1 /*M///////////////////////////////////////////////////////////////////////////////////////
3 // IMPORTANT: READ BEFORE DOWNLOADING, COPYING, INSTALLING OR USING.
5 // By downloading, copying, installing or using the software you agree to this license.
6 // If you do not agree to this license, do not download, install,
7 // copy or use the software.
11 // For Open Source Computer Vision Library
13 // Copyright (C) 2013, OpenCV Foundation, all rights reserved.
14 // Third party copyrights are property of their respective owners.
16 // Redistribution and use in source and binary forms, with or without modification,
17 // are permitted provided that the following conditions are met:
19 // * Redistribution's of source code must retain the above copyright notice,
20 // this list of conditions and the following disclaimer.
22 // * Redistribution's in binary form must reproduce the above copyright notice,
23 // this list of conditions and the following disclaimer in the documentation
24 // and/or other materials provided with the distribution.
26 // * The name of the copyright holders may not be used to endorse or promote products
27 // derived from this software without specific prior written permission.
29 // This software is provided by the copyright holders and contributors "as is" and
30 // any express or implied warranties, including, but not limited to, the implied
31 // warranties of merchantability and fitness for a particular purpose are disclaimed.
32 // In no event shall the Intel Corporation or contributors be liable for any direct,
33 // indirect, incidental, special, exemplary, or consequential damages
34 // (including, but not limited to, procurement of substitute goods or services;
35 // loss of use, data, or profits; or business interruption) however caused
36 // and on any theory of liability, whether in contract, strict liability,
37 // or tort (including negligence or otherwise) arising in any way out of
38 // the use of this software, even if advised of the possibility of such damage.
42 #include "test_precomp.hpp"
43 #include "npy_blob.hpp"
44 #include <opencv2/dnn/shape_utils.hpp>
45 #include <opencv2/dnn/layer.details.hpp> // CV_DNN_REGISTER_LAYER_CLASS
51 using namespace testing;
53 using namespace cv::dnn;
55 template<typename TStr>
56 static std::string _tf(TStr filename, bool inTorchDir = true, bool required = true)
62 return findDataFile(path, required);
65 TEST(Torch_Importer, simple_read)
68 ASSERT_NO_THROW(net = readNetFromTorch(_tf("net_simple_net.txt"), false));
69 ASSERT_FALSE(net.empty());
72 class Test_Torch_layers : public DNNTestLayer
75 void runTorchNet(const String& prefix, String outLayerName = "",
76 bool check2ndBlob = false, bool isBinary = false, bool evaluate = true,
77 double l1 = 0.0, double lInf = 0.0)
79 String suffix = (isBinary) ? ".dat" : ".txt";
82 ASSERT_NO_THROW( inp = readTorchBlob(_tf(prefix + "_input" + suffix), isBinary) );
83 ASSERT_NO_THROW( outRef = readTorchBlob(_tf(prefix + "_output" + suffix), isBinary) );
85 checkBackend(backend, target, &inp, &outRef);
87 Net net = readNetFromTorch(_tf(prefix + "_net" + suffix), isBinary, evaluate);
