mmr/tkDNN
Archived
1
0
Fork 0
Code Issues 46 Pull Requests 12 Releases 4 Activity

namespace change

Browse Source
This commit is contained in:
Francesco Gatti
2018-12-14 21:55:16 +01:00
parent 443179359d
commit 6249956469
32 changed files with 374 additions and 374 deletions
Download Patch File Download Diff File Expand all files Collapse all files
tests/mnist/test_mnist.cpp
+12 -12
View File
@@ -11,18 +11,18 @@ const char *output_bin = "../tests/mnist/output.bin";
int main() {
// Network layout
tkDNN::dataDim_t dim(1, 1, 28, 28, 1);
tkDNN::Network net(dim);
tkDNN::Conv2d l0(&net, 20, 5, 5, 1, 1, 0, 0, c0_bin);
tkDNN::Pooling l1(&net, 2, 2, 2, 2, tkDNN::POOLING_MAX);
tkDNN::Conv2d l2(&net, 50, 5, 5, 1, 1, 0, 0, c1_bin);
tkDNN::Pooling l3(&net, 2, 2, 2, 2, tkDNN::POOLING_MAX);
tkDNN::Dense l4(&net, 500, d2_bin);
tkDNN::Activation l5(&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Dense l6(&net, 10, d3_bin);
tkDNN::Softmax l7(&net);
tk::dnn::dataDim_t dim(1, 1, 28, 28, 1);
tk::dnn::Network net(dim);
tk::dnn::Conv2d l0(&net, 20, 5, 5, 1, 1, 0, 0, c0_bin);
tk::dnn::Pooling l1(&net, 2, 2, 2, 2, tk::dnn::POOLING_MAX);
tk::dnn::Conv2d l2(&net, 50, 5, 5, 1, 1, 0, 0, c1_bin);
tk::dnn::Pooling l3(&net, 2, 2, 2, 2, tk::dnn::POOLING_MAX);
tk::dnn::Dense l4(&net, 500, d2_bin);
tk::dnn::Activation l5(&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Dense l6(&net, 10, d3_bin);
tk::dnn::Softmax l7(&net);
tkDNN::NetworkRT netRT(&net, "mnist.rt");
tk::dnn::NetworkRT netRT(&net, "mnist.rt");
// Load input
dnnType *data;
@@ -43,7 +43,7 @@ int main() {
//std::cout<<"\n======= CUDNN RESULT =======\n";
//printDeviceVector(10, out_data);
tkDNN::dataDim_t dim2(1, 1, 28, 28, 1);
tk::dnn::dataDim_t dim2(1, 1, 28, 28, 1);
std::cout<<"TENSORRT inference:\n"; {
dim2.print();
tests/mnist/test_mnistRT.cpp
+14 -14
View File
@@ -27,16 +27,16 @@ int main() {
std::cout<<"\n==== CUDNN ====\n";
// Network layout
tkDNN::dataDim_t dim(1, 1, 28, 28, 1);
tkDNN::Network net(dim);
tkDNN::Conv2d l0(&net, 20, 5, 5, 1, 1, 0, 0, c0_bin);
tkDNN::Pooling l1(&net, 2, 2, 2, 2, tkDNN::POOLING_MAX);
tkDNN::Conv2d l2(&net, 50, 5, 5, 1, 1, 0, 0, c1_bin);
tkDNN::Pooling l3(&net, 2, 2, 2, 2, tkDNN::POOLING_MAX);
tkDNN::Dense l4(&net, 500, d2_bin);
tkDNN::Activation l5(&net, CUDNN_ACTIVATION_RELU);
tkDNN::Dense l6(&net, 10, d3_bin);
tkDNN::Softmax l7(&net);
tk::dnn::dataDim_t dim(1, 1, 28, 28, 1);
tk::dnn::Network net(dim);
tk::dnn::Conv2d l0(&net, 20, 5, 5, 1, 1, 0, 0, c0_bin);
tk::dnn::Pooling l1(&net, 2, 2, 2, 2, tk::dnn::POOLING_MAX);
tk::dnn::Conv2d l2(&net, 50, 5, 5, 1, 1, 0, 0, c1_bin);
tk::dnn::Pooling l3(&net, 2, 2, 2, 2, tk::dnn::POOLING_MAX);
tk::dnn::Dense l4(&net, 500, d2_bin);
tk::dnn::Activation l5(&net, CUDNN_ACTIVATION_RELU);
tk::dnn::Dense l6(&net, 10, d3_bin);
tk::dnn::Softmax l7(&net);
// Load input
dnnType *data;
@@ -71,7 +71,7 @@ int main() {
auto input = network->addInput("data", dt, DimsCHW{ 1, 28, 28});
assert(input != nullptr);
tkDNN::Conv2d *c0 = &l0;
tk::dnn::Conv2d *c0 = &l0;
Weights w { dt, c0->data_h, c0->inputs*c0->outputs*c0->kernelH*c0->kernelW};
Weights b { dt, c0->bias_h, c0->outputs};
// Add a convolution layer with 20 outputs and a 5x5 filter.
@@ -84,7 +84,7 @@ int main() {
assert(pool1 != nullptr);
pool1->setStride(DimsHW{2, 2});
tkDNN::Conv2d *c1 = &l2;
tk::dnn::Conv2d *c1 = &l2;
Weights w1 { dt, c1->data_h, c1->inputs*c1->outputs*c1->kernelH*c1->kernelW};
Weights b1 { dt, c1->bias_h, c1->outputs};
// Add a second convolution layer with 50 outputs and a 5x5 filter.
@@ -97,7 +97,7 @@ int main() {
assert(pool2 != nullptr);
pool2->setStride(DimsHW{2, 2});
tkDNN::Dense *d2 = &l4;
tk::dnn::Dense *d2 = &l4;
Weights w2 { dt, d2->data_h, d2->inputs*d2->outputs};
Weights b2 { dt, d2->bias_h, d2->outputs};
// Add a fully connected layer with 500 outputs.
@@ -108,7 +108,7 @@ int main() {
auto relu1 = network->addActivation(*ip1->getOutput(0), ActivationType::kRELU);
assert(relu1 != nullptr);
tkDNN::Dense *d3 = &l6;
tk::dnn::Dense *d3 = &l6;
Weights w3 { dt, d3->data_h, d3->inputs*d3->outputs};
Weights b3 { dt, d3->bias_h, d3->outputs};
// Add a second fully connected layer with 20 outputs.
