kernel can be executed on.
Definition: IKernel.cpp:28
void enqueue(cl::CommandQueue &queue, ICLKernel &kernel, const Window &window, const cl::NDRange &lws_hint=CLKernelLibrary::get().default_ndrange(), bool use_dummy_work_items=false)
Add the kernel to the command queue with the given window.
Definition: ICLKernel.cpp:32
const StringSet & options() const
Gets the current options list set.
cl::NDRange lws_hint() const
Return the Local-Workgroup-Size hint.
Definition: ICLKernel.h:318
#define ARM_COMPUTE_RETURN_ON_ERROR(status)
Checks if a status contains an error and returns it.
Definition: Error.h:204
std::string to_string(T &&value)
Convert integer and float values to string.
virtual DataType data_type() const =0
Data type used for each element of the tensor.
1 channel, 1 F32 per channel
#define ARM_COMPUTE_ERROR_ON(cond)
If the condition is true then an error message is printed and an exception thrown.
Definition: Error.h:466
Store the tensor&#39;s metadata.
Definition: ITensorInfo.h:40
#define ARM_COMPUTE_ERROR_THROW_ON(status)
Definition: Error.h:455
Quantization info when assuming per layer quantization.
Describe one of the image&#39;s dimensions with a start, end and step.
Definition: Window.h:77
quantized, asymmetric fixed-point 16-bit number
Manages all the OpenCL kernels compilation and caching, provides accessors for the OpenCL Context...
Status class.
Definition: Error.h:52
#define ARM_COMPUTE_RETURN_ERROR_ON(cond)
If the condition is true, an error is returned.
Definition: Error.h:296
void add_3D_tensor_argument(unsigned int &idx, const ICLTensor *tensor, const Window &window)
Add the passed 3D tensor&#39;s parameters to the object&#39;s kernel&#39;s arguments starting from the index idx...
Definition: ICLKernel.h:214
SimpleTensor< float > src
Definition: DFT.cpp:155
Copyright (c) 2017-2021 Arm Limited.
1 channel, 1 F16 per channel
#define ARM_COMPUTE_RETURN_ERROR_ON_NULLPTR(...)
Definition: Validate.h:159
void add_option(std::string option)
Adds option to the existing build option list.
T x() const
Alias to access the size of the first dimension.
Definition: Dimensions.h:87
const ITensor * get_const_tensor(int id) const
Get constant tensor of a given id.
Definition: ITensorPack.cpp:54
cl::Kernel create_kernel(const CLCompileContext &ctx, const std::string &kernel_name, const std::set< std::string > &build_opts=std::set< std::string >())
Creates an opencl kernel using a compile context.
Definition: CLHelpers.cpp:391
std::pair< int, int > get_min_max_values_from_quantized_data_type(DataType data_type)
Get minimum and maximum values for the input quantized data type.
static constexpr size_t DimX
Alias for dimension 0 also known as X dimension.
Definition: Window.h:43
Window collapse_if_possible(const Window &full_window, size_t first, size_t last, bool *has_collapsed=nullptr) const
Collapse the dimensions between first and last if possible.
Definition: Window.inl:68
std::string float_to_string_with_full_precision(float val)
Create a string with the float in full precision.
Definition: Utils.h:1075
virtual const TensorShape & tensor_shape() const =0
Size for each dimension of the tensor.
auto ceil_to_multiple(S value, T divisor) -> decltype(((value+divisor - 1)/divisor) *divisor)
Computes the smallest number larger or equal to value that is a multiple of divisor.
Definition: Utils.h:71
quantized, asymmetric fixed-point 8-bit number unsigned
Class to describe a number of elements in each dimension.
Definition: Steps.h:40
void run_op(ITensorPack &tensors, const Window &window, cl::CommandQueue &queue) override
Enqueue the OpenCL kernel to process the given window on the passed OpenCL command queue...
UniformQuantizationInfo uniform() const
Return per layer quantization info.
std::string get_cl_type_from_data_type(const DataType &dt)
Translates a tensor data type to the appropriate OpenCL type.
Definition: CLHelpers.cpp:39
void add_option_if(bool cond, std::string option)
Adds option if a given condition is true;.
virtual size_t element_size() const =0
Element size in bytes calculated as data_size() * num_channels()
Elementeise CL kernel type.
Definition: CLTypes.h:84
virtual QuantizationInfo quantization_info() const =0
Get the quantization settings (scale and offset) of the tensor.
#define ARM_COMPUTE_ERROR_ON_UNCONFIGURED_KERNEL(k)
Definition: Validate.h:915
bool has_padding_changed(const std::unordered_map< const ITensorInfo *, PaddingSize > &padding_map)
Check if the previously stored padding info has changed after configuring a kernel.
Definition: Utils.cpp:533
CLCompileContext class.
bool is_data_type_quantized_asymmetric(DataType dt)
Check if a given data type is of asymmetric quantized type.
Definition: Utils.h:1003
ITensor * get_tensor(int id)
Get tensor of a given id from the pac.
Definition: ITensorPack.cpp:64
#define ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_SHAPES(...)
Definition: Validate.h:439
const QuantizationInfo qinfo
Definition: Im2Col.cpp:155
static Status validate(const ITensorInfo *src, const ITensorInfo *dst)
Static function to check if given info will lead to a valid configuration.
#define ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(t, c,...)
Definition: Validate.h:788
std::unordered_map< const ITensorInfo *, PaddingSize > get_padding_info(std::initializer_list< const ITensorInfo *> infos)
Stores padding information before configuring a kernel.
Definition: Utils.cpp:518
Tensor packing service.
Definition: ITensorPack.h:39
#define ARM_COMPUTE_ERROR_ON_NULLPTR(...)
Definition: Validate.h:157
quantized, asymmetric fixed-point 8-bit number signed
DataType
Available data types.
Definition: Types.h:79
void configure(const CLCompileContext &compile_context, const ITensorInfo *src, ITensorInfo *dst)
Set the input, output.
Describe a multidimensional execution window.
Definition: Window.h:39
bool is_data_type_float(DataType dt)
Check if a given data type is of floating point type.
Definition: Utils.h:961
#define ARM_COMPUTE_ERROR_ON_INVALID_SUBWINDOW(f, s)
Definition: Validate.h:201
SimpleTensor< T > slice(const SimpleTensor< T > &src, Coordinates starts, Coordinates ends)