inference-engine/tests/unit/engines/mkldnn/graph/layers/internal/graph_reshape_test.cpp

   1 // Copyright (C) 2018-2019 Intel Corporation
   2 // SPDX-License-Identifier: Apache-2.0
   3 //
   4
   5 #include <gtest/gtest.h>
   6 #include <gmock/gmock-spec-builders.h>
   7 #include "mkldnn_plugin/mkldnn_graph.h"
   8 #include "single_layer_common.hpp"
   9
  10 #include "test_graph.hpp"
  11
  12 #include <mkldnn_plugin/mkldnn_extension_utils.h>
  13 #include "tests_common.hpp"
  14
  15
  16 using namespace ::testing;
  17 using namespace std;
  18 using namespace mkldnn;
  19
  20
  21 struct reshape_test_params {
  22     InferenceEngine::SizeVector in;
  23     InferenceEngine::SizeVector out;
  24     std::vector<size_t> shape;
  25
  26     int axis;
  27     int num_axes;
  28
  29     size_t num_prim_desc;
  30
  31     MKLDNNPlugin::impl_desc_type selectedType;
  32
  33     std::vector<std::function<void(MKLDNNPlugin::PrimitiveDescInfo)>> comp;
  34 };
  35
  36 template <typename data_t>
  37 void ref_reshape(const InferenceEngine::TBlob<data_t> &src, InferenceEngine::TBlob<data_t> &dst) {
  38     const data_t *src_data = src.readOnly();
  39     data_t *dst_data = dst.data();
  40
  41     for (int i=0; i < src.size(); i++)
  42         dst_data[i] = src_data[i];
  43 }
  44
  45 class MKLDNNGraphReshapeTests: public TestsCommon,
  46                                      public WithParamInterface<reshape_test_params> {
  47     std::string model_t = R"V0G0N(
  48 <Net Name="Reshape_Only" version="2" precision="FP32" batch="1">
  49     <layers>
  50         <layer name="in1" type="Input" precision="FP32" id="0">
  51             <output>
  52                 <port id="0">
  53 __SRC_DIMS__
  54                 </port>
  55             </output>
  56         </layer>
  57         <layer name="norm" id="1" type="Reshape" precision="FP32">
  58             <data dim="_SHAPE_" axis="_AX_" num_axes="_NAX_"/>
  59
  60             <input>
  61                 <port id="1">
  62 __SRC_DIMS__
  63                 </port>
  64             </input>
  65             <output>
  66                 <port id="2">
  67 __DST_DIMS__
  68                 </port>
  69             </output>
  70         </layer>
  71     </layers>
  72     <edges>
  73         <edge from-layer="0" from-port="0" to-layer="1" to-port="1"/>
  74     </edges>
  75 </Net>
  76 )V0G0N";
  77
  78     std::string getModel(reshape_test_params p) {
  79         std::string model = model_t;
  80
  81                 std::string src_dims;
  82                 for (auto& dim : p.in) {
  83                         src_dims += "                    <dim>";
  84                         src_dims += std::to_string(dim) + "</dim>\n";
  85                 }
  86                 REPLACE_WITH_STR(model, "__SRC_DIMS__", src_dims);
  87
  88                 std::string dst_dims;
  89                 for (auto& dim : p.out) {
  90                         dst_dims += "\t\t<dim>";
  91                         dst_dims += std::to_string(dim) + "</dim>\n";
  92                 }
  93                 REPLACE_WITH_STR(model, "__DST_DIMS__", dst_dims);
  94
  95         REPLACE_WITH_NUM(model, "_AX_", p.axis);
  96         REPLACE_WITH_NUM(model, "_NAX_", p.num_axes);
  97
  98         std::string shape_str;
  99         for (auto& dim : p.shape) {
 100             if (!shape_str.empty())
