注:待笔者源码解读完毕,后续会为大家奉上详细的注释。
(1) darknet.c
此文件是darknet的main函数所在之处,各项命令在其源码都能找到出处。
#include <time.h>
#include <stdlib.h>
#include <stdio.h>
#if defined(_MSC_VER) && defined(_DEBUG)
#include <crtdbg.h>
#endif
#include "darknet.h"
#include "parser.h"
#include "utils.h"
#include "cuda.h"
#include "blas.h"
#include "connected_layer.h"
#ifdef OPENCV
#include <opencv2/highgui/highgui_c.h>
#endif
extern void predict_classifier(char *datacfg, char *cfgfile, char *weightfile, char *filename, int top);
extern void run_voxel(int argc, char **argv);
extern void run_yolo(int argc, char **argv);
extern void run_detector(int argc, char **argv);
extern void run_coco(int argc, char **argv);
extern void run_writing(int argc, char **argv);
extern void run_captcha(int argc, char **argv);
extern void run_nightmare(int argc, char **argv);
extern void run_dice(int argc, char **argv);
extern void run_compare(int argc, char **argv);
extern void run_classifier(int argc, char **argv);
extern void run_char_rnn(int argc, char **argv);
extern void run_vid_rnn(int argc, char **argv);
extern void run_tag(int argc, char **argv);
extern void run_cifar(int argc, char **argv);
extern void run_go(int argc, char **argv);
extern void run_art(int argc, char **argv);
extern void run_super(int argc, char **argv);
void average(int argc, char *argv[])
{
char *cfgfile = argv[2];
char *outfile = argv[3];
gpu_index = -1;
network net = parse_network_cfg(cfgfile);
network sum = parse_network_cfg(cfgfile);
char *weightfile = argv[4];
load_weights(&sum, weightfile);
int i, j;
int n = argc - 5;
for(i = 0; i < n; ++i){
weightfile = argv[i+5];
load_weights(&net, weightfile);
for(j = 0; j < net.n; ++j){
layer l = net.layers[j];
layer out = sum.layers[j];
if(l.type == CONVOLUTIONAL){
int num = l.n*l.c*l.size*l.size;
axpy_cpu(l.n, 1, l.biases, 1, out.biases, 1);
axpy_cpu(num, 1, l.weights, 1, out.weights, 1);
if(l.batch_normalize){
axpy_cpu(l.n, 1, l.scales, 1, out.scales, 1);
axpy_cpu(l.n, 1, l.rolling_mean, 1, out.rolling_mean, 1);
axpy_cpu(l.n, 1, l.rolling_variance, 1, out.rolling_variance, 1);
}
}
if(l.type == CONNECTED){
axpy_cpu(l.outputs, 1, l.biases, 1, out.biases, 1);
axpy_cpu(l.outputs*l.inputs, 1, l.weights, 1, out.weights, 1);
}
}
}
n = n+1;
for(j = 0; j < net.n; ++j){
layer l = sum.layers[j];
if(l.type == CONVOLUTIONAL){
int num = l.n*l.c*l.size*l.size;
scal_cpu(l.n, 1./n, l.biases, 1);
scal_cpu(num, 1./n, l.weights, 1);
if(l.batch_normalize){
scal_cpu(l.n, 1./n, l.scales, 1);
scal_cpu(l.n, 1./n, l.rolling_mean, 1);
scal_cpu(l.n, 1./n, l.rolling_variance, 1);
}
}
if(l.type == CONNECTED){
scal_cpu(l.outputs, 1./n, l.biases, 1);
scal_cpu(l.outputs*l.inputs, 1./n, l.weights, 1);
}
}
save_weights(sum, outfile);
}
void speed(char *cfgfile, int tics)
{
if (tics == 0) tics = 1000;
network net = parse_network_cfg(cfgfile);
set_batch_network(&net, 1);
int i;
time_t start = time(0);
image im = make_image(net.w, net.h, net.c);
for(i = 0; i < tics; ++i){
network_predict(net, im.data);
}
double t = difftime(time(0), start);
printf("\n%d evals, %f Seconds\n", tics, t);
printf("Speed: %f sec/eval\n", t/tics);
printf("Speed: %f Hz\n", tics/t);
}
void operations(char *cfgfile)
{
gpu_index = -1;
network net = parse_network_cfg(cfgfile);
int i;
long ops = 0;
for(i = 0; i < net.n; ++i){
layer l = net.layers[i];
if(l.type == CONVOLUTIONAL){
ops += 2l * l.n * l.size*l.size*l.c * l.out_h*l.out_w;
} else if(l.type == CONNECTED){
ops += 2l * l.inputs * l.outputs;
} else if (l.type == RNN){
ops += 2l * l.input_layer->inputs * l.input_layer->outputs;
ops += 2l * l.self_layer->inputs * l.self_layer->outputs;
ops += 2l * l.output_layer->inputs * l.output_layer->outputs;
} else if (l.type == GRU){
ops += 2l * l.uz->inputs * l.uz->outputs;
ops += 2l * l.uh->inputs * l.uh->outputs;
ops += 2l * l.ur->inputs * l.ur->outputs;
ops += 2l * l.wz->inputs * l.wz->outputs;
ops += 2l * l.wh->inputs * l.wh->outputs;
ops += 2l * l.wr->inputs * l.wr->outputs;
} else if (l.type == LSTM){
ops += 2l * l.uf->inputs * l.uf->outputs;
ops += 2l * l.ui->inputs * l.ui->outputs;
ops += 2l * l.ug->inputs * l.ug->outputs;
ops += 2l * l.uo->inputs * l.uo->outputs;
ops += 2l * l.wf->inputs * l.wf->outputs;
ops += 2l * l.wi->inputs * l.wi->outputs;
ops += 2l * l.wg->inputs * l.wg->outputs;
ops += 2l * l.wo->inputs * l.wo->outputs;
}
}
printf("Floating Point Operations: %ld\n", ops);
printf("Floating Point Operations: %.2f Bn\n", (float)ops/1000000000.);
}
void oneoff(char *cfgfile, char *weightfile, char *outfile)
{
gpu_index = -1;
network net = parse_network_cfg(cfgfile);
int oldn = net.layers[net.n - 2].n;
int c = net.layers[net.n - 2].c;
net.layers[net.n - 2].n = 9372;
net.layers[net.n - 2].biases += 5;
net.layers[net.n - 2].weights += 5*c;
if(weightfile){
load_weights(&net, weightfile);
}
net.layers[net.n - 2].biases -= 5;
net.layers[net.n - 2].weights -= 5*c;
net.layers[net.n - 2].n = oldn;
printf("%d\n", oldn);
layer l = net.layers[net.n - 2];
copy_cpu(l.n/3, l.biases, 1, l.biases + l.n/3, 1);
copy_cpu(l.n/3, l.biases, 1, l.biases + 2*l.n/3, 1);
copy_cpu(l.n/3*l.c, l.weights, 1, l.weights + l.n/3*l.c, 1);
copy_cpu(l.n/3*l.c, l.weights, 1, l.weights + 2*l.n/3*l.c, 1);
*net.seen = 0;
save_weights(net, outfile);
}
void partial(char *cfgfile, char *weightfile, char *outfile, int max)
{
gpu_index = -1;
network net = parse_network_cfg(cfgfile);
if(weightfile){
load_weights_upto(&net, weightfile, max);
}
*net.seen = 0;
save_weights_upto(net, outfile, max);
}
#include "convolutional_layer.h"
void rescale_net(char *cfgfile, char *weightfile, char *outfile)
{
gpu_index = -1;
network net = parse_network_cfg(cfgfile);
if(weightfile){
load_weights(&net, weightfile);
}
int i;
for(i = 0; i < net.n; ++i){