tests/simple/test_simple.cpp
+9 -9
View File
@@ -10,15 +10,15 @@ const char *output_bin = "../tests/simple/output.bin";
int main() {
// Network layout
tkDNN::dataDim_t dim(1, 1, 10, 10, 1);
tkDNN::Network net(dim);
tkDNN::Conv2d l0(&net, 2, 4, 4, 2, 2, 0, 0, c0_bin);
tkDNN::Activation l1(&net, CUDNN_ACTIVATION_RELU);
tkDNN::Conv2d l2(&net, 4, 2, 2, 1, 1, 0, 0, c1_bin);
tkDNN::Activation l3(&net, CUDNN_ACTIVATION_RELU);
tkDNN::Flatten l4(&net);
tkDNN::Dense l5(&net, 4, d2_bin);
tkDNN::Activation l6(&net, CUDNN_ACTIVATION_RELU);
tk::dnn::dataDim_t dim(1, 1, 10, 10, 1);
tk::dnn::Network net(dim);
tk::dnn::Conv2d l0(&net, 2, 4, 4, 2, 2, 0, 0, c0_bin);
tk::dnn::Activation l1(&net, CUDNN_ACTIVATION_RELU);
tk::dnn::Conv2d l2(&net, 4, 2, 2, 1, 1, 0, 0, c1_bin);
tk::dnn::Activation l3(&net, CUDNN_ACTIVATION_RELU);
tk::dnn::Flatten l4(&net);
tk::dnn::Dense l5(&net, 4, d2_bin);
tk::dnn::Activation l6(&net, CUDNN_ACTIVATION_RELU);
// Load input
dnnType *data;
tests/test_rtinference/rtinference.cpp
+1 -1
View File
@@ -11,7 +11,7 @@ int main(int argc, char *argv[]) {
srand (0);
//convert network to tensorRT
tkDNN::NetworkRT netRT(NULL, argv[1]);
tk::dnn::NetworkRT netRT(NULL, argv[1]);
dnnType *input = new float[netRT.input_dim.tot()];
dnnType *output = new float[netRT.input_dim.tot()];
tests/yolo/yolo.cpp
+62 -62
View File
@@ -31,75 +31,75 @@ const char *output_bin = "../tests/yolo/layers/output.bin";
int main() {
// Network layout
tkDNN::dataDim_t dim(1, 3, 608, 608, 1);
tkDNN::Network net(dim);
tk::dnn::dataDim_t dim(1, 3, 608, 608, 1);
tk::dnn::Network net(dim);
tkDNN::Conv2d c0 (&net, 32, 3, 3, 1, 1, 1, 1, c0_bin, true);
tkDNN::Activation a0 (&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Pooling p1 (&net, 2, 2, 2, 2, tkDNN::POOLING_MAX);
tk::dnn::Conv2d c0 (&net, 32, 3, 3, 1, 1, 1, 1, c0_bin, true);
tk::dnn::Activation a0 (&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Pooling p1 (&net, 2, 2, 2, 2, tk::dnn::POOLING_MAX);
tkDNN::Conv2d c2 (&net, 64, 3, 3, 1, 1, 1, 1, c2_bin, true);
tkDNN::Activation a2 (&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Pooling p3 (&net, 2, 2, 2, 2, tkDNN::POOLING_MAX);
tk::dnn::Conv2d c2 (&net, 64, 3, 3, 1, 1, 1, 1, c2_bin, true);
tk::dnn::Activation a2 (&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Pooling p3 (&net, 2, 2, 2, 2, tk::dnn::POOLING_MAX);
tkDNN::Conv2d c4 (&net, 128, 3, 3, 1, 1, 1, 1, c4_bin, true);
tkDNN::Activation a4 (&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Conv2d c5 (&net, 64, 1, 1, 1, 1, 0, 0, c5_bin, true);
tkDNN::Activation a5 (&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Conv2d c6 (&net, 128, 3, 3, 1, 1, 1, 1, c6_bin, true);
tkDNN::Activation a6 (&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Pooling p7 (&net, 2, 2, 2, 2, tkDNN::POOLING_MAX);
tk::dnn::Conv2d c4 (&net, 128, 3, 3, 1, 1, 1, 1, c4_bin, true);
tk::dnn::Activation a4 (&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Conv2d c5 (&net, 64, 1, 1, 1, 1, 0, 0, c5_bin, true);
tk::dnn::Activation a5 (&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Conv2d c6 (&net, 128, 3, 3, 1, 1, 1, 1, c6_bin, true);
tk::dnn::Activation a6 (&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Pooling p7 (&net, 2, 2, 2, 2, tk::dnn::POOLING_MAX);
tkDNN::Conv2d c8 (&net, 256, 3, 3, 1, 1, 1, 1, c8_bin, true);
tkDNN::Activation a8 (&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Conv2d c9 (&net, 128, 1, 1, 1, 1, 0, 0, c9_bin, true);
tkDNN::Activation a9 (&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Conv2d c10(&net, 256, 3, 3, 1, 1, 1, 1, c10_bin, true);
tkDNN::Activation a10(&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Pooling p11(&net, 2, 2, 2, 2, tkDNN::POOLING_MAX);
tk::dnn::Conv2d c8 (&net, 256, 3, 3, 1, 1, 1, 1, c8_bin, true);
tk::dnn::Activation a8 (&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Conv2d c9 (&net, 128, 1, 1, 1, 1, 0, 0, c9_bin, true);
tk::dnn::Activation a9 (&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Conv2d c10(&net, 256, 3, 3, 1, 1, 1, 1, c10_bin, true);
tk::dnn::Activation a10(&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Pooling p11(&net, 2, 2, 2, 2, tk::dnn::POOLING_MAX);
tkDNN::Conv2d c12(&net, 512, 3, 3, 1, 1, 1, 1, c12_bin, true);
tkDNN::Activation a12(&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Conv2d c13(&net, 256, 1, 1, 1, 1, 0, 0, c13_bin, true);
tkDNN::Activation a13(&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Conv2d c14(&net, 512, 3, 3, 1, 1, 1, 1, c14_bin, true);
tkDNN::Activation a14(&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Conv2d c15(&net, 256, 1, 1, 1, 1, 0, 0, c15_bin, true);
tkDNN::Activation a15(&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Conv2d c16(&net, 512, 3, 3, 1, 1, 1, 1, c16_bin, true);
tkDNN::Activation a16(&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Pooling p17(&net, 2, 2, 2, 2, tkDNN::POOLING_MAX);
tk::dnn::Conv2d c12(&net, 512, 3, 3, 1, 1, 1, 1, c12_bin, true);
tk::dnn::Activation a12(&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Conv2d c13(&net, 256, 1, 1, 1, 1, 0, 0, c13_bin, true);
tk::dnn::Activation a13(&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Conv2d c14(&net, 512, 3, 3, 1, 1, 1, 1, c14_bin, true);