 101                 shape_str += ",";
 102             shape_str += std::to_string(dim);
 103         }
 104         REPLACE_WITH_STR(model, "_SHAPE_", shape_str);
 105         return model;
 106     }
 107
 108 protected:
 109     virtual void TearDown() {
 110     }
 111
 112     virtual void SetUp() {
 113         try {
 114             TestsCommon::SetUp();
 115             reshape_test_params p = ::testing::WithParamInterface<reshape_test_params>::GetParam();
 116             std::string model = getModel(p);
 117
 118             InferenceEngine::CNNNetReader net_reader;
 119             ASSERT_NO_THROW(net_reader.ReadNetwork(model.data(), model.length()));
 120
 121             MKLDNNGraphTestClass graph;
 122             graph.CreateGraph(net_reader.getNetwork());
 123             auto& nodes = graph.getNodes();
 124             for (int i = 0; i < nodes.size(); i++) {
 125                 if (nodes[i]->getType() == MKLDNNPlugin::Reshape) {
 126                     ASSERT_EQ(p.num_prim_desc, nodes[i]->getSupportedPrimitiveDescriptors().size());
 127                     for (size_t j = 0; j < p.num_prim_desc && j < p.comp.size(); j++) {
 128                         p.comp.at(j)(nodes[i]->getSupportedPrimitiveDescriptors().at(j));
 129                     }
 130                     ASSERT_NE(nullptr, nodes[i]->getSelectedPrimitiveDescriptor());
 131                     ASSERT_EQ(p.selectedType, nodes[i]->getSelectedPrimitiveDescriptor()->getImplementationType());
 132                 }
 133             }
 134
 135             InferenceEngine::Blob::Ptr src = InferenceEngine::make_shared_blob<float, const InferenceEngine::SizeVector>(InferenceEngine::Precision::FP32, InferenceEngine::ANY, p.in);
 136             src->allocate();
 137             fill_data(src->buffer(), src->size());
 138
 139             InferenceEngine::TBlob<float>* srcPtr = dynamic_cast<InferenceEngine::TBlob<float>*>(src.get());
 140
 141             if (srcPtr == nullptr)
 142                 FAIL() << "Cannot cast blob to TBlob<float>.";
 143
 144             InferenceEngine::BlobMap srcs;
 145             srcs.insert(std::pair<std::string, InferenceEngine::Blob::Ptr>("in1", src));
 146
 147             InferenceEngine::OutputsDataMap out;
 148             out = net_reader.getNetwork().getOutputsInfo();
 149             InferenceEngine::BlobMap outputBlobs;
 150
 151             std::pair<std::string, InferenceEngine::DataPtr> item = *out.begin();
 152
 153             InferenceEngine::TBlob<float>::Ptr output;
 154             output = InferenceEngine::make_shared_blob<float>(item.second->getTensorDesc());
 155             output->allocate();
 156             outputBlobs[item.first] = output;
 157
 158             graph.Infer(srcs, outputBlobs);
 159
 160             InferenceEngine::TBlob<float> dst_ref(item.second->getTensorDesc());
 161             dst_ref.allocate();
 162
 163             ref_reshape(*srcPtr, dst_ref);
 164
 165             compare(*output, dst_ref);
 166         } catch (const InferenceEngine::details::InferenceEngineException &e) {
 167             FAIL() << e.what();
 168         }
 169     }
 170 };
 171
 172 TEST_P(MKLDNNGraphReshapeTests, TestsReshape) {}
 173
 174
 175 INSTANTIATE_TEST_CASE_P(
 176         TestsReshape, MKLDNNGraphReshapeTests,
 177         ::testing::Values(
 178         reshape_test_params{ {1, 3, 228, 228}, {1, 24, 2, 3249}, {1, 24, 2, 3249}, 0, -1, 1,