layer l = net.layers[i];
if(l.type == CONVOLUTIONAL){
rescale_weights(l, 2, -.5);
break;
}
}
save_weights(net, outfile);
}
void rgbgr_net(char *cfgfile, char *weightfile, char *outfile)
{
gpu_index = -1;
network net = parse_network_cfg(cfgfile);
if(weightfile){
load_weights(&net, weightfile);
}
int i;
for(i = 0; i < net.n; ++i){
layer l = net.layers[i];
if(l.type == CONVOLUTIONAL){
rgbgr_weights(l);
break;
}
}
save_weights(net, outfile);
}
void reset_normalize_net(char *cfgfile, char *weightfile, char *outfile)
{
gpu_index = -1;
network net = parse_network_cfg(cfgfile);
if (weightfile) {
load_weights(&net, weightfile);
}
int i;
for (i = 0; i < net.n; ++i) {
layer l = net.layers[i];
if (l.type == CONVOLUTIONAL && l.batch_normalize) {
denormalize_convolutional_layer(l);
}
if (l.type == CONNECTED && l.batch_normalize) {
denormalize_connected_layer(l);
}
if (l.type == GRU && l.batch_normalize) {
denormalize_connected_layer(*l.input_z_layer);
denormalize_connected_layer(*l.input_r_layer);
denormalize_connected_layer(*l.input_h_layer);
denormalize_connected_layer(*l.state_z_layer);
denormalize_connected_layer(*l.state_r_layer);
denormalize_connected_layer(*l.state_h_layer);
}
if (l.type == LSTM && l.batch_normalize) {
denormalize_connected_layer(*l.wf);
denormalize_connected_layer(*l.wi);
denormalize_connected_layer(*l.wg);
denormalize_connected_layer(*l.wo);
denormalize_connected_layer(*l.uf);
denormalize_connected_layer(*l.ui);
denormalize_connected_layer(*l.ug);
denormalize_connected_layer(*l.uo);
}
}
save_weights(net, outfile);
}
layer normalize_layer(layer l, int n)
{
int j;
l.batch_normalize=1;
l.scales = (float*)calloc(n, sizeof(float));
for(j = 0; j < n; ++j){
l.scales[j] = 1;
}
l.rolling_mean = (float*)calloc(n, sizeof(float));
l.rolling_variance = (float*)calloc(n, sizeof(float));
return l;
}
void normalize_net(char *cfgfile, char *weightfile, char *outfile)
{
gpu_index = -1;
network net = parse_network_cfg(cfgfile);
if(weightfile){
load_weights(&net, weightfile);
}
int i;
for(i = 0; i < net.n; ++i){
layer l = net.layers[i];
if(l.type == CONVOLUTIONAL && !l.batch_normalize){
net.layers[i] = normalize_layer(l, l.n);
}
if (l.type == CONNECTED && !l.batch_normalize) {
net.layers[i] = normalize_layer(l, l.outputs);
}
if (l.type == GRU && l.batch_normalize) {
*l.input_z_layer = normalize_layer(*l.input_z_layer, l.input_z_layer->outputs);
*l.input_r_layer = normalize_layer(*l.input_r_layer, l.input_r_layer->outputs);
*l.input_h_layer = normalize_layer(*l.input_h_layer, l.input_h_layer->outputs);
*l.state_z_layer = normalize_layer(*l.state_z_layer, l.state_z_layer->outputs);
*l.state_r_layer = normalize_layer(*l.state_r_layer, l.state_r_layer->outputs);
*l.state_h_layer = normalize_layer(*l.state_h_layer, l.state_h_layer->outputs);
net.layers[i].batch_normalize=1;
}
if (l.type == LSTM && l.batch_normalize) {
*l.wf = normalize_layer(*l.wf, l.wf->outputs);
*l.wi = normalize_layer(*l.wi, l.wi->outputs);
*l.wg = normalize_layer(*l.wg, l.wg->outputs);
*l.wo = normalize_layer(*l.wo, l.wo->outputs);
*l.uf = normalize_layer(*l.uf, l.uf->outputs);
*l.ui = normalize_layer(*l.ui, l.ui->outputs);
*l.ug = normalize_layer(*l.ug, l.ug->outputs);
*l.uo = normalize_layer(*l.uo, l.uo->outputs);
net.layers[i].batch_normalize=1;
}
}
save_weights(net, outfile);
}
void statistics_net(char *cfgfile, char *weightfile)
{
gpu_index = -1;
network net = parse_network_cfg(cfgfile);
if (weightfile) {
load_weights(&net, weightfile);
}
int i;
for (i = 0; i < net.n; ++i) {
layer l = net.layers[i];
if (l.type == CONNECTED && l.batch_normalize) {
printf("Connected Layer %d\n", i);
statistics_connected_layer(l);
}
if (l.type == GRU && l.batch_normalize) {
printf("GRU Layer %d\n", i);
printf("Input Z\n");
statistics_connected_layer(*l.input_z_layer);
printf("Input R\n");
statistics_connected_layer(*l.input_r_layer);
printf("Input H\n");
statistics_connected_layer(*l.input_h_layer);
printf("State Z\n");
statistics_connected_layer(*l.state_z_layer);
printf("State R\n");
statistics_connected_layer(*l.state_r_layer);
printf("State H\n");
statistics_connected_layer(*l.state_h_layer);
}
if (l.type == LSTM && l.batch_normalize) {
printf("LSTM Layer %d\n", i);
printf("wf\n");
statistics_connected_layer(*l.wf);
printf("wi\n");
statistics_connected_layer(*l.wi);
printf("wg\n");
statistics_connected_layer(*l.wg);
printf("wo\n");
statistics_connected_layer(*l.wo);
printf("uf\n");
statistics_connected_layer(*l.uf);
printf("ui\n");
statistics_connected_layer(*l.ui);
printf("ug\n");
statistics_connected_layer(*l.ug);
printf("uo\n");
statistics_connected_layer(*l.uo);
}
printf("\n");
}
}
void denormalize_net(char *cfgfile, char *weightfile, char *outfile)
{
gpu_index = -1;
network net = parse_network_cfg(cfgfile);
if (weightfile) {
load_weights(&net, weightfile);
}
int i;
for (i = 0; i < net.n; ++i) {
layer l = net.layers[i];
if (l.type == CONVOLUTIONAL && l.batch_normalize) {
denormalize_convolutional_layer(l);
net.layers[i].batch_normalize=0;
}
if (l.type == CONNECTED && l.batch_normalize) {
denormalize_connected_layer(l);
net.layers[i].batch_normalize=0;
}
if (l.type == GRU && l.batch_normalize) {
denormalize_connected_layer(*l.input_z_layer);
denormalize_connected_layer(*l.input_r_layer);
denormalize_connected_layer(*l.input_h_layer);
denormalize_connected_layer(*l.state_z_layer);
denormalize_connected_layer(*l.state_r_layer);
denormalize_connected_layer(*l.state_h_layer);
l.input_z_layer->batch_normalize = 0;
l.input_r_layer->batch_normalize = 0;
l.input_h_layer->batch_normalize = 0;
l.state_z_layer->batch_normalize = 0;
l.state_r_layer->batch_normalize = 0;
l.state_h_layer->batch_normalize = 0;
net.layers[i].batch_normalize=0;
}
