IntenseWebs/openvino
3
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

[GPU] Add supporting blocked format for reverse operation (#12032)

Browse Source
This commit is contained in:
OlehKravchyshyn
2022-10-20 08:18:00 +03:00
committed by GitHub
parent 09b349607e
commit 3500d75187
6 changed files with 266 additions and 179 deletions
Download Patch File Download Diff File Expand all files Collapse all files

50
src/plugins/intel_gpu/src/graph/impls/ocl/reverse.cpp
View File

@@ -50,33 +50,33 @@ public:
namespace detail {
attach_reverse_impl::attach_reverse_impl() {
implementation_map<reverse>::add(impl_types::ocl,
reverse_impl::create,
{
std::make_tuple(data_types::f16, format::bfyx),
std::make_tuple(data_types::f16, format::bfzyx),
std::make_tuple(data_types::f16, format::bfwzyx),
static const auto types =
{data_types::f16, data_types::f32, data_types::i8, data_types::u8, data_types::i32, data_types::i64};
static const auto formats = {
format::bfyx,
format::b_fs_yx_fsv16,
format::b_fs_yx_fsv32,
format::bs_fs_yx_bsv16_fsv16,
format::bs_fs_yx_bsv32_fsv32,
format::bs_fs_yx_bsv32_fsv16,
std::make_tuple(data_types::f32, format::bfyx),
std::make_tuple(data_types::f32, format::bfzyx),
std::make_tuple(data_types::f32, format::bfwzyx),
format::bfzyx,
format::b_fs_zyx_fsv16,
format::b_fs_zyx_fsv32,
format::bs_fs_zyx_bsv16_fsv32,
format::bs_fs_zyx_bsv16_fsv16,
format::bs_fs_zyx_bsv32_fsv32,
format::bs_fs_zyx_bsv32_fsv16,
std::make_tuple(data_types::i32, format::bfyx),
std::make_tuple(data_types::i32, format::bfzyx),
std::make_tuple(data_types::i32, format::bfwzyx),
std::make_tuple(data_types::i64, format::bfyx),
std::make_tuple(data_types::i64, format::bfzyx),
std::make_tuple(data_types::i64, format::bfwzyx),
std::make_tuple(data_types::i8, format::bfyx),
std::make_tuple(data_types::i8, format::bfzyx),
std::make_tuple(data_types::i8, format::bfwzyx),
std::make_tuple(data_types::u8, format::bfyx),
std::make_tuple(data_types::u8, format::bfzyx),
std::make_tuple(data_types::u8, format::bfwzyx)
});
format::bfwzyx,
};
std::set<std::tuple<data_types, format::type>> keys;
for (const auto t : types) {
for (const auto f : formats) {
keys.emplace(t, f);
}
}
implementation_map<reverse>::add(impl_types::ocl, reverse_impl::create, keys);
}
} // namespace detail

7
src/plugins/intel_gpu/src/graph/program.cpp
View File

@@ -61,6 +61,7 @@
#include "region_yolo_inst.h"
#include "strided_slice_inst.h"
#include "loop_inst.h"
#include "reverse_inst.h"
#include "to_string_utils.h"
#include "runtime/cldnn_itt.hpp"
#include "runtime/kernels_cache.hpp"
@@ -1445,7 +1446,8 @@ void program::set_layout_optimizer_attributes(layout_optimizer& lo) {
prim.type() != cldnn::prior_box::type_id() &&
prim.type() != cldnn::resample::type_id() &&
prim.type() != cldnn::eye::type_id() &&
prim.type() != cldnn::generate_proposals::type_id()) {
prim.type() != cldnn::generate_proposals::type_id() &&
prim.type() != cldnn::reverse::type_id()) {
can_use_fsv16 = false;
}
@@ -1483,7 +1485,8 @@ void program::set_layout_optimizer_attributes(layout_optimizer& lo) {
prim.type() != cldnn::resample::type_id() &&
prim.type() != cldnn::prior_box::type_id() &&
prim.type() != cldnn::eye::type_id() &&
prim.type() != cldnn::generate_proposals::type_id()) {
prim.type() != cldnn::generate_proposals::type_id() &&
prim.type() != cldnn::reverse::type_id()) {
can_use_bs_fs_yx_bsv16_fsv16 = false;
}
}