88 ASSERT_FALSE(net.empty());
90 net.setPreferableBackend(backend);
91 net.setPreferableTarget(target);
93 if (outLayerName.empty())
94 outLayerName = net.getLayerNames().back();
97 std::vector<Mat> outBlobs;
98 net.forward(outBlobs, outLayerName);
99 l1 = l1 ? l1 : default_l1;
100 lInf = lInf ? lInf : default_lInf;
101 normAssert(outRef, outBlobs[0], "", l1, lInf);
103 if (check2ndBlob && backend == DNN_BACKEND_OPENCV)
105 Mat out2 = outBlobs[1];
106 Mat ref2 = readTorchBlob(_tf(prefix + "_output_2" + suffix), isBinary);
107 normAssert(out2, ref2, "", l1, lInf);
112 TEST_P(Test_Torch_layers, run_convolution)
114 // Output reference values are in range [23.4018, 72.0181]
115 double l1 = default_l1, lInf = default_lInf;
116 if (target == DNN_TARGET_OPENCL_FP16 || target == DNN_TARGET_MYRIAD)
121 else if (target == DNN_TARGET_CUDA_FP16)
126 runTorchNet("net_conv", "", false, true, true, l1, lInf);
129 TEST_P(Test_Torch_layers, run_pool_max)
131 if (backend == DNN_BACKEND_OPENCV && target == DNN_TARGET_OPENCL_FP16)
132 applyTestTag(CV_TEST_TAG_DNN_SKIP_OPENCL_FP16);
133 if (target == DNN_TARGET_CUDA_FP16)
134 applyTestTag(CV_TEST_TAG_DNN_SKIP_CUDA_FP16);
135 double l1 = 0.0, lInf = 0.0;
136 runTorchNet("net_pool_max", "", true, false, true, l1, lInf);
139 TEST_P(Test_Torch_layers, run_pool_ave)
141 runTorchNet("net_pool_ave");
144 TEST_P(Test_Torch_layers, run_reshape_change_batch_size)
146 runTorchNet("net_reshape");
149 TEST_P(Test_Torch_layers, run_reshape)
151 if (backend == DNN_BACKEND_INFERENCE_ENGINE_NN_BUILDER_2019 && target == DNN_TARGET_MYRIAD)
152 applyTestTag(CV_TEST_TAG_DNN_SKIP_IE_MYRIAD, CV_TEST_TAG_DNN_SKIP_IE_NN_BUILDER);
153 runTorchNet("net_reshape_batch");
154 runTorchNet("net_reshape_channels", "", false, true);
157 TEST_P(Test_Torch_layers, run_reshape_single_sample)
159 // Reference output values in range [14.4586, 18.4492].
160 double l1 = default_l1, lInf = default_lInf;
161 if (target == DNN_TARGET_OPENCL_FP16 || target == DNN_TARGET_MYRIAD)
166 else if (target == DNN_TARGET_CUDA_FP16)
170 runTorchNet("net_reshape_single_sample", "", false, false, true, l1, lInf);
173 TEST_P(Test_Torch_layers, run_linear)
175 if (backend == DNN_BACKEND_OPENCV && target == DNN_TARGET_OPENCL_FP16)
176 applyTestTag(CV_TEST_TAG_DNN_SKIP_OPENCL_FP16);
177 runTorchNet("net_linear_2d");
180 TEST_P(Test_Torch_layers, run_concat)
182 runTorchNet("net_concat", "l5_torchMerge");
185 TEST_P(Test_Torch_layers, run_depth_concat)
188 if (target == DNN_TARGET_OPENCL_FP16)
192 else if (target == DNN_TARGET_CUDA_FP16)
196 runTorchNet("net_depth_concat", "", false, true, true, 0.0, lInf);
199 TEST_P(Test_Torch_layers, run_deconv)
201 runTorchNet("net_deconv");
204 TEST_P(Test_Torch_layers, run_batch_norm)
206 runTorchNet("net_batch_norm", "", false, true);
207 runTorchNet("net_batch_norm_train", "", false, true, false);
210 TEST_P(Test_Torch_layers, net_prelu)
212 runTorchNet("net_prelu");
215 TEST_P(Test_Torch_layers, net_cadd_table)
217 runTorchNet("net_cadd_table");
220 TEST_P(Test_Torch_layers, net_softmax)
222 runTorchNet("net_softmax");
223 runTorchNet("net_softmax_spatial");
226 TEST_P(Test_Torch_layers, net_logsoftmax)