if (l.type == GRU && l.batch_normalize) {
denormalize_connected_layer(*l.wf);
denormalize_connected_layer(*l.wi);
denormalize_connected_layer(*l.wg);
denormalize_connected_layer(*l.wo);
denormalize_connected_layer(*l.uf);
denormalize_connected_layer(*l.ui);
denormalize_connected_layer(*l.ug);
denormalize_connected_layer(*l.uo);
l.wf->batch_normalize = 0;
l.wi->batch_normalize = 0;
l.wg->batch_normalize = 0;
l.wo->batch_normalize = 0;
l.uf->batch_normalize = 0;
l.ui->batch_normalize = 0;
l.ug->batch_normalize = 0;
l.uo->batch_normalize = 0;
net.layers[i].batch_normalize=0;
}
}
save_weights(net, outfile);
}
void visualize(char *cfgfile, char *weightfile)
{
network net = parse_network_cfg(cfgfile);
if(weightfile){
load_weights(&net, weightfile);
}
visualize_network(net);
#ifdef OPENCV
cvWaitKey(0);
#endif
}
int main(int argc, char **argv)
{
#ifdef _DEBUG
_CrtSetDbgFlag(_CRTDBG_ALLOC_MEM_DF | _CRTDBG_LEAK_CHECK_DF);
#endif
int i;
for (i = 0; i < argc; ++i) {
if (!argv[i]) continue;
strip_args(argv[i]);
}
//test_resize("data/bad.jpg");
//test_box();
//test_convolutional_layer();
if(argc < 2){
fprintf(stderr, "usage: %s <function>\n", argv[0]);
return 0;
}
gpu_index = find_int_arg(argc, argv, "-i", 0);
if(find_arg(argc, argv, "-nogpu")) {
gpu_index = -1;
printf("\n Currently Darknet doesn't support -nogpu flag. If you want to use CPU - please compile Darknet with GPU=0 in the Makefile, or compile darknet_no_gpu.sln on Windows.\n");
exit(-1);
}
#ifndef GPU
gpu_index = -1;
#else
if(gpu_index >= 0){
cuda_set_device(gpu_index);
CHECK_CUDA(cudaSetDeviceFlags(cudaDeviceScheduleBlockingSync));
}
#endif
if (0 == strcmp(argv[1], "average")){
average(argc, argv);
} else if (0 == strcmp(argv[1], "yolo")){
run_yolo(argc, argv);
} else if (0 == strcmp(argv[1], "voxel")){
run_voxel(argc, argv);
} else if (0 == strcmp(argv[1], "super")){
run_super(argc, argv);
} else if (0 == strcmp(argv[1], "detector")){
run_detector(argc, argv);
} else if (0 == strcmp(argv[1], "detect")){
float thresh = find_float_arg(argc, argv, "-thresh", .24);
int ext_output = find_arg(argc, argv, "-ext_output");
char *filename = (argc > 4) ? argv[4]: 0;
test_detector("cfg/coco.data", argv[2], argv[3], filename, thresh, 0.5, 0, 1, 0, NULL);
} else if (0 == strcmp(argv[1], "cifar")){
run_cifar(argc, argv);
} else if (0 == strcmp(argv[1], "go")){
run_go(argc, argv);
} else if (0 == strcmp(argv[1], "rnn")){
run_char_rnn(argc, argv);
} else if (0 == strcmp(argv[1], "vid")){
run_vid_rnn(argc, argv);
} else if (0 == strcmp(argv[1], "coco")){
run_coco(argc, argv);
} else if (0 == strcmp(argv[1], "classify")){
predict_classifier("cfg/imagenet1k.data", argv[2], argv[3], argv[4], 5);
} else if (0 == strcmp(argv[1], "classifier")){
run_classifier(argc, argv);
} else if (0 == strcmp(argv[1], "art")){
run_art(argc, argv);
} else if (0 == strcmp(argv[1], "tag")){
run_tag(argc, argv);
} else if (0 == strcmp(argv[1], "compare")){
run_compare(argc, argv);
} else if (0 == strcmp(argv[1], "dice")){
run_dice(argc, argv);
} else if (0 == strcmp(argv[1], "writing")){
run_writing(argc, argv);
} else if (0 == strcmp(argv[1], "3d")){
composite_3d(argv[2], argv[3], argv[4], (argc > 5) ? atof(argv[5]) : 0);
} else if (0 == strcmp(argv[1], "test")){
test_resize(argv[2]);
} else if (0 == strcmp(argv[1], "captcha")){
run_captcha(argc, argv);
} else if (0 == strcmp(argv[1], "nightmare")){
run_nightmare(argc, argv);
} else if (0 == strcmp(argv[1], "rgbgr")){
rgbgr_net(argv[2], argv[3], argv[4]);
} else if (0 == strcmp(argv[1], "reset")){
reset_normalize_net(argv[2], argv[3], argv[4]);
} else if (0 == strcmp(argv[1], "denormalize")){
denormalize_net(argv[2], argv[3], argv[4]);
} else if (0 == strcmp(argv[1], "statistics")){
statistics_net(argv[2], argv[3]);
} else if (0 == strcmp(argv[1], "normalize")){
normalize_net(argv[2], argv[3], argv[4]);
} else if (0 == strcmp(argv[1], "rescale")){
rescale_net(argv[2], argv[3], argv[4]);
} else if (0 == strcmp(argv[1], "ops")){
operations(argv[2]);
} else if (0 == strcmp(argv[1], "speed")){
speed(argv[2], (argc > 3 && argv[3]) ? atoi(argv[3]) : 0);
} else if (0 == strcmp(argv[1], "oneoff")){
oneoff(argv[2], argv[3], argv[4]);
} else if (0 == strcmp(argv[1], "partial")){
partial(argv[2], argv[3], argv[4], atoi(argv[5]));
} else if (0 == strcmp(argv[1], "average")){
average(argc, argv);
} else if (0 == strcmp(argv[1], "visualize")){
visualize(argv[2], (argc > 3) ? argv[3] : 0);
} else if (0 == strcmp(argv[1], "imtest")){
test_resize(argv[2]);
} else {
fprintf(stderr, "Not an option: %s\n", argv[1]);
}
return 0;
}
(2) detector.c
此文件是darknet训练测试各功能函数的所在地。
#include "darknet.h"
#include "network.h"
#include "region_layer.h"
#include "cost_layer.h"
#include "utils.h"
#include "parser.h"
#include "box.h"
#include "demo.h"
#include "option_list.h"
#ifndef __COMPAR_FN_T
#define __COMPAR_FN_T
typedef int (*__compar_fn_t)(const void*, const void*);
#ifdef __USE_GNU
typedef __compar_fn_t comparison_fn_t;
#endif
#endif
#ifdef OPENCV
#include <opencv2/highgui/highgui_c.h>
#include <opencv2/core/core_c.h>
//#include "opencv2/core/core.hpp"
#include <opencv2/core/version.hpp>
#include <opencv2/imgproc/imgproc_c.h>
#ifndef CV_VERSION_EPOCH
#include <opencv2/videoio/videoio_c.h>
#define OPENCV_VERSION CVAUX_STR(CV_VERSION_MAJOR)"" CVAUX_STR(CV_VERSION_MINOR)"" CVAUX_STR(CV_VERSION_REVISION)
#ifndef USE_CMAKE_LIBS
#pragma comment(lib, "opencv_world" OPENCV_VERSION ".lib")
#endif // USE_CMAKE_LIBS
#else
#define OPENCV_VERSION CVAUX_STR(CV_VERSION_EPOCH)"" CVAUX_STR(CV_VERSION_MAJOR)"" CVAUX_STR(CV_VERSION_MINOR)
#ifndef USE_CMAKE_LIBS
#pragma comment(lib, "opencv_core" OPENCV_VERSION ".lib")
#pragma comment(lib, "opencv_imgproc" OPENCV_VERSION ".lib")