tk::dnn::Activation a14(&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Conv2d c15(&net, 256, 1, 1, 1, 1, 0, 0, c15_bin, true);
tk::dnn::Activation a15(&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Conv2d c16(&net, 512, 3, 3, 1, 1, 1, 1, c16_bin, true);
tk::dnn::Activation a16(&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Pooling p17(&net, 2, 2, 2, 2, tk::dnn::POOLING_MAX);
tkDNN::Conv2d c18(&net, 1024, 3, 3, 1, 1, 1, 1, c18_bin, true);
tkDNN::Activation a18(&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Conv2d c19(&net, 512, 1, 1, 1, 1, 0, 0, c19_bin, true);
tkDNN::Activation a19(&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Conv2d c20(&net, 1024, 3, 3, 1, 1, 1, 1, c20_bin, true);
tkDNN::Activation a20(&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Conv2d c21(&net, 512, 1, 1, 1, 1, 0, 0, c21_bin, true);
tkDNN::Activation a21(&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Conv2d c22(&net, 1024, 3, 3, 1, 1, 1, 1, c22_bin, true);
tkDNN::Activation a22(&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Conv2d c23(&net, 1024, 3, 3, 1, 1, 1, 1, c23_bin, true);
tkDNN::Activation a23(&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Conv2d c24(&net, 1024, 3, 3, 1, 1, 1, 1, c24_bin, true);
tkDNN::Activation a24(&net, tkDNN::ACTIVATION_LEAKY);
tk::dnn::Conv2d c18(&net, 1024, 3, 3, 1, 1, 1, 1, c18_bin, true);
tk::dnn::Activation a18(&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Conv2d c19(&net, 512, 1, 1, 1, 1, 0, 0, c19_bin, true);
tk::dnn::Activation a19(&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Conv2d c20(&net, 1024, 3, 3, 1, 1, 1, 1, c20_bin, true);
tk::dnn::Activation a20(&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Conv2d c21(&net, 512, 1, 1, 1, 1, 0, 0, c21_bin, true);
tk::dnn::Activation a21(&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Conv2d c22(&net, 1024, 3, 3, 1, 1, 1, 1, c22_bin, true);
tk::dnn::Activation a22(&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Conv2d c23(&net, 1024, 3, 3, 1, 1, 1, 1, c23_bin, true);
tk::dnn::Activation a23(&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Conv2d c24(&net, 1024, 3, 3, 1, 1, 1, 1, c24_bin, true);
tk::dnn::Activation a24(&net, tk::dnn::ACTIVATION_LEAKY);
tkDNN::Layer *m25_layers[1] = { &a16 };
tkDNN::Route m25(&net, m25_layers, 1);
tkDNN::Conv2d c26(&net, 64, 1, 1, 1, 1, 0, 0, c26_bin, true);
tkDNN::Activation a26(&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Reorg r27(&net, 2);
tk::dnn::Layer *m25_layers[1] = { &a16 };
tk::dnn::Route m25(&net, m25_layers, 1);
tk::dnn::Conv2d c26(&net, 64, 1, 1, 1, 1, 0, 0, c26_bin, true);
tk::dnn::Activation a26(&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Reorg r27(&net, 2);
tkDNN::Layer *m28_layers[2] = { &r27, &a24 };
tkDNN::Route m28(&net, m28_layers, 2);
tk::dnn::Layer *m28_layers[2] = { &r27, &a24 };
tk::dnn::Route m28(&net, m28_layers, 2);
tkDNN::Conv2d c29(&net, 1024, 3, 3, 1, 1, 1, 1, c29_bin, true);
tkDNN::Activation a29(&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Conv2d c30(&net, 425, 1, 1, 1, 1, 0, 0, c30_bin, false);
tkDNN::Region g31(&net, 80, 4, 5);
tk::dnn::Conv2d c29(&net, 1024, 3, 3, 1, 1, 1, 1, c29_bin, true);
tk::dnn::Activation a29(&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Conv2d c30(&net, 425, 1, 1, 1, 1, 0, 0, c30_bin, false);
tk::dnn::Region g31(&net, 80, 4, 5);
tkDNN::RegionInterpret rI(dim, g31.output_dim, 80, 4, 5, 0.6f, g31_bin);
tk::dnn::RegionInterpret rI(dim, g31.output_dim, 80, 4, 5, 0.6f, g31_bin);
// Load input
dnnType *data;
@@ -110,11 +110,11 @@ int main() {
net.print();
//convert network to tensorRT
tkDNN::NetworkRT netRT(&net, "yolo.rt");
tk::dnn::NetworkRT netRT(&net, "yolo.rt");
dnnType *out_data, *out_data2; // cudnn output, tensorRT output
tkDNN::dataDim_t dim1 = dim; //input dim
tk::dnn::dataDim_t dim1 = dim; //input dim
printCenteredTitle(" CUDNN inference ", '=', 30); {
dim1.print();
TIMER_START
@@ -123,7 +123,7 @@ int main() {
dim1.print();
}
tkDNN::dataDim_t dim2 = dim;
tk::dnn::dataDim_t dim2 = dim;
printCenteredTitle(" TENSORRT inference ", '=', 30); {
dim2.print();
TIMER_START
tests/yolo_224/yolo_224.cpp
+62 -62
View File
@@ -31,75 +31,75 @@ const char *output_bin = "../tests/yolo_224/layers/output.bin";
int main() {
// Network layout
tkDNN::dataDim_t dim(1, 3, 224, 224, 1);
tkDNN::Network net(dim);
tk::dnn::dataDim_t dim(1, 3, 224, 224, 1);
tk::dnn::Network net(dim);
tkDNN::Conv2d c0 (&net, 32, 3, 3, 1, 1, 1, 1, c0_bin, true);
tkDNN::Activation a0 (&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Pooling p1 (&net, 2, 2, 2, 2, tkDNN::POOLING_MAX);
tk::dnn::Conv2d c0 (&net, 32, 3, 3, 1, 1, 1, 1, c0_bin, true);
tk::dnn::Activation a0 (&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Pooling p1 (&net, 2, 2, 2, 2, tk::dnn::POOLING_MAX);
tkDNN::Conv2d c2 (&net, 64, 3, 3, 1, 1, 1, 1, c2_bin, true);
tkDNN::Activation a2 (&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Pooling p3 (&net, 2, 2, 2, 2, tkDNN::POOLING_MAX);
tk::dnn::Conv2d c2 (&net, 64, 3, 3, 1, 1, 1, 1, c2_bin, true);
tk::dnn::Activation a2 (&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Pooling p3 (&net, 2, 2, 2, 2, tk::dnn::POOLING_MAX);
tkDNN::Conv2d c4 (&net, 128, 3, 3, 1, 1, 1, 1, c4_bin, true);