 179             MKLDNNPlugin::impl_desc_type::unknown, { [](MKLDNNPlugin::PrimitiveDescInfo impl) {
 180                 ASSERT_EQ(MKLDNNPlugin::impl_desc_type::unknown, impl.getImplementationType());
 181                 ASSERT_EQ(1, impl.getConfig().inConfs.size());
 182                 ASSERT_EQ(1, impl.getConfig().outConfs.size());
 183                 ASSERT_EQ(InferenceEngine::Layout::NCHW, impl.getConfig().inConfs.at(0).desc.getLayout());
 184                 ASSERT_EQ(InferenceEngine::Layout::NCHW, impl.getConfig().outConfs.at(0).desc.getLayout());
 185         } } },
 186         reshape_test_params{ { 4 },{ 2, 2 },{ 2, 2 }, 0, -1, 1,
 187             MKLDNNPlugin::impl_desc_type::unknown,{ [](MKLDNNPlugin::PrimitiveDescInfo impl) {
 188             ASSERT_EQ(MKLDNNPlugin::impl_desc_type::unknown, impl.getImplementationType());
 189             ASSERT_EQ(1, impl.getConfig().inConfs.size());
 190             ASSERT_EQ(1, impl.getConfig().outConfs.size());
 191             ASSERT_EQ(InferenceEngine::Layout::C, impl.getConfig().inConfs.at(0).desc.getLayout());
 192             ASSERT_EQ(InferenceEngine::Layout::NC, impl.getConfig().outConfs.at(0).desc.getLayout());
 193         } } },
 194         reshape_test_params{ { 4 },{ 1, 2, 2 },{ 1, 2, 2 }, 0, -1, 1,
 195             MKLDNNPlugin::impl_desc_type::unknown,{ [](MKLDNNPlugin::PrimitiveDescInfo impl) {
 196             ASSERT_EQ(MKLDNNPlugin::impl_desc_type::unknown, impl.getImplementationType());
 197             ASSERT_EQ(1, impl.getConfig().inConfs.size());
 198             ASSERT_EQ(1, impl.getConfig().outConfs.size());
 199             ASSERT_EQ(InferenceEngine::Layout::C, impl.getConfig().inConfs.at(0).desc.getLayout());
 200             ASSERT_EQ(InferenceEngine::Layout::CHW, impl.getConfig().outConfs.at(0).desc.getLayout());
 201         } } },
 202         reshape_test_params{ { 4 },{ 1, 4, 1, 1 },{ 1, 4, 1, 1 }, 0, -1, 1,
 203             MKLDNNPlugin::impl_desc_type::unknown,{ [](MKLDNNPlugin::PrimitiveDescInfo impl) {
 204             ASSERT_EQ(MKLDNNPlugin::impl_desc_type::unknown, impl.getImplementationType());
 205             ASSERT_EQ(1, impl.getConfig().inConfs.size());
 206             ASSERT_EQ(1, impl.getConfig().outConfs.size());
 207             ASSERT_EQ(InferenceEngine::Layout::C, impl.getConfig().inConfs.at(0).desc.getLayout());
 208             ASSERT_EQ(InferenceEngine::Layout::NCHW, impl.getConfig().outConfs.at(0).desc.getLayout());
 209         } } },
 210         reshape_test_params{ { 4, 4 },{ 1, 4, 4 },{ 1, 4, 4 }, 0, -1, 1,
 211             MKLDNNPlugin::impl_desc_type::unknown,{ [](MKLDNNPlugin::PrimitiveDescInfo impl) {
 212             ASSERT_EQ(MKLDNNPlugin::impl_desc_type::unknown, impl.getImplementationType());
 213             ASSERT_EQ(1, impl.getConfig().inConfs.size());
 214             ASSERT_EQ(1, impl.getConfig().outConfs.size());
 215             ASSERT_EQ(InferenceEngine::Layout::NC, impl.getConfig().inConfs.at(0).desc.getLayout());
 216             ASSERT_EQ(InferenceEngine::Layout::CHW, impl.getConfig().outConfs.at(0).desc.getLayout());
 217         } } },
 218         reshape_test_params{ { 4, 4 },{ 1, 4, 2, 2 },{ 1, 4, 2, 2 }, 0, -1, 1,
 219             MKLDNNPlugin::impl_desc_type::unknown,{ [](MKLDNNPlugin::PrimitiveDescInfo impl) {
 220             ASSERT_EQ(MKLDNNPlugin::impl_desc_type::unknown, impl.getImplementationType());