228 runTorchNet("net_logsoftmax");
229 runTorchNet("net_logsoftmax_spatial");
232 TEST_P(Test_Torch_layers, net_lp_pooling_square)
234 runTorchNet("net_lp_pooling_square", "", false, true);
236 TEST_P(Test_Torch_layers, net_lp_pooling_power)
238 if (backend == DNN_BACKEND_INFERENCE_ENGINE_NGRAPH && target == DNN_TARGET_MYRIAD)
239 applyTestTag(CV_TEST_TAG_DNN_SKIP_IE_MYRIAD, CV_TEST_TAG_DNN_SKIP_IE_NGRAPH);
240 runTorchNet("net_lp_pooling_power", "", false, true);
243 TEST_P(Test_Torch_layers, net_conv_gemm_lrn)
245 if (backend == DNN_BACKEND_INFERENCE_ENGINE_NN_BUILDER_2019 && target == DNN_TARGET_MYRIAD)
246 applyTestTag(CV_TEST_TAG_DNN_SKIP_IE_MYRIAD, CV_TEST_TAG_DNN_SKIP_IE_NN_BUILDER);
247 if (backend == DNN_BACKEND_INFERENCE_ENGINE_NGRAPH && target == DNN_TARGET_MYRIAD)
248 applyTestTag(CV_TEST_TAG_DNN_SKIP_IE_MYRIAD, CV_TEST_TAG_DNN_SKIP_IE_NGRAPH);
249 double l1 = 0.0, lInf = 0.0;
250 if (target == DNN_TARGET_OPENCL_FP16)
255 else if (target == DNN_TARGET_CUDA_FP16)
260 runTorchNet("net_conv_gemm_lrn", "", false, true, true, l1, lInf);
263 TEST_P(Test_Torch_layers, net_inception_block)
265 runTorchNet("net_inception_block", "", false, true);
268 TEST_P(Test_Torch_layers, net_normalize)
270 if(backend == DNN_BACKEND_CUDA)
271 applyTestTag(CV_TEST_TAG_DNN_SKIP_CUDA); /* only L1 and L2 norms are supported */
272 runTorchNet("net_normalize", "", false, true);
275 TEST_P(Test_Torch_layers, net_padding)
277 runTorchNet("net_padding", "", false, true);
278 runTorchNet("net_spatial_zero_padding", "", false, true);
279 runTorchNet("net_spatial_reflection_padding", "", false, true);
282 TEST_P(Test_Torch_layers, net_non_spatial)
284 if (backend == DNN_BACKEND_INFERENCE_ENGINE_NN_BUILDER_2019 &&
285 (target == DNN_TARGET_OPENCL || target == DNN_TARGET_OPENCL_FP16))
286 applyTestTag(target == DNN_TARGET_OPENCL ? CV_TEST_TAG_DNN_SKIP_IE_OPENCL : CV_TEST_TAG_DNN_SKIP_IE_OPENCL_FP16,
287 CV_TEST_TAG_DNN_SKIP_IE_NN_BUILDER, CV_TEST_TAG_DNN_SKIP_IE_VERSION);
288 runTorchNet("net_non_spatial", "", false, true);
291 TEST_P(Test_Torch_layers, run_paralel)
293 if (backend != DNN_BACKEND_OPENCV || target != DNN_TARGET_CPU)
294 throw SkipTestException(""); // TODO: Check this
295 runTorchNet("net_parallel", "l5_torchMerge");
298 TEST_P(Test_Torch_layers, net_residual)
300 #if defined(INF_ENGINE_RELEASE) && INF_ENGINE_RELEASE == 2018050000
301 if (backend == DNN_BACKEND_INFERENCE_ENGINE_NN_BUILDER_2019 && (target == DNN_TARGET_OPENCL ||
302 target == DNN_TARGET_OPENCL_FP16))
303 applyTestTag(target == DNN_TARGET_OPENCL ? CV_TEST_TAG_DNN_SKIP_IE_OPENCL : CV_TEST_TAG_DNN_SKIP_IE_OPENCL_FP16,
304 CV_TEST_TAG_DNN_SKIP_IE_NN_BUILDER, CV_TEST_TAG_DNN_SKIP_IE_VERSION);
306 runTorchNet("net_residual", "", false, true);
309 class Test_Torch_nets : public DNNTestLayer {};
311 TEST_P(Test_Torch_nets, OpenFace_accuracy)
313 #if defined(INF_ENGINE_RELEASE) && INF_ENGINE_VER_MAJOR_EQ(2018050000)
314 if (backend == DNN_BACKEND_INFERENCE_ENGINE_NN_BUILDER_2019 && target == DNN_TARGET_MYRIAD)
315 applyTestTag(CV_TEST_TAG_DNN_SKIP_IE_MYRIAD, CV_TEST_TAG_DNN_SKIP_IE_NN_BUILDER);
319 const string model = findDataFile("dnn/openface_nn4.small2.v1.t7", false);