tkDNN::Activation a4 (&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Conv2d c5 (&net, 64, 1, 1, 1, 1, 0, 0, c5_bin, true);
tkDNN::Activation a5 (&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Conv2d c6 (&net, 128, 3, 3, 1, 1, 1, 1, c6_bin, true);
tkDNN::Activation a6 (&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Pooling p7 (&net, 2, 2, 2, 2, tkDNN::POOLING_MAX);
tk::dnn::Conv2d c4 (&net, 128, 3, 3, 1, 1, 1, 1, c4_bin, true);
tk::dnn::Activation a4 (&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Conv2d c5 (&net, 64, 1, 1, 1, 1, 0, 0, c5_bin, true);
tk::dnn::Activation a5 (&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Conv2d c6 (&net, 128, 3, 3, 1, 1, 1, 1, c6_bin, true);
tk::dnn::Activation a6 (&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Pooling p7 (&net, 2, 2, 2, 2, tk::dnn::POOLING_MAX);
tkDNN::Conv2d c8 (&net, 256, 3, 3, 1, 1, 1, 1, c8_bin, true);
tkDNN::Activation a8 (&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Conv2d c9 (&net, 128, 1, 1, 1, 1, 0, 0, c9_bin, true);
tkDNN::Activation a9 (&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Conv2d c10(&net, 256, 3, 3, 1, 1, 1, 1, c10_bin, true);
tkDNN::Activation a10(&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Pooling p11(&net, 2, 2, 2, 2, tkDNN::POOLING_MAX);
tk::dnn::Conv2d c8 (&net, 256, 3, 3, 1, 1, 1, 1, c8_bin, true);
tk::dnn::Activation a8 (&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Conv2d c9 (&net, 128, 1, 1, 1, 1, 0, 0, c9_bin, true);
tk::dnn::Activation a9 (&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Conv2d c10(&net, 256, 3, 3, 1, 1, 1, 1, c10_bin, true);
tk::dnn::Activation a10(&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Pooling p11(&net, 2, 2, 2, 2, tk::dnn::POOLING_MAX);
tkDNN::Conv2d c12(&net, 512, 3, 3, 1, 1, 1, 1, c12_bin, true);
tkDNN::Activation a12(&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Conv2d c13(&net, 256, 1, 1, 1, 1, 0, 0, c13_bin, true);
tkDNN::Activation a13(&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Conv2d c14(&net, 512, 3, 3, 1, 1, 1, 1, c14_bin, true);
tkDNN::Activation a14(&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Conv2d c15(&net, 256, 1, 1, 1, 1, 0, 0, c15_bin, true);
tkDNN::Activation a15(&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Conv2d c16(&net, 512, 3, 3, 1, 1, 1, 1, c16_bin, true);
tkDNN::Activation a16(&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Pooling p17(&net, 2, 2, 2, 2, tkDNN::POOLING_MAX);
tk::dnn::Conv2d c12(&net, 512, 3, 3, 1, 1, 1, 1, c12_bin, true);
tk::dnn::Activation a12(&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Conv2d c13(&net, 256, 1, 1, 1, 1, 0, 0, c13_bin, true);
tk::dnn::Activation a13(&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Conv2d c14(&net, 512, 3, 3, 1, 1, 1, 1, c14_bin, true);
tk::dnn::Activation a14(&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Conv2d c15(&net, 256, 1, 1, 1, 1, 0, 0, c15_bin, true);
tk::dnn::Activation a15(&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Conv2d c16(&net, 512, 3, 3, 1, 1, 1, 1, c16_bin, true);
tk::dnn::Activation a16(&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Pooling p17(&net, 2, 2, 2, 2, tk::dnn::POOLING_MAX);
tkDNN::Conv2d c18(&net, 1024, 3, 3, 1, 1, 1, 1, c18_bin, true);
tkDNN::Activation a18(&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Conv2d c19(&net, 512, 1, 1, 1, 1, 0, 0, c19_bin, true);
tkDNN::Activation a19(&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Conv2d c20(&net, 1024, 3, 3, 1, 1, 1, 1, c20_bin, true);
tkDNN::Activation a20(&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Conv2d c21(&net, 512, 1, 1, 1, 1, 0, 0, c21_bin, true);
tkDNN::Activation a21(&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Conv2d c22(&net, 1024, 3, 3, 1, 1, 1, 1, c22_bin, true);
tkDNN::Activation a22(&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Conv2d c23(&net, 1024, 3, 3, 1, 1, 1, 1, c23_bin, true);
tkDNN::Activation a23(&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Conv2d c24(&net, 1024, 3, 3, 1, 1, 1, 1, c24_bin, true);
tkDNN::Activation a24(&net, tkDNN::ACTIVATION_LEAKY);
tk::dnn::Conv2d c18(&net, 1024, 3, 3, 1, 1, 1, 1, c18_bin, true);
tk::dnn::Activation a18(&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Conv2d c19(&net, 512, 1, 1, 1, 1, 0, 0, c19_bin, true);
tk::dnn::Activation a19(&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Conv2d c20(&net, 1024, 3, 3, 1, 1, 1, 1, c20_bin, true);
tk::dnn::Activation a20(&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Conv2d c21(&net, 512, 1, 1, 1, 1, 0, 0, c21_bin, true);
tk::dnn::Activation a21(&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Conv2d c22(&net, 1024, 3, 3, 1, 1, 1, 1, c22_bin, true);
tk::dnn::Activation a22(&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Conv2d c23(&net, 1024, 3, 3, 1, 1, 1, 1, c23_bin, true);
tk::dnn::Activation a23(&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Conv2d c24(&net, 1024, 3, 3, 1, 1, 1, 1, c24_bin, true);
tk::dnn::Activation a24(&net, tk::dnn::ACTIVATION_LEAKY);
tkDNN::Layer *m25_layers[1] = { &a16 };
tkDNN::Route m25(&net, m25_layers, 1);
tkDNN::Conv2d c26(&net, 64, 1, 1, 1, 1, 0, 0, c26_bin, true);