2
src/plugins/intel_gpu/src/kernel_selector/kernels/reverse/reverse_kernel_ref.cpp
View File

@@ -26,6 +26,8 @@ ParamsKey ReverseKernelRef::GetSupportedKey() const {
k.EnableOutputDataType(Datatype::UINT8);
k.EnableAllInputLayout();
k.EnableAllOutputLayout();
k.EnableTensorOffset();
k.EnableTensorPitches();
k.EnableBatching();
k.EnableDifferentTypes();
return k;

353
src/plugins/intel_gpu/tests/test_cases/reverse_gpu_test.cpp
View File

@@ -47,31 +47,41 @@ public:
auto data_type = type_to_data_type<T>::value;
ReverseParams<T, mode> params = testing::TestWithParam<ReverseParams<T, mode>>::GetParam();
auto& engine = get_test_engine();
tensor t;
auto reverse_input = engine.allocate_memory({data_type, params.input_format, params.input_tensor});
format fmt = generic_test::get_plain_format_for(params.input_format);
bool reorder_needed = fmt != params.input_format;
auto reverse_input = engine.allocate_memory({data_type, fmt, params.input_tensor});
auto reverse_axes = engine.allocate_memory(
{ReverseModeTraits<mode>::data_type, format::bfyx, tensor(batch(1), feature(params.axis.size()))});
{ReverseModeTraits<mode>::data_type, fmt, tensor(batch(1), feature(params.axis.size()))});
set_values(reverse_input, params.input);
set_values(reverse_axes, params.axis);
const std::string reverse_id = "reverse";
const std::string reverse_input_id = "reverse_input";
const std::string axes_id = "reverse_axes";
topology topology;
topology.add(input_layout(reverse_input_id, reverse_input->get_layout()));
topology.add(input_layout(axes_id, reverse_axes->get_layout()));
topology.add(reverse(reverse_id, reverse_input_id, axes_id, mode));
network network(engine, topology);
topology tp;
tp.add(input_layout(reverse_input_id, reverse_input->get_layout()));
tp.add(input_layout(axes_id, reverse_axes->get_layout()));
const std::string reverse_id = "reverse";
std::string ouput_op_name{reverse_id};
if (reorder_needed) {
const std::string r_reverse_input_id = "r_reverse_input";
const std::string r_axes_id = "r_reverse_axes";
tp.add(reorder(r_reverse_input_id, reverse_input_id, params.input_format, type_to_data_type<T>::value));
tp.add(reorder(r_axes_id, axes_id, params.input_format, type_to_data_type<T>::value));
tp.add(reverse(reverse_id, r_reverse_input_id, r_axes_id, mode));
ouput_op_name = "reversed_result";
tp.add(reorder(ouput_op_name, reverse_id, fmt, type_to_data_type<T>::value));
} else {
tp.add(reverse(reverse_id, reverse_input_id, axes_id, mode));
}
network network(engine, tp);
network.set_input_data(reverse_input_id, reverse_input);
network.set_input_data(axes_id, reverse_axes);
auto result = network.execute();
auto out_mem = result.at(reverse_id).get_memory();
auto out_mem = result.at(ouput_op_name).get_memory();
cldnn::mem_lock<T> out_ptr(out_mem, get_test_stream());
ASSERT_EQ(params.expected_out.size(), out_ptr.size());
@@ -90,7 +100,7 @@ struct PrintToStringParamName {
for (auto val : param.param.axis) {
buf << val << ",";
}
buf << "} format" << param.param.input_format.to_string();
buf << "} format " << param.param.input_format.to_string();
return buf.str();
}
};
@@ -156,161 +166,200 @@ TEST_P(reverse_gpu_test_f16_index, reverse_f16_index) {
ASSERT_NO_FATAL_FAILURE(test());
}
namespace {
const auto four_d_formats = {
format::bfyx,
format::b_fs_yx_fsv16,
format::b_fs_yx_fsv32,
format::bs_fs_yx_bsv16_fsv16,
format::bs_fs_yx_bsv32_fsv32,
format::bs_fs_yx_bsv32_fsv16,
};
const auto five_d_formats = {
format::bfzyx,
format::b_fs_zyx_fsv16,
format::b_fs_zyx_fsv32,
format::bs_fs_zyx_bsv16_fsv32,
format::bs_fs_zyx_bsv16_fsv16,
format::bs_fs_zyx_bsv32_fsv32,
format::bs_fs_zyx_bsv32_fsv16,
};
} // namespace
template <typename T>
std::vector<ReverseParams<T, reverse_mode::mask>> generateMaskParams() {
std::vector<ReverseParams<T, reverse_mode::mask>> params{
// reverse_2d_1_mask
{tensor(batch(4), feature(3)),
format::bfyx,
std::vector<T>{0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11},
{false, true},
std::vector<T>{2, 1, 0, 5, 4, 3, 8, 7, 6, 11, 10, 9}},
{tensor(batch(4), feature(3)),
format::bfyx,
std::vector<T>{0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11},
{true, true},