320 Net net = readNetFromTorch(model);
322 net.setPreferableBackend(backend);
323 net.setPreferableTarget(target);
325 Mat sample = imread(findDataFile("cv/shared/lena.png"));
326 Mat sampleF32(sample.size(), CV_32FC3);
327 sample.convertTo(sampleF32, sampleF32.type());
329 resize(sampleF32, sampleF32, Size(96, 96), 0, 0, INTER_NEAREST);
331 Mat inputBlob = blobFromImage(sampleF32, 1.0, Size(), Scalar(), /*swapRB*/true);
333 net.setInput(inputBlob);
334 Mat out = net.forward();
336 // Reference output values are in range [-0.17212, 0.263492]
337 // on Myriad problem layer: l4_Pooling - does not use pads_begin
338 float l1 = 1e-5, lInf = 1e-3;
339 if (target == DNN_TARGET_OPENCL_FP16 || target == DNN_TARGET_MYRIAD)
344 else if (target == DNN_TARGET_CUDA_FP16)
349 Mat outRef = readTorchBlob(_tf("net_openface_output.dat"), true);
350 normAssert(out, outRef, "", l1, lInf);
353 static Mat getSegmMask(const Mat& scores)
355 const int rows = scores.size[2];
356 const int cols = scores.size[3];
357 const int numClasses = scores.size[1];
359 Mat maxCl = Mat::zeros(rows, cols, CV_8UC1);
360 Mat maxVal(rows, cols, CV_32FC1, Scalar(0));
361 for (int ch = 0; ch < numClasses; ch++)
363 for (int row = 0; row < rows; row++)
365 const float *ptrScore = scores.ptr<float>(0, ch, row);
366 uint8_t *ptrMaxCl = maxCl.ptr<uint8_t>(row);
367 float *ptrMaxVal = maxVal.ptr<float>(row);
368 for (int col = 0; col < cols; col++)
370 if (ptrScore[col] > ptrMaxVal[col])
372 ptrMaxVal[col] = ptrScore[col];
373 ptrMaxCl[col] = (uchar)ch;
381 // Computer per-class intersection over union metric.
382 static void normAssertSegmentation(const Mat& ref, const Mat& test)
384 CV_Assert_N(ref.dims == 4, test.dims == 4);
385 const int numClasses = ref.size[1];
386 CV_Assert(numClasses == test.size[1]);
388 Mat refMask = getSegmMask(ref);
389 Mat testMask = getSegmMask(test);
390 EXPECT_EQ(countNonZero(refMask != testMask), 0);
393 TEST_P(Test_Torch_nets, ENet_accuracy)
395 applyTestTag(target == DNN_TARGET_CPU ? "" : CV_TEST_TAG_MEMORY_512MB);
397 if (backend == DNN_BACKEND_OPENCV && target == DNN_TARGET_OPENCL_FP16)
398 throw SkipTestException("");
399 if (backend == DNN_BACKEND_CUDA && target == DNN_TARGET_CUDA_FP16)
400 applyTestTag(CV_TEST_TAG_DNN_SKIP_CUDA_FP16);
401 #if defined(INF_ENGINE_RELEASE) && INF_ENGINE_VER_MAJOR_EQ(2020010000)
402 if (backend == DNN_BACKEND_INFERENCE_ENGINE_NN_BUILDER_2019)
403 applyTestTag(CV_TEST_TAG_DNN_SKIP_IE_NN_BUILDER, CV_TEST_TAG_DNN_SKIP_IE_VERSION);
405 if (backend == DNN_BACKEND_INFERENCE_ENGINE_NN_BUILDER_2019 && target != DNN_TARGET_CPU)
407 if (target == DNN_TARGET_OPENCL_FP16) applyTestTag(CV_TEST_TAG_DNN_SKIP_IE_OPENCL_FP16, CV_TEST_TAG_DNN_SKIP_IE_NN_BUILDER);
408 if (target == DNN_TARGET_OPENCL) applyTestTag(CV_TEST_TAG_DNN_SKIP_IE_OPENCL, CV_TEST_TAG_DNN_SKIP_IE_NN_BUILDER);
409 if (target == DNN_TARGET_MYRIAD) applyTestTag(CV_TEST_TAG_DNN_SKIP_IE_MYRIAD, CV_TEST_TAG_DNN_SKIP_IE_NN_BUILDER);
410 throw SkipTestException("");
413 #if defined(INF_ENGINE_RELEASE) && INF_ENGINE_VER_MAJOR_GE(2021010000)
414 if (backend == DNN_BACKEND_INFERENCE_ENGINE_NGRAPH)
415 applyTestTag(CV_TEST_TAG_DNN_SKIP_IE_NGRAPH);