#pragma comment(lib, "opencv_highgui" OPENCV_VERSION ".lib")
#endif // USE_CMAKE_LIBS
#endif
IplImage* draw_train_chart(float max_img_loss, int max_batches, int number_of_lines, int img_size, int dont_show);
void draw_train_loss(IplImage* img, int img_size, float avg_loss, float max_img_loss, int current_batch, int max_batches,
float precision, int draw_precision, char *accuracy_name, int dont_show, int mjpeg_port);
#endif // OPENCV
#ifndef CV_RGB
#define CV_RGB(r, g, b) cvScalar( (b), (g), (r), 0 )
#endif // OPENCV
#include "http_stream.h"
int check_mistakes;
static int coco_ids[] = { 1,2,3,4,5,6,7,8,9,10,11,13,14,15,16,17,18,19,20,21,22,23,24,25,27,28,31,32,33,34,35,36,37,38,39,40,41,42,43,44,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,67,70,72,73,74,75,76,77,78,79,80,81,82,84,85,86,87,88,89,90 };
void train_detector(char *datacfg, char *cfgfile, char *weightfile, int *gpus, int ngpus, int clear, int dont_show, int calc_map, int mjpeg_port)
{
list *options = read_data_cfg(datacfg);
char *train_images = option_find_str(options, "train", "data/train.txt");
char *valid_images = option_find_str(options, "valid", train_images);
char *backup_directory = option_find_str(options, "backup", "/backup/");
int train_images_num = 0;
network net_map;
if (calc_map) {
FILE* valid_file = fopen(valid_images, "r");
if (!valid_file) {
printf("\n Error: There is no %s file for mAP calculation!\n Don't use -map flag.\n Or set valid=%s in your %s file. \n", valid_images, train_images, datacfg);
getchar();
exit(-1);
}
else fclose(valid_file);
list *plist = get_paths(train_images);
train_images_num = plist->size;
free_list(plist);
cuda_set_device(gpus[0]);
printf(" Prepare additional network for mAP calculation...\n");
net_map = parse_network_cfg_custom(cfgfile, 1, 1);
int k; // free memory unnecessary arrays
for (k = 0; k < net_map.n; ++k) {
free_layer(net_map.layers[k]);
}
/*
#ifdef GPU
cuda_free(net_map.workspace);
cuda_free(net_map.input_state_gpu);
if (*net_map.input16_gpu) cuda_free(*net_map.input16_gpu);
if (*net_map.output16_gpu) cuda_free(*net_map.output16_gpu);
#else
free(net_map.workspace);
#endif
*/
}
srand(time(0));
char *base = basecfg(cfgfile);
printf("%s\n", base);
float avg_loss = -1;
network* nets = (network*)calloc(ngpus, sizeof(network));
srand(time(0));
int seed = rand();
int i;
for (i = 0; i < ngpus; ++i) {
srand(seed);
#ifdef GPU
cuda_set_device(gpus[i]);
#endif
nets[i] = parse_network_cfg(cfgfile);
if (weightfile) {
load_weights(&nets[i], weightfile);
}
if (clear) *nets[i].seen = 0;
nets[i].learning_rate *= ngpus;
}
srand(time(0));
network net = nets[0];
const int actual_batch_size = net.batch * net.subdivisions;
if (actual_batch_size == 1) {
printf("\n Error: You set incorrect value batch=1 for Training! You should set batch=64 subdivision=64 \n");
getchar();
}
else if (actual_batch_size < 64) {
printf("\n Warning: You set batch=%d lower than 64! It is recommended to set batch=64 subdivision=64 \n", actual_batch_size);
}
int imgs = net.batch * net.subdivisions * ngpus;
printf("Learning Rate: %g, Momentum: %g, Decay: %g\n", net.learning_rate, net.momentum, net.decay);
data train, buffer;
layer l = net.layers[net.n - 1];
int classes = l.classes;
float jitter = l.jitter;
list *plist = get_paths(train_images);
//int N = plist->size;
char **paths = (char **)list_to_array(plist);
int init_w = net.w;
int init_h = net.h;
int iter_save, iter_save_last, iter_map;
iter_save = get_current_batch(net);
iter_save_last = get_current_batch(net);
iter_map = get_current_batch(net);
float mean_average_precision = -1;
load_args args = { 0 };
args.w = net.w;
args.h = net.h;
args.c = net.c;
args.paths = paths;
args.n = imgs;
args.m = plist->size;
args.classes = classes;
args.flip = net.flip;
args.jitter = jitter;
args.num_boxes = l.max_boxes;
args.d = &buffer;
args.type = DETECTION_DATA;
args.threads = 64; // 16 or 64
args.angle = net.angle;
args.exposure = net.exposure;
args.saturation = net.saturation;
args.hue = net.hue;
#ifdef OPENCV
args.threads = 3 * ngpus; // Amazon EC2 Tesla V100: p3.2xlarge (8 logical cores) - p3.16xlarge
//args.threads = 12 * ngpus; // Ryzen 7 2700X (16 logical cores)
IplImage* img = NULL;
float max_img_loss = 5;
int number_of_lines = 100;
int img_size = 1000;
img = draw_train_chart(max_img_loss, net.max_batches, number_of_lines, img_size, dont_show);
#endif //OPENCV
if (net.track) {
args.track = net.track;
args.augment_speed = net.augment_speed;
args.threads = net.subdivisions * ngpus; // 2 * ngpus;
args.mini_batch = net.batch / net.time_steps;
printf("\n Tracking! batch = %d, subdiv = %d, time_steps = %d, mini_batch = %d \n", net.batch, net.subdivisions, net.time_steps, args.mini_batch);
}
//printf(" imgs = %d \n", imgs);
pthread_t load_thread = load_data(args);
double time;
int count = 0;
//while(i*imgs < N*120){
while (get_current_batch(net) < net.max_batches) {
if (l.random && count++ % 10 == 0) {
printf("Resizing\n");
float random_val = rand_scale(1.4); // *x or /x
int dim_w = roundl(random_val*init_w / 32 + 1) * 32;
int dim_h = roundl(random_val*init_h / 32 + 1) * 32;
// at the beginning
if (avg_loss < 0) {
dim_w = roundl(1.4*init_w / 32 + 1) * 32;
dim_h = roundl(1.4*init_h / 32 + 1) * 32;
}
if (dim_w < 32) dim_w = 32;
if (dim_h < 32) dim_h = 32;
printf("%d x %d \n", dim_w, dim_h);
args.w = dim_w;
args.h = dim_h;
pthread_join(load_thread, 0);
train = buffer;
free_data(train);
load_thread = load_data(args);
for (i = 0; i < ngpus; ++i) {
resize_network(nets + i, dim_w, dim_h);
}
net = nets[0];
}
time = what_time_is_it_now();
pthread_join(load_thread, 0);
train = buffer;
load_thread = load_data(args);
/*
int k;
for(k = 0; k < l.max_boxes; ++k){