tkDNN::Activation a26(&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Reorg r27(&net, 2);
tk::dnn::Layer *m25_layers[1] = { &a16 };
tk::dnn::Route m25(&net, m25_layers, 1);
tk::dnn::Conv2d c26(&net, 64, 1, 1, 1, 1, 0, 0, c26_bin, true);
tk::dnn::Activation a26(&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Reorg r27(&net, 2);
tkDNN::Layer *m28_layers[2] = { &r27, &a24 };
tkDNN::Route m28(&net, m28_layers, 2);
tk::dnn::Layer *m28_layers[2] = { &r27, &a24 };
tk::dnn::Route m28(&net, m28_layers, 2);
tkDNN::Conv2d c29(&net, 1024, 3, 3, 1, 1, 1, 1, c29_bin, true);
tkDNN::Activation a29(&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Conv2d c30(&net, 425, 1, 1, 1, 1, 0, 0, c30_bin, false);
tkDNN::Region g31(&net, 80, 4, 5);
tk::dnn::Conv2d c29(&net, 1024, 3, 3, 1, 1, 1, 1, c29_bin, true);
tk::dnn::Activation a29(&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Conv2d c30(&net, 425, 1, 1, 1, 1, 0, 0, c30_bin, false);
tk::dnn::Region g31(&net, 80, 4, 5);
tkDNN::RegionInterpret rI(dim, g31.output_dim, 80, 4, 5, 0.6f, g31_bin);
tk::dnn::RegionInterpret rI(dim, g31.output_dim, 80, 4, 5, 0.6f, g31_bin);
// Load input
dnnType *data;
@@ -110,11 +110,11 @@ int main() {
net.print();
//convert network to tensorRT
tkDNN::NetworkRT netRT(&net, "yolo_224.rt");
tk::dnn::NetworkRT netRT(&net, "yolo_224.rt");
dnnType *out_data, *out_data2; // cudnn output, tensorRT output
tkDNN::dataDim_t dim1 = dim; //input dim
tk::dnn::dataDim_t dim1 = dim; //input dim
printCenteredTitle(" CUDNN inference ", '=', 30); {
dim1.print();
TIMER_START
@@ -123,7 +123,7 @@ int main() {
dim1.print();
}
tkDNN::dataDim_t dim2 = dim;
tk::dnn::dataDim_t dim2 = dim;
printCenteredTitle(" TENSORRT inference ", '=', 30); {
dim2.print();
TIMER_START
tests/yolo_relu/yolo_relu.cpp
+62 -62
View File
@@ -31,75 +31,75 @@ const char *output_bin = "../tests/yolo_relu/layers/output.bin";
int main() {
// Network layout
tkDNN::dataDim_t dim(1, 3, 608, 608, 1);
tkDNN::Network net(dim);
tk::dnn::dataDim_t dim(1, 3, 608, 608, 1);
tk::dnn::Network net(dim);
tkDNN::Conv2d c0 (&net, 32, 3, 3, 1, 1, 1, 1, c0_bin, true);
tkDNN::Activation a0 (&net, CUDNN_ACTIVATION_RELU);
tkDNN::Pooling p1 (&net, 2, 2, 2, 2, tkDNN::POOLING_MAX);
tk::dnn::Conv2d c0 (&net, 32, 3, 3, 1, 1, 1, 1, c0_bin, true);
tk::dnn::Activation a0 (&net, CUDNN_ACTIVATION_RELU);
tk::dnn::Pooling p1 (&net, 2, 2, 2, 2, tk::dnn::POOLING_MAX);
tkDNN::Conv2d c2 (&net, 64, 3, 3, 1, 1, 1, 1, c2_bin, true);
tkDNN::Activation a2 (&net, CUDNN_ACTIVATION_RELU);
tkDNN::Pooling p3 (&net, 2, 2, 2, 2, tkDNN::POOLING_MAX);
tk::dnn::Conv2d c2 (&net, 64, 3, 3, 1, 1, 1, 1, c2_bin, true);
tk::dnn::Activation a2 (&net, CUDNN_ACTIVATION_RELU);
tk::dnn::Pooling p3 (&net, 2, 2, 2, 2, tk::dnn::POOLING_MAX);
tkDNN::Conv2d c4 (&net, 128, 3, 3, 1, 1, 1, 1, c4_bin, true);
tkDNN::Activation a4 (&net, CUDNN_ACTIVATION_RELU);
tkDNN::Conv2d c5 (&net, 64, 1, 1, 1, 1, 0, 0, c5_bin, true);
tkDNN::Activation a5 (&net, CUDNN_ACTIVATION_RELU);
tkDNN::Conv2d c6 (&net, 128, 3, 3, 1, 1, 1, 1, c6_bin, true);
tkDNN::Activation a6 (&net, CUDNN_ACTIVATION_RELU);
tkDNN::Pooling p7 (&net, 2, 2, 2, 2, tkDNN::POOLING_MAX);
tk::dnn::Conv2d c4 (&net, 128, 3, 3, 1, 1, 1, 1, c4_bin, true);
tk::dnn::Activation a4 (&net, CUDNN_ACTIVATION_RELU);
tk::dnn::Conv2d c5 (&net, 64, 1, 1, 1, 1, 0, 0, c5_bin, true);
tk::dnn::Activation a5 (&net, CUDNN_ACTIVATION_RELU);
tk::dnn::Conv2d c6 (&net, 128, 3, 3, 1, 1, 1, 1, c6_bin, true);
tk::dnn::Activation a6 (&net, CUDNN_ACTIVATION_RELU);
tk::dnn::Pooling p7 (&net, 2, 2, 2, 2, tk::dnn::POOLING_MAX);
tkDNN::Conv2d c8 (&net, 256, 3, 3, 1, 1, 1, 1, c8_bin, true);
tkDNN::Activation a8 (&net, CUDNN_ACTIVATION_RELU);
tkDNN::Conv2d c9 (&net, 128, 1, 1, 1, 1, 0, 0, c9_bin, true);
tkDNN::Activation a9 (&net, CUDNN_ACTIVATION_RELU);
tkDNN::Conv2d c10(&net, 256, 3, 3, 1, 1, 1, 1, c10_bin, true);
tkDNN::Activation a10(&net, CUDNN_ACTIVATION_RELU);
tkDNN::Pooling p11(&net, 2, 2, 2, 2, tkDNN::POOLING_MAX);
tk::dnn::Conv2d c8 (&net, 256, 3, 3, 1, 1, 1, 1, c8_bin, true);
tk::dnn::Activation a8 (&net, CUDNN_ACTIVATION_RELU);
tk::dnn::Conv2d c9 (&net, 128, 1, 1, 1, 1, 0, 0, c9_bin, true);
tk::dnn::Activation a9 (&net, CUDNN_ACTIVATION_RELU);
tk::dnn::Conv2d c10(&net, 256, 3, 3, 1, 1, 1, 1, c10_bin, true);
tk::dnn::Activation a10(&net, CUDNN_ACTIVATION_RELU);
tk::dnn::Pooling p11(&net, 2, 2, 2, 2, tk::dnn::POOLING_MAX);
tkDNN::Conv2d c12(&net, 512, 3, 3, 1, 1, 1, 1, c12_bin, true);
tkDNN::Activation a12(&net, CUDNN_ACTIVATION_RELU);
tkDNN::Conv2d c13(&net, 256, 1, 1, 1, 1, 0, 0, c13_bin, true);
tkDNN::Activation a13(&net, CUDNN_ACTIVATION_RELU);
tkDNN::Conv2d c14(&net, 512, 3, 3, 1, 1, 1, 1, c14_bin, true);
tkDNN::Activation a14(&net, CUDNN_ACTIVATION_RELU);
tkDNN::Conv2d c15(&net, 256, 1, 1, 1, 1, 0, 0, c15_bin, true);
tkDNN::Activation a15(&net, CUDNN_ACTIVATION_RELU);
tkDNN::Conv2d c16(&net, 512, 3, 3, 1, 1, 1, 1, c16_bin, true);
tkDNN::Activation a16(&net, CUDNN_ACTIVATION_RELU);
tkDNN::Pooling p17(&net, 2, 2, 2, 2, tkDNN::POOLING_MAX);
tk::dnn::Conv2d c12(&net, 512, 3, 3, 1, 1, 1, 1, c12_bin, true);
tk::dnn::Activation a12(&net, CUDNN_ACTIVATION_RELU);
tk::dnn::Conv2d c13(&net, 256, 1, 1, 1, 1, 0, 0, c13_bin, true);
tk::dnn::Activation a13(&net, CUDNN_ACTIVATION_RELU);