417 if (backend == DNN_BACKEND_INFERENCE_ENGINE_NGRAPH && target != DNN_TARGET_CPU)
419 if (target == DNN_TARGET_OPENCL_FP16) applyTestTag(CV_TEST_TAG_DNN_SKIP_IE_OPENCL_FP16, CV_TEST_TAG_DNN_SKIP_IE_NGRAPH);
420 if (target == DNN_TARGET_OPENCL) applyTestTag(CV_TEST_TAG_DNN_SKIP_IE_OPENCL, CV_TEST_TAG_DNN_SKIP_IE_NGRAPH);
421 if (target == DNN_TARGET_MYRIAD) applyTestTag(CV_TEST_TAG_DNN_SKIP_IE_MYRIAD, CV_TEST_TAG_DNN_SKIP_IE_NGRAPH);
422 throw SkipTestException("");
427 const string model = findDataFile("dnn/Enet-model-best.net", false);
428 net = readNetFromTorch(model, true);
429 ASSERT_TRUE(!net.empty());
432 net.setPreferableBackend(backend);
433 net.setPreferableTarget(target);
435 Mat sample = imread(_tf("street.png", false));
436 Mat inputBlob = blobFromImage(sample, 1./255, Size(), Scalar(), /*swapRB*/true);
438 net.setInput(inputBlob, "");
439 Mat out = net.forward();
440 Mat ref = blobFromNPY(_tf("torch_enet_prob.npy", false));
441 // Due to numerical instability in Pooling-Unpooling layers (indexes jittering)
442 // thresholds for ENet must be changed. Accuracy of results was checked on
443 // Cityscapes dataset and difference in mIOU with Torch is 10E-4%
444 normAssert(ref, out, "", 0.00044, /*target == DNN_TARGET_CPU ? 0.453 : */0.552);
445 normAssertSegmentation(ref, out);
448 for (int i = 0; i < N; i++)
450 net.setInput(inputBlob, "");
451 Mat out = net.forward();
452 normAssert(ref, out, "", 0.00044, /*target == DNN_TARGET_CPU ? 0.453 : */0.552);
453 normAssertSegmentation(ref, out);
457 // Check accuracy of style transfer models from https://github.com/jcjohnson/fast-neural-style
458 // th fast_neural_style.lua \
459 // -input_image ~/opencv_extra/testdata/dnn/googlenet_1.png \
460 // -output_image lena.png \
461 // -median_filter 0 \
463 // -model models/eccv16/starry_night.t7
464 // th fast_neural_style.lua \
465 // -input_image ~/opencv_extra/testdata/dnn/googlenet_1.png \
466 // -output_image lena.png \
467 // -median_filter 0 \
469 // -model models/instance_norm/feathers.t7
470 TEST_P(Test_Torch_nets, FastNeuralStyle_accuracy)
472 #if defined INF_ENGINE_RELEASE
473 if (backend == DNN_BACKEND_INFERENCE_ENGINE_NN_BUILDER_2019 && target == DNN_TARGET_MYRIAD
474 && getInferenceEngineVPUType() == CV_DNN_INFERENCE_ENGINE_VPU_TYPE_MYRIAD_X)
475 applyTestTag(CV_TEST_TAG_DNN_SKIP_IE_MYRIAD_X, CV_TEST_TAG_DNN_SKIP_IE_NN_BUILDER);
476 if (backend == DNN_BACKEND_INFERENCE_ENGINE_NGRAPH && target == DNN_TARGET_MYRIAD
477 && getInferenceEngineVPUType() == CV_DNN_INFERENCE_ENGINE_VPU_TYPE_MYRIAD_X)
478 applyTestTag(CV_TEST_TAG_DNN_SKIP_IE_MYRIAD_X, CV_TEST_TAG_DNN_SKIP_IE_NGRAPH);
483 #if defined(INF_ENGINE_RELEASE)
484 #if INF_ENGINE_RELEASE <= 2018050000
485 if (backend == DNN_BACKEND_INFERENCE_ENGINE_NN_BUILDER_2019 && target == DNN_TARGET_OPENCL)
486 applyTestTag(CV_TEST_TAG_DNN_SKIP_IE_OPENCL, CV_TEST_TAG_DNN_SKIP_IE_NN_BUILDER, CV_TEST_TAG_DNN_SKIP_IE_VERSION);
490 std::string models[] = {"dnn/fast_neural_style_eccv16_starry_night.t7",
491 "dnn/fast_neural_style_instance_norm_feathers.t7"};
492 std::string targets[] = {"dnn/lena_starry_night.png", "dnn/lena_feathers.png"};