 221             ASSERT_EQ(1, impl.getConfig().inConfs.size());
 222             ASSERT_EQ(1, impl.getConfig().outConfs.size());
 223             ASSERT_EQ(InferenceEngine::Layout::NC, impl.getConfig().inConfs.at(0).desc.getLayout());
 224             ASSERT_EQ(InferenceEngine::Layout::NCHW, impl.getConfig().outConfs.at(0).desc.getLayout());
 225         } } },
 226         reshape_test_params{ { 4, 2, 2 },{ 1, 4, 2, 2 },{ 1, 4, 2, 2 }, 0, -1, 1,
 227             MKLDNNPlugin::impl_desc_type::unknown,{ [](MKLDNNPlugin::PrimitiveDescInfo impl) {
 228             ASSERT_EQ(MKLDNNPlugin::impl_desc_type::unknown, impl.getImplementationType());
 229             ASSERT_EQ(1, impl.getConfig().inConfs.size());
 230             ASSERT_EQ(1, impl.getConfig().outConfs.size());
 231             ASSERT_EQ(InferenceEngine::Layout::CHW, impl.getConfig().inConfs.at(0).desc.getLayout());
 232             ASSERT_EQ(InferenceEngine::Layout::NCHW, impl.getConfig().outConfs.at(0).desc.getLayout());
 233         } } },
 234         reshape_test_params{ { 2, 2 },{ 4 },{ 4 }, 0, -1, 1,
 235             MKLDNNPlugin::impl_desc_type::unknown,{ [](MKLDNNPlugin::PrimitiveDescInfo impl) {
 236             ASSERT_EQ(MKLDNNPlugin::impl_desc_type::unknown, impl.getImplementationType());
 237             ASSERT_EQ(1, impl.getConfig().inConfs.size());
 238             ASSERT_EQ(1, impl.getConfig().outConfs.size());
 239             ASSERT_EQ(InferenceEngine::Layout::NC, impl.getConfig().inConfs.at(0).desc.getLayout());
 240             ASSERT_EQ(InferenceEngine::Layout::C, impl.getConfig().outConfs.at(0).desc.getLayout());
 241         } } },
 242         reshape_test_params{ { 1, 2, 2 },{ 4 },{ 4 }, 0, -1, 1,
 243             MKLDNNPlugin::impl_desc_type::unknown,{ [](MKLDNNPlugin::PrimitiveDescInfo impl) {
 244             ASSERT_EQ(MKLDNNPlugin::impl_desc_type::unknown, impl.getImplementationType());
 245             ASSERT_EQ(1, impl.getConfig().inConfs.size());
 246             ASSERT_EQ(1, impl.getConfig().outConfs.size());
 247             ASSERT_EQ(InferenceEngine::Layout::CHW, impl.getConfig().inConfs.at(0).desc.getLayout());
 248             ASSERT_EQ(InferenceEngine::Layout::C, impl.getConfig().outConfs.at(0).desc.getLayout());
 249         } } },
 250         reshape_test_params{ { 1, 1, 2, 2 },{ 4 },{ 4 }, 0, -1, 1,
 251             MKLDNNPlugin::impl_desc_type::unknown,{ [](MKLDNNPlugin::PrimitiveDescInfo impl) {
 252             ASSERT_EQ(MKLDNNPlugin::impl_desc_type::unknown, impl.getImplementationType());
 253             ASSERT_EQ(1, impl.getConfig().inConfs.size());
 254             ASSERT_EQ(1, impl.getConfig().outConfs.size());
 255             ASSERT_EQ(InferenceEngine::Layout::NCHW, impl.getConfig().inConfs.at(0).desc.getLayout());
 256             ASSERT_EQ(InferenceEngine::Layout::C, impl.getConfig().outConfs.at(0).desc.getLayout());
 257         } } },
 258         reshape_test_params{ { 4, 2, 2 },{ 4, 4 },{ 4, 4 }, 0, -1, 1,
 259             MKLDNNPlugin::impl_desc_type::unknown,{ [](MKLDNNPlugin::PrimitiveDescInfo impl) {
 260             ASSERT_EQ(MKLDNNPlugin::impl_desc_type::unknown, impl.getImplementationType());
 261             ASSERT_EQ(1, impl.getConfig().inConfs.size());
 262             ASSERT_EQ(1, impl.getConfig().outConfs.size());
 263             ASSERT_EQ(InferenceEngine::Layout::CHW, impl.getConfig().inConfs.at(0).desc.getLayout());