tk::dnn::Conv2d c14(&net, 512, 3, 3, 1, 1, 1, 1, c14_bin, true);
tk::dnn::Activation a14(&net, CUDNN_ACTIVATION_RELU);
tk::dnn::Conv2d c15(&net, 256, 1, 1, 1, 1, 0, 0, c15_bin, true);
tk::dnn::Activation a15(&net, CUDNN_ACTIVATION_RELU);
tk::dnn::Conv2d c16(&net, 512, 3, 3, 1, 1, 1, 1, c16_bin, true);
tk::dnn::Activation a16(&net, CUDNN_ACTIVATION_RELU);
tk::dnn::Pooling p17(&net, 2, 2, 2, 2, tk::dnn::POOLING_MAX);
tkDNN::Conv2d c18(&net, 1024, 3, 3, 1, 1, 1, 1, c18_bin, true);
tkDNN::Activation a18(&net, CUDNN_ACTIVATION_RELU);
tkDNN::Conv2d c19(&net, 512, 1, 1, 1, 1, 0, 0, c19_bin, true);
tkDNN::Activation a19(&net, CUDNN_ACTIVATION_RELU);
tkDNN::Conv2d c20(&net, 1024, 3, 3, 1, 1, 1, 1, c20_bin, true);
tkDNN::Activation a20(&net, CUDNN_ACTIVATION_RELU);
tkDNN::Conv2d c21(&net, 512, 1, 1, 1, 1, 0, 0, c21_bin, true);
tkDNN::Activation a21(&net, CUDNN_ACTIVATION_RELU);
tkDNN::Conv2d c22(&net, 1024, 3, 3, 1, 1, 1, 1, c22_bin, true);
tkDNN::Activation a22(&net, CUDNN_ACTIVATION_RELU);
tkDNN::Conv2d c23(&net, 1024, 3, 3, 1, 1, 1, 1, c23_bin, true);
tkDNN::Activation a23(&net, CUDNN_ACTIVATION_RELU);
tkDNN::Conv2d c24(&net, 1024, 3, 3, 1, 1, 1, 1, c24_bin, true);
tkDNN::Activation a24(&net, CUDNN_ACTIVATION_RELU);
tk::dnn::Conv2d c18(&net, 1024, 3, 3, 1, 1, 1, 1, c18_bin, true);
tk::dnn::Activation a18(&net, CUDNN_ACTIVATION_RELU);
tk::dnn::Conv2d c19(&net, 512, 1, 1, 1, 1, 0, 0, c19_bin, true);
tk::dnn::Activation a19(&net, CUDNN_ACTIVATION_RELU);
tk::dnn::Conv2d c20(&net, 1024, 3, 3, 1, 1, 1, 1, c20_bin, true);
tk::dnn::Activation a20(&net, CUDNN_ACTIVATION_RELU);
tk::dnn::Conv2d c21(&net, 512, 1, 1, 1, 1, 0, 0, c21_bin, true);
tk::dnn::Activation a21(&net, CUDNN_ACTIVATION_RELU);
tk::dnn::Conv2d c22(&net, 1024, 3, 3, 1, 1, 1, 1, c22_bin, true);
tk::dnn::Activation a22(&net, CUDNN_ACTIVATION_RELU);
tk::dnn::Conv2d c23(&net, 1024, 3, 3, 1, 1, 1, 1, c23_bin, true);
tk::dnn::Activation a23(&net, CUDNN_ACTIVATION_RELU);
tk::dnn::Conv2d c24(&net, 1024, 3, 3, 1, 1, 1, 1, c24_bin, true);
tk::dnn::Activation a24(&net, CUDNN_ACTIVATION_RELU);
tkDNN::Layer *m25_layers[1] = { &a16 };
tkDNN::Route m25(&net, m25_layers, 1);
tkDNN::Conv2d c26(&net, 64, 1, 1, 1, 1, 0, 0, c26_bin, true);
tkDNN::Activation a26(&net, CUDNN_ACTIVATION_RELU);
tkDNN::Reorg r27(&net, 2);
tk::dnn::Layer *m25_layers[1] = { &a16 };
tk::dnn::Route m25(&net, m25_layers, 1);
tk::dnn::Conv2d c26(&net, 64, 1, 1, 1, 1, 0, 0, c26_bin, true);
tk::dnn::Activation a26(&net, CUDNN_ACTIVATION_RELU);
tk::dnn::Reorg r27(&net, 2);
tkDNN::Layer *m28_layers[2] = { &r27, &a24 };
tkDNN::Route m28(&net, m28_layers, 2);
tk::dnn::Layer *m28_layers[2] = { &r27, &a24 };
tk::dnn::Route m28(&net, m28_layers, 2);
tkDNN::Conv2d c29(&net, 1024, 3, 3, 1, 1, 1, 1, c29_bin, true);
tkDNN::Activation a29(&net, CUDNN_ACTIVATION_RELU);
tkDNN::Conv2d c30(&net, 425, 1, 1, 1, 1, 0, 0, c30_bin, false);
tkDNN::Region g31(&net, 80, 4, 5);
tk::dnn::Conv2d c29(&net, 1024, 3, 3, 1, 1, 1, 1, c29_bin, true);
tk::dnn::Activation a29(&net, CUDNN_ACTIVATION_RELU);
tk::dnn::Conv2d c30(&net, 425, 1, 1, 1, 1, 0, 0, c30_bin, false);
tk::dnn::Region g31(&net, 80, 4, 5);
tkDNN::RegionInterpret rI(dim, g31.output_dim, 80, 4, 5, 0.3f, g31_bin);
tk::dnn::RegionInterpret rI(dim, g31.output_dim, 80, 4, 5, 0.3f, g31_bin);
// Load input
dnnType *data;
@@ -110,11 +110,11 @@ int main() {
net.print();
//convert network to tensorRT
tkDNN::NetworkRT netRT(&net, "yolo_relu.rt");
tk::dnn::NetworkRT netRT(&net, "yolo_relu.rt");
dnnType *out_data, *out_data2; // cudnn output, tensorRT output
tkDNN::dataDim_t dim1 = dim; //input dim
tk::dnn::dataDim_t dim1 = dim; //input dim
printCenteredTitle(" CUDNN inference ", '=', 30); {
dim1.print();
TIMER_START
@@ -123,7 +123,7 @@ int main() {
dim1.print();
}
tkDNN::dataDim_t dim2 = dim;
tk::dnn::dataDim_t dim2 = dim;
printCenteredTitle(" TENSORRT inference ", '=', 30); {
dim2.print();
TIMER_START
tests/yolo_tiny/yolo_tiny.cpp
+28 -28
View File
@@ -18,38 +18,38 @@ const char *output_bin = "../tests/yolo_tiny/layers/output.bin";
int main() {
// Network layout
tkDNN::dataDim_t dim(1, 3, 416, 416, 1);
tkDNN::Network net(dim);
tk::dnn::dataDim_t dim(1, 3, 416, 416, 1);
tk::dnn::Network net(dim);
tkDNN::Conv2d c0 (&net, 16, 3, 3, 1, 1, 1, 1, c0_bin, true);
tkDNN::Activation a0 (&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Pooling p1 (&net, 2, 2, 2, 2, tkDNN::POOLING_MAX);
tk::dnn::Conv2d c0 (&net, 16, 3, 3, 1, 1, 1, 1, c0_bin, true);
tk::dnn::Activation a0 (&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Pooling p1 (&net, 2, 2, 2, 2, tk::dnn::POOLING_MAX);
tkDNN::Conv2d c2 (&net, 32, 3, 3, 1, 1, 1, 1, c2_bin, true);
tkDNN::Activation a2 (&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Pooling p3 (&net, 2, 2, 2, 2, tkDNN::POOLING_MAX);
tk::dnn::Conv2d c2 (&net, 32, 3, 3, 1, 1, 1, 1, c2_bin, true);
tk::dnn::Activation a2 (&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Pooling p3 (&net, 2, 2, 2, 2, tk::dnn::POOLING_MAX);
tkDNN::Conv2d c4 (&net, 64, 3, 3, 1, 1, 1, 1, c4_bin, true);
tkDNN::Activation a4 (&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Pooling p5 (&net, 2, 2, 2, 2, tkDNN::POOLING_MAX);
tk::dnn::Conv2d c4 (&net, 64, 3, 3, 1, 1, 1, 1, c4_bin, true);