494 for (int i = 0; i < 2; ++i)
496 const string model = findDataFile(models[i], false);
497 Net net = readNetFromTorch(model);
499 net.setPreferableBackend(backend);
500 net.setPreferableTarget(target);
502 Mat img = imread(findDataFile("dnn/googlenet_1.png"));
503 Mat inputBlob = blobFromImage(img, 1.0, Size(), Scalar(103.939, 116.779, 123.68), false);
505 net.setInput(inputBlob);
506 Mat out = net.forward();
509 getPlane(out, 0, 0) += 103.939;
510 getPlane(out, 0, 1) += 116.779;
511 getPlane(out, 0, 2) += 123.68;
512 out = cv::min(cv::max(0, out), 255);
514 Mat ref = imread(findDataFile(targets[i]));
515 Mat refBlob = blobFromImage(ref, 1.0, Size(), Scalar(), false);
517 if (target == DNN_TARGET_OPENCL_FP16 || target == DNN_TARGET_MYRIAD)
519 double normL1 = cvtest::norm(refBlob, out, cv::NORM_L1) / refBlob.total();
520 if (target == DNN_TARGET_MYRIAD)
521 EXPECT_LE(normL1, 4.0f);
523 EXPECT_LE(normL1, 0.6f);
525 else if(target == DNN_TARGET_CUDA_FP16)
527 normAssert(out, refBlob, "", 0.6, 25);
530 normAssert(out, refBlob, "", 0.5, 1.1);
534 INSTANTIATE_TEST_CASE_P(/**/, Test_Torch_nets, dnnBackendsAndTargets());
536 // Test a custom layer
537 // https://github.com/torch/nn/blob/master/doc/convolution.md#nn.SpatialUpSamplingNearest
538 class SpatialUpSamplingNearestLayer CV_FINAL : public Layer
541 SpatialUpSamplingNearestLayer(const LayerParams ¶ms) : Layer(params)
543 scale = params.get<int>("scale_factor");
546 static Ptr<Layer> create(LayerParams& params)
548 return Ptr<Layer>(new SpatialUpSamplingNearestLayer(params));
551 virtual bool getMemoryShapes(const std::vector<std::vector<int> > &inputs,
552 const int requiredOutputs,
553 std::vector<std::vector<int> > &outputs,
554 std::vector<std::vector<int> > &internals) const CV_OVERRIDE
556 std::vector<int> outShape(4);
557 outShape[0] = inputs[0][0]; // batch size
558 outShape[1] = inputs[0][1]; // number of channels
559 outShape[2] = scale * inputs[0][2];
560 outShape[3] = scale * inputs[0][3];
561 outputs.assign(1, outShape);
565 void forward(InputArrayOfArrays inputs_arr, OutputArrayOfArrays outputs_arr, OutputArrayOfArrays) CV_OVERRIDE
568 CV_TRACE_ARG_VALUE(name, "name", name.c_str());
570 std::vector<Mat> inputs, outputs;
571 inputs_arr.getMatVector(inputs);
572 outputs_arr.getMatVector(outputs);
574 Mat& inp = inputs[0];
575 Mat& out = outputs[0];
576 const int outHeight = out.size[2];
577 const int outWidth = out.size[3];
578 for (size_t n = 0; n < inp.size[0]; ++n)
580 for (size_t ch = 0; ch < inp.size[1]; ++ch)
582 resize(getPlane(inp, n, ch), getPlane(out, n, ch),
583 Size(outWidth, outHeight), 0, 0, INTER_NEAREST);
592 TEST_P(Test_Torch_layers, upsampling_nearest)
594 // Test a custom layer.
595 CV_DNN_REGISTER_LAYER_CLASS(SpatialUpSamplingNearest, SpatialUpSamplingNearestLayer);
598 runTorchNet("net_spatial_upsampling_nearest", "", false, true);
602 LayerFactory::unregisterLayer("SpatialUpSamplingNearest");
605 LayerFactory::unregisterLayer("SpatialUpSamplingNearest");
607 // Test an implemented layer.
608 runTorchNet("net_spatial_upsampling_nearest", "", false, true);
611 INSTANTIATE_TEST_CASE_P(/**/, Test_Torch_layers, dnnBackendsAndTargets());