std::vector<T>{11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0}},
{tensor(batch(4), feature(3)),
format::bfyx,
std::vector<T>{0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11},
{false, false},
std::vector<T>{0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11}},
};
std::vector<ReverseParams<T, reverse_mode::mask>> params;
for (const auto f : four_d_formats) {
params.push_back({tensor(batch(4), feature(3)),
f,
std::vector<T>{0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11},
{false, true},
std::vector<T>{2, 1, 0, 5, 4, 3, 8, 7, 6, 11, 10, 9}});
params.push_back({tensor(batch(4), feature(3)),
f,
std::vector<T>{0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11},
{true, true},
std::vector<T>{11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0}});
params.push_back({tensor(batch(4), feature(3)),
f,
std::vector<T>{0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11},
{false, false},
std::vector<T>{0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11}});
}
return params;
}
template <typename T>
std::vector<ReverseParams<T, reverse_mode::index>> generateIndexParams() {
std::vector<ReverseParams<T, reverse_mode::index>> params{
std::vector<ReverseParams<T, reverse_mode::index>> params;
for (const auto fmt : four_d_formats) {
std::vector<ReverseParams<T, reverse_mode::index>> local_params{
//{tensor(batch(8)), format::bfyx, std::vector<T>{0, 1, 2, 3, 4, 5, 6, 7}, {},
// std::vector<T>{0, 1, 2, 3, 4, 5, 6, 7}},
{tensor(batch(8)),
fmt,
std::vector<T>{0, 1, 2, 3, 4, 5, 6, 7},
{0},
std::vector<T>{7, 6, 5, 4, 3, 2, 1, 0}},
{tensor(batch(4), feature(3)),
fmt,
std::vector<T>{0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11},
{0},
std::vector<T>{9, 10, 11, 6, 7, 8, 3, 4, 5, 0, 1, 2}},
{tensor(batch(4), feature(3)),
fmt,
std::vector<T>{0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11},
{1},
std::vector<T>{2, 1, 0, 5, 4, 3, 8, 7, 6, 11, 10, 9}},
{tensor(batch(4), feature(3)),
fmt,
std::vector<T>{0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11},
{0, 1},
std::vector<T>{11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0}},
{tensor{2, 4, 1, 3},
fmt,
std::vector<T>{0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23},
{0, 1},
std::vector<T>{21, 22, 23, 18, 19, 20, 15, 16, 17, 12, 13, 14, 9, 10, 11, 6, 7, 8, 3, 4, 5, 0, 1, 2}},
{tensor{
2,
4,
1,
3,
},
fmt,
std::vector<T>{0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23},
{0, 2},
std::vector<T>{14, 13, 12, 17, 16, 15, 20, 19, 18, 23, 22, 21, 2, 1, 0, 5, 4, 3, 8, 7, 6, 11, 10, 9}},
{tensor{2, 4, 1, 3},
fmt,
std::vector<T>{0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23},
{1, 2},
std::vector<T>{11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, 23, 22, 21, 20, 19, 18, 17, 16, 15, 14, 13, 12}},
{tensor{2, 4, 1, 3},
fmt,
std::vector<T>{0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23},
{0, 1, 2},
std::vector<T>{23, 22, 21, 20, 19, 18, 17, 16, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0}}};
std::move(local_params.begin(), local_params.end(), std::back_inserter(params));
}
//{tensor(batch(8)), format::bfyx, std::vector<T>{0, 1, 2, 3, 4, 5, 6, 7}, {},
// std::vector<T>{0, 1, 2, 3, 4, 5, 6, 7}},
{tensor(batch(8)),
format::bfyx,
std::vector<T>{0, 1, 2, 3, 4, 5, 6, 7},
{0},
std::vector<T>{7, 6, 5, 4, 3, 2, 1, 0}},
{tensor(batch(4), feature(3)),
format::bfyx,
std::vector<T>{0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11},
{0},
std::vector<T>{9, 10, 11, 6, 7, 8, 3, 4, 5, 0, 1, 2}},
{tensor(batch(4), feature(3)),
format::bfyx,
std::vector<T>{0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11},
{1},
std::vector<T>{2, 1, 0, 5, 4, 3, 8, 7, 6, 11, 10, 9}},
{tensor(batch(4), feature(3)),
format::bfyx,
std::vector<T>{0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11},
{0, 1},
std::vector<T>{11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0}},
{tensor{1, 1, 3, 4, 2},
format::bfzyx,
std::vector<T>{0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23},
{2},
std::vector<T>{12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11}},
{tensor{1, 1, 3, 4, 2},
format::bfzyx,
std::vector<T>{0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23},
{3},