 264             ASSERT_EQ(InferenceEngine::Layout::NC, impl.getConfig().outConfs.at(0).desc.getLayout());
 265         } } },
 266         reshape_test_params{ { 1, 4, 2, 2 },{ 4, 4 },{ 4, 4 }, 0, -1, 1,
 267             MKLDNNPlugin::impl_desc_type::unknown,{ [](MKLDNNPlugin::PrimitiveDescInfo impl) {
 268             ASSERT_EQ(MKLDNNPlugin::impl_desc_type::unknown, impl.getImplementationType());
 269             ASSERT_EQ(1, impl.getConfig().inConfs.size());
 270             ASSERT_EQ(1, impl.getConfig().outConfs.size());
 271             ASSERT_EQ(InferenceEngine::Layout::NCHW, impl.getConfig().inConfs.at(0).desc.getLayout());
 272             ASSERT_EQ(InferenceEngine::Layout::NC, impl.getConfig().outConfs.at(0).desc.getLayout());
 273         } } },
 274         reshape_test_params{ { 1, 4, 2, 2 },{ 4, 2, 2 },{ 4, 2, 2 }, 0, -1, 1,
 275             MKLDNNPlugin::impl_desc_type::unknown,{ [](MKLDNNPlugin::PrimitiveDescInfo impl) {
 276             ASSERT_EQ(MKLDNNPlugin::impl_desc_type::unknown, impl.getImplementationType());
 277             ASSERT_EQ(1, impl.getConfig().inConfs.size());
 278             ASSERT_EQ(1, impl.getConfig().outConfs.size());
 279             ASSERT_EQ(InferenceEngine::Layout::NCHW, impl.getConfig().inConfs.at(0).desc.getLayout());
 280             ASSERT_EQ(InferenceEngine::Layout::CHW, impl.getConfig().outConfs.at(0).desc.getLayout());
 281         } } },
 282         reshape_test_params{ { 1, 4, 2, 2 }, { 4, 2, 2, 1, 1 }, { 4, 2, 2, 1, 1 }, 0, -1, 1,
 283             MKLDNNPlugin::impl_desc_type::unknown, { [](MKLDNNPlugin::PrimitiveDescInfo impl) {
 284             ASSERT_EQ(MKLDNNPlugin::impl_desc_type::unknown, impl.getImplementationType());
 285             ASSERT_EQ(1, impl.getConfig().inConfs.size());
 286             ASSERT_EQ(1, impl.getConfig().outConfs.size());
 287             ASSERT_EQ(InferenceEngine::Layout::NCHW, impl.getConfig().inConfs.at(0).desc.getLayout());
 288             ASSERT_EQ(InferenceEngine::Layout::NCDHW, impl.getConfig().outConfs.at(0).desc.getLayout());
 289         } } },
 290         reshape_test_params{ { 4, 2, 2, 1, 1 }, { 1, 4, 2, 2 }, { 1, 4, 2, 2 }, 0, -1, 1,
 291             MKLDNNPlugin::impl_desc_type::unknown, { [](MKLDNNPlugin::PrimitiveDescInfo impl) {
 292             ASSERT_EQ(MKLDNNPlugin::impl_desc_type::unknown, impl.getImplementationType());
 293             ASSERT_EQ(1, impl.getConfig().inConfs.size());
 294             ASSERT_EQ(1, impl.getConfig().outConfs.size());
 295             ASSERT_EQ(InferenceEngine::Layout::NCDHW, impl.getConfig().inConfs.at(0).desc.getLayout());
 296             ASSERT_EQ(InferenceEngine::Layout::NCHW, impl.getConfig().outConfs.at(0).desc.getLayout());
 297         } } },
 298         reshape_test_params{ { 1, 200 }, { 1, 200, 1, 1, 1 }, { 1, 200, 1, 1, 1 }, 0, -1, 1,
 299             MKLDNNPlugin::impl_desc_type::unknown, { [](MKLDNNPlugin::PrimitiveDescInfo impl) {
 300             ASSERT_EQ(MKLDNNPlugin::impl_desc_type::unknown, impl.getImplementationType());
 301             ASSERT_EQ(1, impl.getConfig().inConfs.size());
 302             ASSERT_EQ(1, impl.getConfig().outConfs.size());
 303             ASSERT_EQ(InferenceEngine::Layout::NC, impl.getConfig().inConfs.at(0).desc.getLayout());
 304             ASSERT_EQ(InferenceEngine::Layout::NCDHW, impl.getConfig().outConfs.at(0).desc.getLayout());
 305         } } }
 306 ));