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

Merge branch 'tree' into ipmslam

Browse Source
This commit is contained in:
Francesco Gatti
2020-05-03 15:58:43 +02:00
parent 98537624cf 85bbbf42a1
commit 7f0d9a930d
24 changed files with 4531 additions and 68 deletions
Download Patch File Download Diff File Expand all files Collapse all files
include/tkDNN/ImuOdom.h
+14 -3
View File
@@ -35,7 +35,8 @@ class ImuOdom {
// output eigen CPU
Eigen::MatrixXf deltaP, deltaQ;
Eigen::MatrixXd odomPOS, odomROT;
Eigen::MatrixXd odomPOS, odomEULER;
Eigen::Matrix3d odomROT;
Eigen::Isometry3f tf = Eigen::Isometry3f::Identity();
ImuOdom() {}
@@ -109,7 +110,7 @@ class ImuOdom {
odomPOS = Eigen::MatrixXd::Zero(3, 1);
odomROT = Eigen::MatrixXd::Identity(3, 3);
odomEULER = Eigen::MatrixXd::Zero(3, 1);
return true;
}
@@ -136,8 +137,18 @@ class ImuOdom {
q.x() = deltaQ(1);
q.y() = deltaQ(2);
q.z() = deltaQ(3);
odomPOS = odomPOS + odomROT*deltaP.cast<double>();
odomPOS = odomPOS + deltaP.cast<double>();
odomROT = odomROT * q.normalized().toRotationMatrix();
// compute euler
auto newEULER = odomROT.eulerAngles(0, 1, 2);
for(int i=0; i<3; i++) {
while( fabs(newEULER(i) - odomEULER(i)) > M_PI_2 ) {
newEULER(i) += newEULER(i) - odomEULER(i) > 0 ? -M_PI : +M_PI;
//std::cout<<newEULER(i)<<" "<<odomEULER(i)<<"\n";
}
}
odomEULER = newEULER;
// compose tf
tf.matrix().block(0, 0, 3, 3) = odomROT.cast<float>();
include/tkDNN/Layer.h
+8 -2
View File
@@ -18,6 +18,7 @@ enum layerType_t {
LAYER_ACTIVATION,
LAYER_ACTIVATION_CRELU,
LAYER_ACTIVATION_LEAKY,
LAYER_ACTIVATION_MISH,
LAYER_FLATTEN,
LAYER_RESHAPE,
LAYER_MULADD,
@@ -66,6 +67,7 @@ public:
case LAYER_ACTIVATION: return "Activation";
case LAYER_ACTIVATION_CRELU: return "ActivationCReLU";
case LAYER_ACTIVATION_LEAKY: return "ActivationLeaky";
case LAYER_ACTIVATION_MISH: return "ActivationMish";
case LAYER_FLATTEN: return "Flatten";
case LAYER_RESHAPE: return "Reshape";
case LAYER_MULADD: return "MulAdd";
@@ -168,7 +170,8 @@ public:
*/
typedef enum {
ACTIVATION_ELU = 100,
ACTIVATION_LEAKY = 101
ACTIVATION_LEAKY = 101,
ACTIVATION_MISH = 102
} tkdnnActivationMode_t;
/**
@@ -187,6 +190,8 @@ public:
return LAYER_ACTIVATION_CRELU;
else if (act_mode == ACTIVATION_LEAKY)
return LAYER_ACTIVATION_LEAKY;
else if (act_mode == ACTIVATION_MISH)
return LAYER_ACTIVATION_MISH;
else
return LAYER_ACTIVATION;
};
@@ -561,13 +566,14 @@ public:
int sort_class;
};
Yolo(Network *net, int classes, int num, std::string fname_weights, int n_masks=3);
Yolo(Network *net, int classes, int num, std::string fname_weights,int n_masks=3, float scale_xy=1);
virtual ~Yolo();
virtual layerType_t getLayerType() { return LAYER_YOLO; };
int classes, num, n_masks;
dnnType *mask_h, *mask_d; //anchors
dnnType *bias_h, *bias_d; //anchors
float scaleXY;
std::vector<std::string> classesNames;
virtual dnnType* infer(dataDim_t &dim, dnnType* srcData);
include/tkDNN/NetworkRT.h
+1
View File
@@ -25,6 +25,7 @@ template<typename T> T readBUF(const char*& buffer)
using namespace nvinfer1;
#include "pluginsRT/ActivationLeakyRT.h"
#include "pluginsRT/ActivationReLUCeilingRT.h"
#include "pluginsRT/ActivationMishRT.h"
#include "pluginsRT/ReorgRT.h"
#include "pluginsRT/RegionRT.h"
//#include "pluginsRT/RouteRT.h"
include/tkDNN/kernels.h
+3
View File
@@ -8,6 +8,7 @@ void activationLEAKYForward(dnnType *srcData, dnnType *dstData, int size, cudaSt
void activationReLUCeilingForward(dnnType *srcData, dnnType *dstData, int size, const float ceiling, cudaStream_t stream = cudaStream_t(0));
void activationLOGISTICForward(dnnType *srcData, dnnType *dstData, int size, cudaStream_t stream = cudaStream_t(0));