std::vector<T>{9, 10, 11, 6, 7, 8, 3, 4, 5, 0, 1, 2, 21, 22, 23, 18, 19, 20, 15, 16, 17, 12, 13, 14}},
{tensor{1, 1, 3, 4, 2},
format::bfzyx,
std::vector<T>{0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23},
{4},
std::vector<T>{2, 1, 0, 5, 4, 3, 8, 7, 6, 11, 10, 9, 14, 13, 12, 17, 16, 15, 20, 19, 18, 23, 22, 21}},
{tensor{2, 4, 1, 3},
format::bfyx,
std::vector<T>{0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23},
{0, 1},
std::vector<T>{21, 22, 23, 18, 19, 20, 15, 16, 17, 12, 13, 14, 9, 10, 11, 6, 7, 8, 3, 4, 5, 0, 1, 2}},
{tensor{
2,
4,
1,
3,
},
format::bfyx,
std::vector<T>{0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23},
{0, 2},
std::vector<T>{14, 13, 12, 17, 16, 15, 20, 19, 18, 23, 22, 21, 2, 1, 0, 5, 4, 3, 8, 7, 6, 11, 10, 9}},
{tensor{2, 4, 1, 3},
format::bfyx,
std::vector<T>{0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23},
{1, 2},
std::vector<T>{11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, 23, 22, 21, 20, 19, 18, 17, 16, 15, 14, 13, 12}},
{tensor{2, 4, 1, 3},
format::bfyx,
std::vector<T>{0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23},
{0, 1, 2},
std::vector<T>{23, 22, 21, 20, 19, 18, 17, 16, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0}}};
for (const auto fmt : five_d_formats) {
std::vector<ReverseParams<T, reverse_mode::index>> local_params{
{tensor{1, 1, 3, 4, 2},
fmt,
std::vector<T>{0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23},
{2},
std::vector<T>{12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11}},
{tensor{1, 1, 3, 4, 2},
fmt,
std::vector<T>{0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23},
{3},
std::vector<T>{9, 10, 11, 6, 7, 8, 3, 4, 5, 0, 1, 2, 21, 22, 23, 18, 19, 20, 15, 16, 17, 12, 13, 14}},
{tensor{1, 1, 3, 4, 2},
fmt,
std::vector<T>{0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23},
{4},
std::vector<T>{2, 1, 0, 5, 4, 3, 8, 7, 6, 11, 10, 9, 14, 13, 12, 17, 16, 15, 20, 19, 18, 23, 22, 21}},
};
std::move(local_params.begin(), local_params.end(), std::back_inserter(params));
}
return params;
}
template <>
std::vector<ReverseParams<half_t, reverse_mode::mask>> generateMaskParams() {
std::vector<ReverseParams<half_t, reverse_mode::mask>> params{// reverse_2d_1_mask
{tensor(batch(4), feature(3)),
format::bfyx,
std::vector<half_t>{half_t(0),
half_t(1),
half_t(2),
half_t(3),
half_t(4),
half_t(5),
half_t(6),
half_t(7),
half_t(8),
half_t(9),
half_t(10),
half_t(11)},
{false, true},
std::vector<half_t>{half_t(2),
half_t(1),
half_t(0),
half_t(5),
half_t(4),
half_t(3),
half_t(8),
half_t(7),
half_t(6),
half_t(11),
half_t(10),
half_t(9)}}};
std::vector<ReverseParams<half_t, reverse_mode::mask>> params;
for (const auto fmt : four_d_formats) {
// reverse_2d_1_mask
params.push_back({tensor(batch(4), feature(3)),
fmt,
std::vector<half_t>{half_t(0),
half_t(1),
half_t(2),
half_t(3),
half_t(4),
half_t(5),
half_t(6),
half_t(7),
half_t(8),
half_t(9),
half_t(10),
half_t(11)},
{false, true},
std::vector<half_t>{half_t(2),
half_t(1),
half_t(0),
half_t(5),
half_t(4),
half_t(3),
half_t(8),
half_t(7),
half_t(6),
half_t(11),
half_t(10),
half_t(9)}});
}
return params;
}
template <>
std::vector<ReverseParams<half_t, reverse_mode::index>> generateIndexParams() {
std::vector<ReverseParams<half_t, reverse_mode::index>> params{// reverse_2d_1_mask
{tensor(batch(4), feature(3)),
format::bfyx,
std::vector<half_t>{half_t(0),
half_t(1),
half_t(2),
half_t(3),
half_t(4),
half_t(5),
half_t(6),
half_t(7),
half_t(8),
half_t(9),
half_t(10),
half_t(11)},
{1},
std::vector<half_t>{half_t(2),
half_t(1),
half_t(0),
half_t(5),
half_t(4),
half_t(3),
half_t(8),
half_t(7),
half_t(6),
half_t(11),
half_t(10),
half_t(9)}}};
std::vector<ReverseParams<half_t, reverse_mode::index>> params;
for (const auto fmt : four_d_formats) {
// reverse_2d_1_mask
params.push_back({tensor(batch(4), feature(3)),
fmt,
std::vector<half_t>{half_t(0),
half_t(1),
half_t(2),
half_t(3),
half_t(4),
half_t(5),
half_t(6),
half_t(7),
half_t(8),
half_t(9),
half_t(10),
half_t(11)},
{1},
std::vector<half_t>{half_t(2),
half_t(1),
half_t(0),
half_t(5),
half_t(4),
half_t(3),
half_t(8),
half_t(7),
half_t(6),
half_t(11),
half_t(10),
half_t(9)}});
}
return params;
}