box b = float_to_box(train.y.vals[10] + 1 + k*5);
if(!b.x) break;
printf("loaded: %f %f %f %f\n", b.x, b.y, b.w, b.h);
}
image im = float_to_image(448, 448, 3, train.X.vals[10]);
int k;
for(k = 0; k < l.max_boxes; ++k){
box b = float_to_box(train.y.vals[10] + 1 + k*5);
printf("%d %d %d %d\n", truth.x, truth.y, truth.w, truth.h);
draw_bbox(im, b, 8, 1,0,0);
}
save_image(im, "truth11");
*/
printf("Loaded: %lf seconds\n", (what_time_is_it_now() - time));
time = what_time_is_it_now();
float loss = 0;
#ifdef GPU
if (ngpus == 1) {
loss = train_network(net, train);
}
else {
loss = train_networks(nets, ngpus, train, 4);
}
#else
loss = train_network(net, train);
#endif
if (avg_loss < 0 || avg_loss != avg_loss) avg_loss = loss; // if(-inf or nan)
avg_loss = avg_loss*.9 + loss*.1;
i = get_current_batch(net);
int calc_map_for_each = 4 * train_images_num / (net.batch * net.subdivisions); // calculate mAP for each 4 Epochs
calc_map_for_each = fmax(calc_map_for_each, 100);
int next_map_calc = iter_map + calc_map_for_each;
next_map_calc = fmax(next_map_calc, net.burn_in);
next_map_calc = fmax(next_map_calc, 1000);
if (calc_map) {
printf("\n (next mAP calculation at %d iterations) ", next_map_calc);
if (mean_average_precision > 0) printf("\n Last accuracy mAP@0.5 = %2.2f %% ", mean_average_precision * 100);
}
if (net.cudnn_half) {
if (i < net.burn_in * 3) fprintf(stderr, "\n Tensor Cores are disabled until the first %d iterations are reached.", 3 * net.burn_in);
else fprintf(stderr, "\n Tensor Cores are used.");
}
printf("\n %d: %f, %f avg loss, %f rate, %lf seconds, %d images\n", get_current_batch(net), loss, avg_loss, get_current_rate(net), (what_time_is_it_now() - time), i*imgs);
int draw_precision = 0;
if (calc_map && (i >= next_map_calc || i == net.max_batches)) {
if (l.random) {
printf("Resizing to initial size: %d x %d \n", init_w, init_h);
args.w = init_w;
args.h = init_h;
pthread_join(load_thread, 0);
train = buffer;
load_thread = load_data(args);
int k;
for (k = 0; k < ngpus; ++k) {
resize_network(nets + k, init_w, init_h);
}
net = nets[0];
}
copy_weights_net(net, &net_map);
// combine Training and Validation networks
//network net_combined = combine_train_valid_networks(net, net_map);
iter_map = i;
mean_average_precision = validate_detector_map(datacfg, cfgfile, weightfile, 0.25, 0.5, &net_map);// &net_combined);
printf("\n mean_average_precision (mAP@0.5) = %f \n", mean_average_precision);
draw_precision = 1;
}
#ifdef OPENCV
draw_train_loss(img, img_size, avg_loss, max_img_loss, i, net.max_batches, mean_average_precision, draw_precision, "mAP%", dont_show, mjpeg_port);
#endif // OPENCV
//if (i % 1000 == 0 || (i < 1000 && i % 100 == 0)) {
//if (i % 100 == 0) {
if (i >= (iter_save + 1000) || i % 1000 == 0) {
iter_save = i;
#ifdef GPU
if (ngpus != 1) sync_nets(nets, ngpus, 0);
#endif
char buff[256];
sprintf(buff, "%s/%s_%d.weights", backup_directory, base, i);
save_weights(net, buff);
}
if (i >= (iter_save_last + 100) || i % 100 == 0) {
iter_save_last = i;
#ifdef GPU
if (ngpus != 1) sync_nets(nets, ngpus, 0);
#endif
char buff[256];
sprintf(buff, "%s/%s_last.weights", backup_directory, base, i);
save_weights(net, buff);
}
free_data(train);
}
#ifdef GPU
if (ngpus != 1) sync_nets(nets, ngpus, 0);
#endif
char buff[256];
sprintf(buff, "%s/%s_final.weights", backup_directory, base);
save_weights(net, buff);
#ifdef OPENCV
cvReleaseImage(&img);
cvDestroyAllWindows();
#endif
// free memory
pthread_join(load_thread, 0);
free_data(buffer);
free(base);
free(paths);
free_list_contents(plist);
free_list(plist);
free_list_contents_kvp(options);
free_list(options);
for (i = 0; i < ngpus; ++i) free_network(nets[i]);
free(nets);
//free_network(net);
if (calc_map) {
net_map.n = 0;
free_network(net_map);
}
}
static int get_coco_image_id(char *filename)
{
char *p = strrchr(filename, '/');
char *c = strrchr(filename, '_');
if (c) p = c;
return atoi(p + 1);
}
static void print_cocos(FILE *fp, char *image_path, detection *dets, int num_boxes, int classes, int w, int h)
{
int i, j;
int image_id = get_coco_image_id(image_path);
for (i = 0; i < num_boxes; ++i) {
float xmin = dets[i].bbox.x - dets[i].bbox.w / 2.;
float xmax = dets[i].bbox.x + dets[i].bbox.w / 2.;
float ymin = dets[i].bbox.y - dets[i].bbox.h / 2.;
float ymax = dets[i].bbox.y + dets[i].bbox.h / 2.;
if (xmin < 0) xmin = 0;
if (ymin < 0) ymin = 0;
if (xmax > w) xmax = w;
if (ymax > h) ymax = h;
float bx = xmin;
float by = ymin;
float bw = xmax - xmin;
float bh = ymax - ymin;
for (j = 0; j < classes; ++j) {
if (dets[i].prob[j]) fprintf(fp, "{\"image_id\":%d, \"category_id\":%d, \"bbox\":[%f, %f, %f, %f], \"score\":%f},\n", image_id, coco_ids[j], bx, by, bw, bh, dets[i].prob[j]);
}
}
}
void print_detector_detections(FILE **fps, char *id, detection *dets, int total, int classes, int w, int h)
{
int i, j;
for (i = 0; i < total; ++i) {
float xmin = dets[i].bbox.x - dets[i].bbox.w / 2. + 1;
float xmax = dets[i].bbox.x + dets[i].bbox.w / 2. + 1;
float ymin = dets[i].bbox.y - dets[i].bbox.h / 2. + 1;
float ymax = dets[i].bbox.y + dets[i].bbox.h / 2. + 1;
if (xmin < 1) xmin = 1;
if (ymin < 1) ymin = 1;
if (xmax > w) xmax = w;
if (ymax > h) ymax = h;
for (j = 0; j < classes; ++j) {
if (dets[i].prob[j]) fprintf(fps[j], "%s %f %f %f %f %f\n", id, dets[i].prob[j],
xmin, ymin, xmax, ymax);
}
}
}
void print_imagenet_detections(FILE *fp, int id, detection *dets, int total, int classes, int w, int h)
{
int i, j;
for (i = 0; i < total; ++i) {
float xmin = dets[i].bbox.x - dets[i].bbox.w / 2.;
float xmax = dets[i].bbox.x + dets[i].bbox.w / 2.;
float ymin = dets[i].bbox.y - dets[i].bbox.h / 2.;
float ymax = dets[i].bbox.y + dets[i].bbox.h / 2.;
if (xmin < 0) xmin = 0;
if (ymin < 0) ymin = 0;
