static void runTorchNet(String prefix, String outLayerName = "",
bool check2ndBlob = false, bool isBinary = false)
{
String suffix = (isBinary) ? ".dat" : ".txt";
Net net;
Ptr<Importer> importer = createTorchImporter(_tf(prefix + "_net" + suffix), isBinary);
ASSERT_TRUE(importer != NULL);
importer->populateNet(net);
Blob inp, outRef;
ASSERT_NO_THROW( inp = readTorchBlob(_tf(prefix + "_input" + suffix), isBinary) );
ASSERT_NO_THROW( outRef = readTorchBlob(_tf(prefix + "_output" + suffix), isBinary) );
net.setBlob(".0", inp);
net.forward();
if (outLayerName.empty())
outLayerName = net.getLayerNames().back();
Blob out = net.getBlob(outLayerName);
normAssert(outRef, out);
if (check2ndBlob)
{
Blob out2 = net.getBlob(outLayerName + ".1");
Blob ref2 = readTorchBlob(_tf(prefix + "_output_2" + suffix), isBinary);
normAssert(out2, ref2);
}
}
TEST(Reproducibility_AlexNet, Accuracy)
{
Net net;
{
const string proto = findDataFile("dnn/bvlc_alexnet.prototxt", false);
const string model = findDataFile("dnn/bvlc_alexnet.caffemodel", false);
Ptr<Importer> importer = createCaffeImporter(proto, model);
ASSERT_TRUE(importer != NULL);
importer->populateNet(net);
}
Mat sample = imread(_tf("grace_hopper_227.png"));
ASSERT_TRUE(!sample.empty());
Size inputSize(227, 227);
if (sample.size() != inputSize)
resize(sample, sample, inputSize);
net.setBlob(".data", blobFromImage(sample, 1.));
net.forward();
Mat out = net.getBlob("prob");
Mat ref = blobFromNPY(_tf("caffe_alexnet_prob.npy"));
normAssert(ref, out);
}
TEST(Reproducibility_FCN, Accuracy)
{
Net net;
{
const string proto = findDataFile("dnn/fcn8s-heavy-pascal.prototxt", false);
const string model = findDataFile("dnn/fcn8s-heavy-pascal.caffemodel", false);
Ptr<Importer> importer = createCaffeImporter(proto, model);
ASSERT_TRUE(importer != NULL);
importer->populateNet(net);
}
Mat sample = imread(_tf("street.png"));
ASSERT_TRUE(!sample.empty());
Size inputSize(500, 500);
if (sample.size() != inputSize)
resize(sample, sample, inputSize);
net.setBlob(".data", blobFromImage(sample, 1.));
net.forward();
Mat out = net.getBlob("score");
Mat ref = blobFromNPY(_tf("caffe_fcn8s_prob.npy"));
normAssert(ref, out);
}
TEST(Reproducibility_AlexNet, Accuracy)
{
Net net;
{
Ptr<Importer> importer = createCaffeImporter(_tf("bvlc_alexnet.prototxt"), _tf("bvlc_alexnet.caffemodel"));
ASSERT_TRUE(importer != NULL);
importer->populateNet(net);
}
Mat sample = imread(_tf("grace_hopper_227.png"));
ASSERT_TRUE(!sample.empty());
cv::cvtColor(sample, sample, cv::COLOR_BGR2RGB);
Size inputSize(227, 227);
if (sample.size() != inputSize)
resize(sample, sample, inputSize);
net.setBlob(".data", dnn::Blob::fromImages(sample));
net.forward();
Blob out = net.getBlob("prob");
Blob ref = blobFromNPY(_tf("caffe_alexnet_prob.npy"));
normAssert(ref, out);
}
TEST(Reproducibility_FCN, Accuracy)
{
Net net;
{
Ptr<Importer> importer = createCaffeImporter(_tf("fcn8s-heavy-pascal.prototxt"), _tf("fcn8s-heavy-pascal.caffemodel"));
ASSERT_TRUE(importer != NULL);
importer->populateNet(net);
}
Mat sample = imread(_tf("street.png"));
ASSERT_TRUE(!sample.empty());
Size inputSize(500, 500);
if (sample.size() != inputSize)
resize(sample, sample, inputSize);
cv::cvtColor(sample, sample, cv::COLOR_BGR2RGB);
net.setBlob(".data", dnn::Blob::fromImages(sample));
net.forward();
Blob out = net.getBlob("score");
Blob ref = blobFromNPY(_tf("caffe_fcn8s_prob.npy"));
normAssert(ref, out);
}
TEST(Torch_Importer, ENet_accuracy)
{
Net net;
{
Ptr<Importer> importer = createTorchImporter(_tf("Enet-model-best.net", false));
ASSERT_TRUE(importer != NULL);
importer->populateNet(net);
}
Mat sample = imread(_tf("street.png", false));
cv::cvtColor(sample, sample, cv::COLOR_BGR2RGB);
sample.convertTo(sample, CV_32F, 1/255.0);
dnn::Blob inputBlob = dnn::Blob::fromImages(sample);
net.setBlob("", inputBlob);
net.forward();
dnn::Blob out = net.getBlob(net.getLayerNames().back());
Blob ref = blobFromNPY(_tf("torch_enet_prob.npy", false));
normAssert(ref, out);
}