31
src/plugins/intel_gpu/tests/test_utils/test_utils.cpp
View File

@@ -388,6 +388,37 @@ double default_tolerance(data_types dt) {
IE_THROW() << "Unknown";
}
cldnn::format generic_test::get_plain_format_for(const cldnn::format input) {
cldnn::format fmt{format::bfzyx};
switch (input) {
case format::b_fs_zyx_fsv16:
case format::b_fs_zyx_fsv32:
case format::bs_fs_zyx_bsv16_fsv32:
case format::bs_fs_zyx_bsv16_fsv16:
case format::bs_fs_zyx_bsv32_fsv32:
case format::bs_fs_zyx_bsv32_fsv16:
fmt = format::bfzyx;
break;
case format::b_fs_yx_fsv16:
case format::b_fs_yx_fsv32:
case format::bs_fs_yx_bsv16_fsv16:
case format::bs_fs_yx_bsv32_fsv16:
case format::bs_fs_yx_bsv32_fsv32:
fmt = format::bfyx;
break;
case format::bfyx:
case format::bfzyx:
case format::bfwzyx:
fmt = input;
break;
default:
throw std::runtime_error(std::string("Unsupported format::" + format(input).to_string()));
break;
}
return fmt;
}
std::vector<cldnn::format> generic_test::test_input_formats = { cldnn::format::bfyx , cldnn::format::yxfb, cldnn::format::fyxb, cldnn::format::byxf };
std::vector<int32_t> generic_test::test_batch_sizes = { 1, 2 };// 4, 8, 16};
std::vector<int32_t> generic_test::test_feature_sizes = { 1, 2 };// , 3, 15};

2
src/plugins/intel_gpu/tests/test_utils/test_utils.h
View File

@@ -468,6 +468,8 @@ public:
}
};
static cldnn::format get_plain_format_for(const cldnn::format);
protected:
cldnn::engine& engine = get_test_engine();
std::shared_ptr<test_params> generic_params;