if (xmax > w) xmax = w;
if (ymax > h) ymax = h;
for (j = 0; j < classes; ++j) {
int myclass = j;
if (dets[i].prob[myclass]) fprintf(fp, "%d %d %f %f %f %f %f\n", id, j + 1, dets[i].prob[myclass],
xmin, ymin, xmax, ymax);
}
}
}
void validate_detector(char *datacfg, char *cfgfile, char *weightfile, char *outfile)
{
int j;
list *options = read_data_cfg(datacfg);
char *valid_images = option_find_str(options, "valid", "data/train.list");
char *name_list = option_find_str(options, "names", "data/names.list");
char *prefix = option_find_str(options, "results", "results");
char **names = get_labels(name_list);
char *mapf = option_find_str(options, "map", 0);
int *map = 0;
if (mapf) map = read_map(mapf);
network net = parse_network_cfg_custom(cfgfile, 1, 1); // set batch=1
if (weightfile) {
load_weights(&net, weightfile);
}
//set_batch_network(&net, 1);
fprintf(stderr, "Learning Rate: %g, Momentum: %g, Decay: %g\n", net.learning_rate, net.momentum, net.decay);
srand(time(0));
list *plist = get_paths(valid_images);
char **paths = (char **)list_to_array(plist);
layer l = net.layers[net.n - 1];
int classes = l.classes;
char buff[1024];
char *type = option_find_str(options, "eval", "voc");
FILE *fp = 0;
FILE **fps = 0;
int coco = 0;
int imagenet = 0;
if (0 == strcmp(type, "coco")) {
if (!outfile) outfile = "coco_results";
snprintf(buff, 1024, "%s/%s.json", prefix, outfile);
fp = fopen(buff, "w");
fprintf(fp, "[\n");
coco = 1;
}
else if (0 == strcmp(type, "imagenet")) {
if (!outfile) outfile = "imagenet-detection";
snprintf(buff, 1024, "%s/%s.txt", prefix, outfile);
fp = fopen(buff, "w");
imagenet = 1;
classes = 200;
}
else {
if (!outfile) outfile = "comp4_det_test_";
fps = (FILE**)calloc(classes, sizeof(FILE*));
for (j = 0; j < classes; ++j) {
snprintf(buff, 1024, "%s/%s%s.txt", prefix, outfile, names[j]);
fps[j] = fopen(buff, "w");
}
}
int m = plist->size;
int i = 0;
int t;
float thresh = .005;
float nms = .45;
int nthreads = 4;
if (m < 4) nthreads = m;
image* val = (image*)calloc(nthreads, sizeof(image));
image* val_resized = (image*)calloc(nthreads, sizeof(image));
image* buf = (image*)calloc(nthreads, sizeof(image));
image* buf_resized = (image*)calloc(nthreads, sizeof(image));
pthread_t* thr = (pthread_t*)calloc(nthreads, sizeof(pthread_t));
load_args args = { 0 };
args.w = net.w;
args.h = net.h;
args.c = net.c;
args.type = IMAGE_DATA;
//args.type = LETTERBOX_DATA;
for (t = 0; t < nthreads; ++t) {
args.path = paths[i + t];
args.im = &buf[t];
args.resized = &buf_resized[t];
thr[t] = load_data_in_thread(args);
}
time_t start = time(0);
for (i = nthreads; i < m + nthreads; i += nthreads) {
fprintf(stderr, "%d\n", i);
for (t = 0; t < nthreads && i + t - nthreads < m; ++t) {
pthread_join(thr[t], 0);
val[t] = buf[t];
val_resized[t] = buf_resized[t];
}
for (t = 0; t < nthreads && i + t < m; ++t) {
args.path = paths[i + t];
args.im = &buf[t];
args.resized = &buf_resized[t];
thr[t] = load_data_in_thread(args);
}
for (t = 0; t < nthreads && i + t - nthreads < m; ++t) {
char *path = paths[i + t - nthreads];
char *id = basecfg(path);
float *X = val_resized[t].data;
network_predict(net, X);
int w = val[t].w;
int h = val[t].h;
int nboxes = 0;
int letterbox = (args.type == LETTERBOX_DATA);
detection *dets = get_network_boxes(&net, w, h, thresh, .5, map, 0, &nboxes, letterbox);
if (nms) do_nms_sort(dets, nboxes, classes, nms);
if (coco) {
print_cocos(fp, path, dets, nboxes, classes, w, h);
}
else if (imagenet) {
print_imagenet_detections(fp, i + t - nthreads + 1, dets, nboxes, classes, w, h);
}
else {
print_detector_detections(fps, id, dets, nboxes, classes, w, h);
}
free_detections(dets, nboxes);
free(id);
free_image(val[t]);
free_image(val_resized[t]);
}
}
for (j = 0; j < classes; ++j) {
if (fps) fclose(fps[j]);
}
if (coco) {
fseek(fp, -2, SEEK_CUR);
fprintf(fp, "\n]\n");
fclose(fp);
}
fprintf(stderr, "Total Detection Time: %f Seconds\n", (double)time(0) - start);
}
void validate_detector_recall(char *datacfg, char *cfgfile, char *weightfile)
{
network net = parse_network_cfg_custom(cfgfile, 1, 1); // set batch=1
if (weightfile) {
load_weights(&net, weightfile);
}
//set_batch_network(&net, 1);
fuse_conv_batchnorm(net);
srand(time(0));
//list *plist = get_paths("data/coco_val_5k.list");
list *options = read_data_cfg(datacfg);
char *valid_images = option_find_str(options, "valid", "data/train.txt");
list *plist = get_paths(valid_images);
char **paths = (char **)list_to_array(plist);
layer l = net.layers[net.n - 1];
int j, k;
int m = plist->size;
int i = 0;
float thresh = .001;
float iou_thresh = .5;
float nms = .4;
int total = 0;
int correct = 0;
int proposals = 0;
float avg_iou = 0;
for (i = 0; i < m; ++i) {
char *path = paths[i];
image orig = load_image(path, 0, 0, net.c);
image sized = resize_image(orig, net.w, net.h);
char *id = basecfg(path);
network_predict(net, sized.data);
int nboxes = 0;
int letterbox = 0;
detection *dets = get_network_boxes(&net, sized.w, sized.h, thresh, .5, 0, 1, &nboxes, letterbox);
if (nms) do_nms_obj(dets, nboxes, 1, nms);
char labelpath[4096];
replace_image_to_label(path, labelpath);
int num_labels = 0;
box_label *truth = read_boxes(labelpath, &num_labels);
for (k = 0; k < nboxes; ++k) {
if (dets[k].objectness > thresh) {
++proposals;
}
}
for (j = 0; j < num_labels; ++j) {
++total;
box t = { truth[j].x, truth[j].y, truth[j].w, truth[j].h };
float best_iou = 0;
for (k = 0; k < nboxes; ++k) {
float iou = box_iou(dets[k].bbox, t);
if (dets[k].objectness > thresh && iou > best_iou) {
best_iou = iou;
}
}
avg_iou += best_iou;
if (best_iou > iou_thresh) {
++correct;
}
}
//fprintf(stderr, " %s - %s - ", paths[i], labelpath);