tk::dnn::Activation a4 (&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Pooling p5 (&net, 2, 2, 2, 2, tk::dnn::POOLING_MAX);
tkDNN::Conv2d c6 (&net, 128, 3, 3, 1, 1, 1, 1, c6_bin, true);
tkDNN::Activation a6 (&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Pooling p7(&net, 2, 2, 2, 2, tkDNN::POOLING_MAX);
tk::dnn::Conv2d c6 (&net, 128, 3, 3, 1, 1, 1, 1, c6_bin, true);
tk::dnn::Activation a6 (&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Pooling p7(&net, 2, 2, 2, 2, tk::dnn::POOLING_MAX);
tkDNN::Conv2d c8(&net, 256, 3, 3, 1, 1, 1, 1, c8_bin, true);
tkDNN::Activation a8(&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Pooling p9(&net, 2, 2, 2, 2, tkDNN::POOLING_MAX);
tk::dnn::Conv2d c8(&net, 256, 3, 3, 1, 1, 1, 1, c8_bin, true);
tk::dnn::Activation a8(&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Pooling p9(&net, 2, 2, 2, 2, tk::dnn::POOLING_MAX);
tkDNN::Conv2d c10(&net, 512, 3, 3, 1, 1, 1, 1, c10_bin, true);
tkDNN::Activation a10(&net, tkDNN::ACTIVATION_LEAKY);
tk::dnn::Conv2d c10(&net, 512, 3, 3, 1, 1, 1, 1, c10_bin, true);
tk::dnn::Activation a10(&net, tk::dnn::ACTIVATION_LEAKY);
tkDNN::Conv2d c11(&net, 1024, 3, 3, 1, 1, 1, 1, c11_bin, true);
tkDNN::Activation a11(&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Conv2d c12(&net, 512, 3, 3, 1, 1, 1, 1, c12_bin, true);
tkDNN::Activation a12(&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Conv2d c13(&net, 425, 1, 1, 1, 1, 0, 0, c13_bin, false);
tkDNN::Region g14(&net, 80, 4, 5);
tk::dnn::Conv2d c11(&net, 1024, 3, 3, 1, 1, 1, 1, c11_bin, true);
tk::dnn::Activation a11(&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Conv2d c12(&net, 512, 3, 3, 1, 1, 1, 1, c12_bin, true);
tk::dnn::Activation a12(&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Conv2d c13(&net, 425, 1, 1, 1, 1, 0, 0, c13_bin, false);
tk::dnn::Region g14(&net, 80, 4, 5);
// Load input
dnnType *data;
@@ -60,11 +60,11 @@ int main() {
net.print();
//convert network to tensorRT
tkDNN::NetworkRT netRT(&net, "yolo_tiny.rt");
tk::dnn::NetworkRT netRT(&net, "yolo_tiny.rt");
dnnType *out_data, *out_data2; // cudnn output, tensorRT output
tkDNN::dataDim_t dim1 = dim; //input dim
tk::dnn::dataDim_t dim1 = dim; //input dim
printCenteredTitle(" CUDNN inference ", '=', 30); {
dim1.print();
TIMER_START
@@ -73,7 +73,7 @@ int main() {
dim1.print();
}
tkDNN::dataDim_t dim2 = dim;
tk::dnn::dataDim_t dim2 = dim;
printCenteredTitle(" TENSORRT inference ", '=', 30); {
dim2.print();
TIMER_START
tests/yolo_voc/yolo_voc.cpp
+62 -62
View File
@@ -31,75 +31,75 @@ const char *output_bin = "../tests/yolo_voc/layers/output.bin";
int main() {
// Network layout
tkDNN::dataDim_t dim(1, 3, 416, 416, 1);
tkDNN::Network net(dim);
tk::dnn::dataDim_t dim(1, 3, 416, 416, 1);
tk::dnn::Network net(dim);
tkDNN::Conv2d c0 (&net, 32, 3, 3, 1, 1, 1, 1, c0_bin, true);
tkDNN::Activation a0 (&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Pooling p1 (&net, 2, 2, 2, 2, tkDNN::POOLING_MAX);
tk::dnn::Conv2d c0 (&net, 32, 3, 3, 1, 1, 1, 1, c0_bin, true);
tk::dnn::Activation a0 (&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Pooling p1 (&net, 2, 2, 2, 2, tk::dnn::POOLING_MAX);
tkDNN::Conv2d c2 (&net, 64, 3, 3, 1, 1, 1, 1, c2_bin, true);
tkDNN::Activation a2 (&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Pooling p3 (&net, 2, 2, 2, 2, tkDNN::POOLING_MAX);
tk::dnn::Conv2d c2 (&net, 64, 3, 3, 1, 1, 1, 1, c2_bin, true);
tk::dnn::Activation a2 (&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Pooling p3 (&net, 2, 2, 2, 2, tk::dnn::POOLING_MAX);
tkDNN::Conv2d c4 (&net, 128, 3, 3, 1, 1, 1, 1, c4_bin, true);
tkDNN::Activation a4 (&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Conv2d c5 (&net, 64, 1, 1, 1, 1, 0, 0, c5_bin, true);
tkDNN::Activation a5 (&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Conv2d c6 (&net, 128, 3, 3, 1, 1, 1, 1, c6_bin, true);
tkDNN::Activation a6 (&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Pooling p7 (&net, 2, 2, 2, 2, tkDNN::POOLING_MAX);
tk::dnn::Conv2d c4 (&net, 128, 3, 3, 1, 1, 1, 1, c4_bin, true);
tk::dnn::Activation a4 (&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Conv2d c5 (&net, 64, 1, 1, 1, 1, 0, 0, c5_bin, true);
tk::dnn::Activation a5 (&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Conv2d c6 (&net, 128, 3, 3, 1, 1, 1, 1, c6_bin, true);
tk::dnn::Activation a6 (&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Pooling p7 (&net, 2, 2, 2, 2, tk::dnn::POOLING_MAX);
tkDNN::Conv2d c8 (&net, 256, 3, 3, 1, 1, 1, 1, c8_bin, true);
tkDNN::Activation a8 (&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Conv2d c9 (&net, 128, 1, 1, 1, 1, 0, 0, c9_bin, true);
tkDNN::Activation a9 (&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Conv2d c10(&net, 256, 3, 3, 1, 1, 1, 1, c10_bin, true);
tkDNN::Activation a10(&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Pooling p11(&net, 2, 2, 2, 2, tkDNN::POOLING_MAX);
tk::dnn::Conv2d c8 (&net, 256, 3, 3, 1, 1, 1, 1, c8_bin, true);
tk::dnn::Activation a8 (&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Conv2d c9 (&net, 128, 1, 1, 1, 1, 0, 0, c9_bin, true);
tk::dnn::Activation a9 (&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Conv2d c10(&net, 256, 3, 3, 1, 1, 1, 1, c10_bin, true);