void activationSIGMOIDForward(dnnType *srcData, dnnType *dstData, int size, cudaStream_t stream = cudaStream_t(0));
void activationMishForward(dnnType* srcData, dnnType* dstData, int size, cudaStream_t stream= cudaStream_t(0));
void fill(dnnType *data, int size, dnnType val, cudaStream_t stream = cudaStream_t(0));
@@ -45,4 +46,6 @@ void dcnV2CudaForward(cublasStatus_t stat, cublasHandle_t handle,
const int in_n, const int in_c, const int in_h, const int in_w,
const int out_n, const int out_c, const int out_h, const int out_w,
const int dst_dim, cudaStream_t stream = cudaStream_t(0));
void scalAdd(dnnType* dstData, int size, float alpha, float beta, int inc, cudaStream_t stream = cudaStream_t(0));
#endif //KERNELS_H
include/tkDNN/pluginsRT/ActivationMishRT.h
+60
View File
@@ -0,0 +1,60 @@
#include<cassert>
#include "../kernels.h"
class ActivationMishRT : public IPlugin {
public:
ActivationMishRT() {
}
~ActivationMishRT(){
}
int getNbOutputs() const override {
return 1;
}
Dims getOutputDimensions(int index, const Dims* inputs, int nbInputDims) override {
return inputs[0];
}
void configure(const Dims* inputDims, int nbInputs, const Dims* outputDims, int nbOutputs, int maxBatchSize) override {
size = 1;
for(int i=0; i<outputDims[0].nbDims; i++)
size *= outputDims[0].d[i];
}
int initialize() override {
return 0;
}
virtual void terminate() override {
}
virtual size_t getWorkspaceSize(int maxBatchSize) const override {
return 0;
}
virtual int enqueue(int batchSize, const void*const * inputs, void** outputs, void* workspace, cudaStream_t stream) override {
activationMishForward((dnnType*)reinterpret_cast<const dnnType*>(inputs[0]),
reinterpret_cast<dnnType*>(outputs[0]), batchSize*size, stream);
return 0;
}
virtual size_t getSerializationSize() override {
return 1*sizeof(int);
}
virtual void serialize(void* buffer) override {
char *buf = reinterpret_cast<char*>(buffer);
tk::dnn::writeBUF(buf, size);
}
int size;
};
include/tkDNN/pluginsRT/YoloRT.h
+7 -2
View File
@@ -8,11 +8,12 @@ class YoloRT : public IPlugin {
public:
YoloRT(int classes, int num, tk::dnn::Yolo *yolo = nullptr, int n_masks=3) {
YoloRT(int classes, int num, tk::dnn::Yolo *yolo = nullptr, int n_masks=3, float scale_xy=1) {
this->classes = classes;
this->num = num;
this->n_masks = n_masks;
this->scaleXY = scale_xy;
mask = new dnnType[n_masks];
bias = new dnnType[num*n_masks*2];
@@ -64,6 +65,8 @@ public:
for(int n = 0; n < n_masks; ++n){
int index = entry_index(b, n*w*h, 0);
activationLOGISTICForward(srcData + index, dstData + index, 2*w*h, stream);
if (this->scaleXY != 1) scalAdd(dstData + index, 2 * w*h, this->scaleXY, -0.5*(this->scaleXY - 1), 1);
index = entry_index(b, n*w*h, 4);
activationLOGISTICForward(srcData + index, dstData + index, (1+classes)*w*h, stream);
@@ -76,7 +79,7 @@ public:
virtual size_t getSerializationSize() override {
return 6*sizeof(int) + n_masks*sizeof(dnnType) + num*n_masks*2*sizeof(dnnType) + YOLORT_CLASSNAME_W*classes*sizeof(char);
return 6*sizeof(int) + sizeof(float)+ n_masks*sizeof(dnnType) + num*n_masks*2*sizeof(dnnType) + YOLORT_CLASSNAME_W*classes*sizeof(char);
}
virtual void serialize(void* buffer) override {
@@ -87,6 +90,7 @@ public:
tk::dnn::writeBUF(buf, c);
tk::dnn::writeBUF(buf, h);
tk::dnn::writeBUF(buf, w);
tk::dnn::writeBUF(buf, scaleXY);
for(int i=0; i<n_masks; i++)
tk::dnn::writeBUF(buf, mask[i]);
for(int i=0; i<n_masks*2*num; i++)
@@ -104,6 +108,7 @@ public:
int c, h, w;
int classes, num, n_masks;
float scaleXY;
std::vector<std::string> classesNames;
dnnType *mask;