fprintf(stderr, "%5d %5d %5d\tRPs/Img: %.2f\tIOU: %.2f%%\tRecall:%.2f%%\n", i, correct, total, (float)proposals / (i + 1), avg_iou * 100 / total, 100.*correct / total);
free(id);
free_image(orig);
free_image(sized);
}
}
typedef struct {
box b;
float p;
int class_id;
int image_index;
int truth_flag;
int unique_truth_index;
} box_prob;
int detections_comparator(const void *pa, const void *pb)
{
box_prob a = *(box_prob *)pa;
box_prob b = *(box_prob *)pb;
float diff = a.p - b.p;
if (diff < 0) return 1;
else if (diff > 0) return -1;
return 0;
}
float validate_detector_map(char *datacfg, char *cfgfile, char *weightfile, float thresh_calc_avg_iou, const float iou_thresh, network *existing_net)
{
int j;
list *options = read_data_cfg(datacfg);
char *valid_images = option_find_str(options, "valid", "data/train.txt");
char *difficult_valid_images = option_find_str(options, "difficult", NULL);
char *name_list = option_find_str(options, "names", "data/names.list");
char **names = get_labels(name_list);
char *mapf = option_find_str(options, "map", 0);
int *map = 0;
if (mapf) map = read_map(mapf);
FILE* reinforcement_fd = NULL;
network net;
int initial_batch;
if (existing_net) {
char *train_images = option_find_str(options, "train", "data/train.txt");
char *valid_images = option_find_str(options, "valid", train_images);
net = *existing_net;
}
else {
net = parse_network_cfg_custom(cfgfile, 1, 1); // set batch=1
if (weightfile) {
load_weights(&net, weightfile);
}
//set_batch_network(&net, 1);
fuse_conv_batchnorm(net);
calculate_binary_weights(net);
}
srand(time(0));
printf("\n calculation mAP (mean average precision)...\n");
list *plist = get_paths(valid_images);
char **paths = (char **)list_to_array(plist);
char **paths_dif = NULL;
if (difficult_valid_images) {
list *plist_dif = get_paths(difficult_valid_images);
paths_dif = (char **)list_to_array(plist_dif);
}
layer l = net.layers[net.n - 1];
int classes = l.classes;
int m = plist->size;
int i = 0;
int t;
const float thresh = .005;
const float nms = .45;
//const float iou_thresh = 0.5;
int nthreads = 4;
if (m < 4) nthreads = m;
image* val = (image*)calloc(nthreads, sizeof(image));
image* val_resized = (image*)calloc(nthreads, sizeof(image));
image* buf = (image*)calloc(nthreads, sizeof(image));
image* buf_resized = (image*)calloc(nthreads, sizeof(image));
pthread_t* thr = (pthread_t*)calloc(nthreads, sizeof(pthread_t));
load_args args = { 0 };
args.w = net.w;
args.h = net.h;
args.c = net.c;
args.type = IMAGE_DATA;
//args.type = LETTERBOX_DATA;
//const float thresh_calc_avg_iou = 0.24;
float avg_iou = 0;
int tp_for_thresh = 0;
int fp_for_thresh = 0;
box_prob* detections = (box_prob*)calloc(1, sizeof(box_prob));
int detections_count = 0;
int unique_truth_count = 0;
int* truth_classes_count = (int*)calloc(classes, sizeof(int));
for (t = 0; t < nthreads; ++t) {
args.path = paths[i + t];
args.im = &buf[t];
args.resized = &buf_resized[t];
thr[t] = load_data_in_thread(args);
}
time_t start = time(0);
for (i = nthreads; i < m + nthreads; i += nthreads) {
fprintf(stderr, "\r%d", i);
for (t = 0; t < nthreads && i + t - nthreads < m; ++t) {
pthread_join(thr[t], 0);
val[t] = buf[t];
val_resized[t] = buf_resized[t];
}
for (t = 0; t < nthreads && i + t < m; ++t) {
args.path = paths[i + t];
args.im = &buf[t];
args.resized = &buf_resized[t];
thr[t] = load_data_in_thread(args);
}
for (t = 0; t < nthreads && i + t - nthreads < m; ++t) {
const int image_index = i + t - nthreads;
char *path = paths[image_index];
char *id = basecfg(path);
float *X = val_resized[t].data;
network_predict(net, X);
int nboxes = 0;
float hier_thresh = 0;
detection *dets;
if (args.type == LETTERBOX_DATA) {
int letterbox = 1;
dets = get_network_boxes(&net, val[t].w, val[t].h, thresh, hier_thresh, 0, 1, &nboxes, letterbox);
}
else {
int letterbox = 0;
dets = get_network_boxes(&net, 1, 1, thresh, hier_thresh, 0, 0, &nboxes, letterbox);
}
//detection *dets = get_network_boxes(&net, val[t].w, val[t].h, thresh, hier_thresh, 0, 1, &nboxes, letterbox); // for letterbox=1
if (nms) do_nms_sort(dets, nboxes, l.classes, nms);
char labelpath[4096];
replace_image_to_label(path, labelpath);
int num_labels = 0;
box_label *truth = read_boxes(labelpath, &num_labels);
int i, j;
for (j = 0; j < num_labels; ++j) {
truth_classes_count[truth[j].id]++;
}
// difficult
box_label *truth_dif = NULL;
int num_labels_dif = 0;
if (paths_dif)
{
char *path_dif = paths_dif[image_index];
char labelpath_dif[4096];
replace_image_to_label(path_dif, labelpath_dif);
truth_dif = read_boxes(labelpath_dif, &num_labels_dif);
}
const int checkpoint_detections_count = detections_count;
for (i = 0; i < nboxes; ++i) {
int class_id;
for (class_id = 0; class_id < classes; ++class_id) {
float prob = dets[i].prob[class_id];
if (prob > 0) {
detections_count++;
detections = (box_prob*)realloc(detections, detections_count * sizeof(box_prob));
detections[detections_count - 1].b = dets[i].bbox;
detections[detections_count - 1].p = prob;
detections[detections_count - 1].image_index = image_index;
detections[detections_count - 1].class_id = class_id;
detections[detections_count - 1].truth_flag = 0;
detections[detections_count - 1].unique_truth_index = -1;
int truth_index = -1;
float max_iou = 0;
for (j = 0; j < num_labels; ++j)
{
box t = { truth[j].x, truth[j].y, truth[j].w, truth[j].h };
//printf(" IoU = %f, prob = %f, class_id = %d, truth[j].id = %d \n",
// box_iou(dets[i].bbox, t), prob, class_id, truth[j].id);
float current_iou = box_iou(dets[i].bbox, t);
if (current_iou > iou_thresh && class_id == truth[j].id) {
if (current_iou > max_iou) {
max_iou = current_iou;
truth_index = unique_truth_count + j;
}
}
}
// best IoU
if (truth_index > -1) {
detections[detections_count - 1].truth_flag = 1;