tk::dnn::Activation a10(&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Pooling p11(&net, 2, 2, 2, 2, tk::dnn::POOLING_MAX);
tkDNN::Conv2d c12(&net, 512, 3, 3, 1, 1, 1, 1, c12_bin, true);
tkDNN::Activation a12(&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Conv2d c13(&net, 256, 1, 1, 1, 1, 0, 0, c13_bin, true);
tkDNN::Activation a13(&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Conv2d c14(&net, 512, 3, 3, 1, 1, 1, 1, c14_bin, true);
tkDNN::Activation a14(&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Conv2d c15(&net, 256, 1, 1, 1, 1, 0, 0, c15_bin, true);
tkDNN::Activation a15(&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Conv2d c16(&net, 512, 3, 3, 1, 1, 1, 1, c16_bin, true);
tkDNN::Activation a16(&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Pooling p17(&net, 2, 2, 2, 2, tkDNN::POOLING_MAX);
tk::dnn::Conv2d c12(&net, 512, 3, 3, 1, 1, 1, 1, c12_bin, true);
tk::dnn::Activation a12(&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Conv2d c13(&net, 256, 1, 1, 1, 1, 0, 0, c13_bin, true);
tk::dnn::Activation a13(&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Conv2d c14(&net, 512, 3, 3, 1, 1, 1, 1, c14_bin, true);
tk::dnn::Activation a14(&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Conv2d c15(&net, 256, 1, 1, 1, 1, 0, 0, c15_bin, true);
tk::dnn::Activation a15(&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Conv2d c16(&net, 512, 3, 3, 1, 1, 1, 1, c16_bin, true);
tk::dnn::Activation a16(&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Pooling p17(&net, 2, 2, 2, 2, tk::dnn::POOLING_MAX);
tkDNN::Conv2d c18(&net, 1024, 3, 3, 1, 1, 1, 1, c18_bin, true);
tkDNN::Activation a18(&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Conv2d c19(&net, 512, 1, 1, 1, 1, 0, 0, c19_bin, true);
tkDNN::Activation a19(&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Conv2d c20(&net, 1024, 3, 3, 1, 1, 1, 1, c20_bin, true);
tkDNN::Activation a20(&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Conv2d c21(&net, 512, 1, 1, 1, 1, 0, 0, c21_bin, true);
tkDNN::Activation a21(&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Conv2d c22(&net, 1024, 3, 3, 1, 1, 1, 1, c22_bin, true);
tkDNN::Activation a22(&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Conv2d c23(&net, 1024, 3, 3, 1, 1, 1, 1, c23_bin, true);
tkDNN::Activation a23(&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Conv2d c24(&net, 1024, 3, 3, 1, 1, 1, 1, c24_bin, true);
tkDNN::Activation a24(&net, tkDNN::ACTIVATION_LEAKY);
tk::dnn::Conv2d c18(&net, 1024, 3, 3, 1, 1, 1, 1, c18_bin, true);
tk::dnn::Activation a18(&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Conv2d c19(&net, 512, 1, 1, 1, 1, 0, 0, c19_bin, true);
tk::dnn::Activation a19(&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Conv2d c20(&net, 1024, 3, 3, 1, 1, 1, 1, c20_bin, true);
tk::dnn::Activation a20(&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Conv2d c21(&net, 512, 1, 1, 1, 1, 0, 0, c21_bin, true);
tk::dnn::Activation a21(&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Conv2d c22(&net, 1024, 3, 3, 1, 1, 1, 1, c22_bin, true);
tk::dnn::Activation a22(&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Conv2d c23(&net, 1024, 3, 3, 1, 1, 1, 1, c23_bin, true);
tk::dnn::Activation a23(&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Conv2d c24(&net, 1024, 3, 3, 1, 1, 1, 1, c24_bin, true);
tk::dnn::Activation a24(&net, tk::dnn::ACTIVATION_LEAKY);
tkDNN::Layer *m25_layers[1] = { &a16 };
tkDNN::Route m25(&net, m25_layers, 1);
tkDNN::Conv2d c26(&net, 64, 1, 1, 1, 1, 0, 0, c26_bin, true);
tkDNN::Activation a26(&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Reorg r27(&net, 2);
tk::dnn::Layer *m25_layers[1] = { &a16 };
tk::dnn::Route m25(&net, m25_layers, 1);
tk::dnn::Conv2d c26(&net, 64, 1, 1, 1, 1, 0, 0, c26_bin, true);
tk::dnn::Activation a26(&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Reorg r27(&net, 2);
tkDNN::Layer *m28_layers[2] = { &r27, &a24 };
tkDNN::Route m28(&net, m28_layers, 2);
tk::dnn::Layer *m28_layers[2] = { &r27, &a24 };
tk::dnn::Route m28(&net, m28_layers, 2);
tkDNN::Conv2d c29(&net, 1024, 3, 3, 1, 1, 1, 1, c29_bin, true);
tkDNN::Activation a29(&net, tkDNN::ACTIVATION_LEAKY);
tkDNN::Conv2d c30(&net, 125, 1, 1, 1, 1, 0, 0, c30_bin, false);
tkDNN::Region g31(&net, 20, 4, 5);
tk::dnn::Conv2d c29(&net, 1024, 3, 3, 1, 1, 1, 1, c29_bin, true);
tk::dnn::Activation a29(&net, tk::dnn::ACTIVATION_LEAKY);
tk::dnn::Conv2d c30(&net, 125, 1, 1, 1, 1, 0, 0, c30_bin, false);
tk::dnn::Region g31(&net, 20, 4, 5);
tkDNN::RegionInterpret rI(dim, g31.output_dim, 20, 4, 5, 0.6f, g31_bin);
tk::dnn::RegionInterpret rI(dim, g31.output_dim, 20, 4, 5, 0.6f, g31_bin);
// Load input
dnnType *data;
@@ -110,11 +110,11 @@ int main() {
net.print();
//convert network to tensorRT
tkDNN::NetworkRT netRT(&net, "yolo_voc.rt");
tk::dnn::NetworkRT netRT(&net, "yolo_voc.rt");
dnnType *out_data, *out_data2; // cudnn output, tensorRT output
tkDNN::dataDim_t dim1 = dim; //input dim
tk::dnn::dataDim_t dim1 = dim; //input dim
printCenteredTitle(" CUDNN inference ", '=', 30); {
dim1.print();
TIMER_START
@@ -123,7 +123,7 @@ int main() {
dim1.print();
}
tkDNN::dataDim_t dim2 = dim;
tk::dnn::dataDim_t dim2 = dim;
printCenteredTitle(" TENSORRT inference ", '=', 30); {
dim2.print();
TIMER_START