Add getMemoryUsage function, detection update moved in abstract lass, splitted execution time in pre-inf-post, other minors.

Signed-off-by: Micaela Verucchi <micaelaverucchi@gmail.com>

src/CenternetDetection.cpp

@@ -264,40 +264,14 @@ void CenternetDetection::preprocess(cv::Mat &frame)
 #endif
 }
 
-void CenternetDetection::update(cv::Mat &frame) 
+void CenternetDetection::postprocess()
 {
-    originalSize = frame.size();
-    if(!frame.data) {
-        std::cout<<"CENTERNET: NO IMAGE DATA\n";
-        return;
-    }  
-    TIMER_START
-   
-    preprocess(frame);
-
-    printCenteredTitle(" TENSORRT inference ", '=', 30); {
-        dim2.print();
-        TIMER_START
-        netRT->infer(dim2, input_d);
-        TIMER_STOP
-        dim2.print();
-    }
-
     dnnType *rt_out[4];
     rt_out[0] = (dnnType *)netRT->buffersRT[1];
     rt_out[1] = (dnnType *)netRT->buffersRT[2];
     rt_out[2] = (dnnType *)netRT->buffersRT[3]; 
     rt_out[3] = (dnnType *)netRT->buffersRT[4]; 
 
-    postprocess(rt_out, 4);
-
-    // std::cout<<"TOTAL: \n";
-    TIMER_STOP
-    stats.push_back(t_ns);
-}
-
-void CenternetDetection::postprocess(dnnType **rt_out, const int n_out)
-{
     // auto start_t = std::chrono::steady_clock::now();
     // auto step_t = std::chrono::steady_clock::now();
     // auto end_t = std::chrono::steady_clock::now();

src/MobilenetDetection.cpp

@@ -243,45 +243,15 @@ void MobilenetDetection::preprocess(cv::Mat &frame)
 #endif
 }
 
-void MobilenetDetection::update(cv::Mat &frame)
+void MobilenetDetection::postprocess()
 {
-    TIMER_START
-    if(!frame.data) {
-        std::cout<<"MOBILENET: NO IMAGE DATA\n";
-        return;
-    } 
-    originalSize = frame.size();
-
-    //preprocess
-    preprocess(frame);
-
-    //do inference
-    tk::dnn::dataDim_t dim = tk::dnn::dataDim_t(1, 3, imageSize, imageSize, 1);;
-    printCenteredTitle(" TENSORRT inference ", '=', 30);
-    {
-        dim.print();
-        TIMER_START
-        netRT->infer(dim, input_d);
-        TIMER_STOP
-        dim.print();
-    }
-
     //get confidences and locations_h
     dnnType *rt_out[2];
     rt_out[0] = (dnnType *)netRT->buffersRT[3];
     rt_out[1] = (dnnType *)netRT->buffersRT[4];
 
     detected.clear();
-    //postprocess
-    postprocess(rt_out, 2);
 
-    TIMER_STOP
-    stats.push_back(t_ns);
-}
-
-
-void MobilenetDetection::postprocess(dnnType **rt_out, const int n_out)
-{
     checkCuda(cudaMemcpy(confidences_h, rt_out[0], nPriors * classes * sizeof(float), cudaMemcpyDeviceToHost));
     checkCuda(cudaMemcpy(locations_h, rt_out[1], N_COORDS * nPriors * sizeof(float), cudaMemcpyDeviceToHost));
     convert_locatios_to_boxes_and_center();

src/Yolo3Detection.cpp

@@ -84,52 +84,20 @@ void Yolo3Detection::preprocess(cv::Mat &frame)
 #endif
 }
 
-void Yolo3Detection::update(cv::Mat &frame)
+void Yolo3Detection::postprocess()
 {
-    TIMER_START
-    if(!frame.data) {
-        std::cout<<"YOLO: NO IMAGE DATA\n";
-        return;
-    }   
-
-    originalSize = frame.size();
-    preprocess(frame);
-
-    //do inference
-    tk::dnn::dataDim_t dim = netRT->input_dim;
-
-    // printDeviceVector(netRT->input_dim.tot()*sizeof(dnnType),input_d);
-    
-    
-    printCenteredTitle(" TENSORRT inference ", '=', 30); 
-    {
-        dim.print();
-        TIMER_START
-        netRT->infer(dim, input_d);
-        TIMER_STOP
-        dim.print();
-    }
-
     //get yolo outputs
     dnnType *rt_out[netRT->pluginFactory->n_yolos]; 
     for(int i=0; i<netRT->pluginFactory->n_yolos; i++) {
         rt_out[i] = (dnnType*)netRT->buffersRT[i+1];
     }
 
-    postprocess(rt_out, netRT->pluginFactory->n_yolos);
-        
-    TIMER_STOP
-    stats.push_back(t_ns);
-}
-
-void Yolo3Detection::postprocess(dnnType **rt_out, const int n_out)
-{
     float x_ratio =  float(originalSize.width) / float(netRT->input_dim.w);
     float y_ratio =  float(originalSize.height) / float(netRT->input_dim.h);
 
     // compute dets
     nDets = 0;
-    for(int i=0; i<n_out; i++) {
+    for(int i=0; i<netRT->pluginFactory->n_yolos; i++) {
         yolo[i]->dstData = rt_out[i];
         yolo[i]->computeDetections(dets, nDets, netRT->input_dim.w, netRT->input_dim.h, confThreshold);
     }

src/utils.cpp

@@ -168,3 +168,36 @@ void matrixMulAdd(  cublasHandle_t handle, dnnType* srcData, dnnType* dstData,
     checkERROR( cublasSaxpy(handle, dim, &alpha, srcData, 1, dstData, 1));
 
 }
+
+
+void getMemUsage(double& vm_usage_kb, double& resident_set_kb)
+{
+   using std::ios_base;
+   using std::ifstream;
+   using std::string;
+
+   vm_usage_kb     = 0.0;
+   resident_set_kb = 0.0;
+
+   ifstream stat_stream("/proc/self/stat",ios_base::in);
+
+   //all the stats
+   string pid, comm, state, ppid, pgrp, session, tty_nr;
+   string tpgid, flags, minflt, cminflt, majflt, cmajflt;
+   string utime, stime, cutime, cstime, priority, nice;
+   string O, itrealvalue, starttime;
+
+   unsigned long vsize;
+   long rss;
+
+   stat_stream >> pid >> comm >> state >> ppid >> pgrp >> session >> tty_nr
+               >> tpgid >> flags >> minflt >> cminflt >> majflt >> cmajflt
+               >> utime >> stime >> cutime >> cstime >> priority >> nice
+               >> O >> itrealvalue >> starttime >> vsize >> rss;
+
+   stat_stream.close();
+
+   long page_size_kb = sysconf(_SC_PAGE_SIZE) / 1024; // in case x86-64 is configured to use 2MB pages
+   vm_usage_kb     = vsize / 1024.0;  
+   resident_set_kb = rss * page_size_kb;
+}