detections[detections_count - 1].unique_truth_index = truth_index;
}
else {
// if object is difficult then remove detection
for (j = 0; j < num_labels_dif; ++j) {
box t = { truth_dif[j].x, truth_dif[j].y, truth_dif[j].w, truth_dif[j].h };
float current_iou = box_iou(dets[i].bbox, t);
if (current_iou > iou_thresh && class_id == truth_dif[j].id) {
--detections_count;
break;
}
}
}
// calc avg IoU, true-positives, false-positives for required Threshold
if (prob > thresh_calc_avg_iou) {
int z, found = 0;
for (z = checkpoint_detections_count; z < detections_count - 1; ++z)
if (detections[z].unique_truth_index == truth_index) {
found = 1; break;
}
if (truth_index > -1 && found == 0) {
avg_iou += max_iou;
++tp_for_thresh;
}
else
fp_for_thresh++;
}
}
}
}
unique_truth_count += num_labels;
//static int previous_errors = 0;
//int total_errors = fp_for_thresh + (unique_truth_count - tp_for_thresh);
//int errors_in_this_image = total_errors - previous_errors;
//previous_errors = total_errors;
//if(reinforcement_fd == NULL) reinforcement_fd = fopen("reinforcement.txt", "wb");
//char buff[1000];
//sprintf(buff, "%s\n", path);
//if(errors_in_this_image > 0) fwrite(buff, sizeof(char), strlen(buff), reinforcement_fd);
free_detections(dets, nboxes);
free(id);
free_image(val[t]);
free_image(val_resized[t]);
}
}
if ((tp_for_thresh + fp_for_thresh) > 0)
avg_iou = avg_iou / (tp_for_thresh + fp_for_thresh);
// SORT(detections)
qsort(detections, detections_count, sizeof(box_prob), detections_comparator);
typedef struct {
double precision;
double recall;
int tp, fp, fn;
} pr_t;
// for PR-curve
pr_t** pr = (pr_t**)calloc(classes, sizeof(pr_t*));
for (i = 0; i < classes; ++i) {
pr[i] = (pr_t*)calloc(detections_count, sizeof(pr_t));
}
printf("\n detections_count = %d, unique_truth_count = %d \n", detections_count, unique_truth_count);
int* truth_flags = (int*)calloc(unique_truth_count, sizeof(int));
int rank;
for (rank = 0; rank < detections_count; ++rank) {
if (rank % 100 == 0)
printf(" rank = %d of ranks = %d \r", rank, detections_count);
if (rank > 0) {
int class_id;
for (class_id = 0; class_id < classes; ++class_id) {
pr[class_id][rank].tp = pr[class_id][rank - 1].tp;
pr[class_id][rank].fp = pr[class_id][rank - 1].fp;
}
}
box_prob d = detections[rank];
// if (detected && isn't detected before)
if (d.truth_flag == 1) {
if (truth_flags[d.unique_truth_index] == 0)
{
truth_flags[d.unique_truth_index] = 1;
pr[d.class_id][rank].tp++; // true-positive
}
}
else {
pr[d.class_id][rank].fp++; // false-positive
}
for (i = 0; i < classes; ++i)
{
const int tp = pr[i][rank].tp;
const int fp = pr[i][rank].fp;
const int fn = truth_classes_count[i] - tp; // false-negative = objects - true-positive
pr[i][rank].fn = fn;
if ((tp + fp) > 0) pr[i][rank].precision = (double)tp / (double)(tp + fp);
else pr[i][rank].precision = 0;
if ((tp + fn) > 0) pr[i][rank].recall = (double)tp / (double)(tp + fn);
else pr[i][rank].recall = 0;
}
}
free(truth_flags);
double mean_average_precision = 0;
for (i = 0; i < classes; ++i) {
double avg_precision = 0;
int point;
for (point = 0; point < 11; ++point) {
double cur_recall = point * 0.1;
double cur_precision = 0;
for (rank = 0; rank < detections_count; ++rank)
{
if (pr[i][rank].recall >= cur_recall) { // > or >=
if (pr[i][rank].precision > cur_precision) {
cur_precision = pr[i][rank].precision;
}
}
}
//printf("class_id = %d, point = %d, cur_recall = %.4f, cur_precision = %.4f \n", i, point, cur_recall, cur_precision);
avg_precision += cur_precision;
}
avg_precision = avg_precision / 11;
printf("class_id = %d, name = %s, \t ap = %2.2f %% \n", i, names[i], avg_precision * 100);
mean_average_precision += avg_precision;
}
const float cur_precision = (float)tp_for_thresh / ((float)tp_for_thresh + (float)fp_for_thresh);
const float cur_recall = (float)tp_for_thresh / ((float)tp_for_thresh + (float)(unique_truth_count - tp_for_thresh));
const float f1_score = 2.F * cur_precision * cur_recall / (cur_precision + cur_recall);
printf(" for thresh = %1.2f, precision = %1.2f, recall = %1.2f, F1-score = %1.2f \n",
thresh_calc_avg_iou, cur_precision, cur_recall, f1_score);
printf(" for thresh = %0.2f, TP = %d, FP = %d, FN = %d, average IoU = %2.2f %% \n",
thresh_calc_avg_iou, tp_for_thresh, fp_for_thresh, unique_truth_count - tp_for_thresh, avg_iou * 100);
mean_average_precision = mean_average_precision / classes;
printf("\n IoU threshold = %2.0f %% \n", iou_thresh * 100);
printf(" mean average precision (mAP@%0.2f) = %f, or %2.2f %% \n", iou_thresh, mean_average_precision, mean_average_precision * 100);
for (i = 0; i < classes; ++i) {
free(pr[i]);
}
free(pr);
free(detections);
free(truth_classes_count);
fprintf(stderr, "Total Detection Time: %f Seconds\n", (double)(time(0) - start));
if (reinforcement_fd != NULL) fclose(reinforcement_fd);
// free memory
free_ptrs((void**)names, net.layers[net.n - 1].classes);
free_list_contents_kvp(options);
free_list(options);
if (existing_net) {
//set_batch_network(&net, initial_batch);
}
else {
free_network(net);
}
return mean_average_precision;
}
//#ifdef OPENCV
typedef struct {
float w, h;
} anchors_t;
int anchors_comparator(const void *pa, const void *pb)
{
anchors_t a = *(anchors_t *)pa;
anchors_t b = *(anchors_t *)pb;
float diff = b.w*b.h - a.w*a.h;
if (diff < 0) return 1;
else if (diff > 0) return -1;
return 0;
}
int anchors_data_comparator(const float **pa, const float **pb)
{
float *a = (float *)*pa;
float *b = (float *)*pb;
float diff = b[0] * b[1] - a[0] * a[1];
if (diff < 0) return 1;
else if (diff > 0) return -1;
return 0;
}
void calc_anchors(char *datacfg, int num_of_clusters, int width, int height, int show)
{
printf("\n num_of_clusters = %d, width = %d, height = %d \n", num_of_clusters, width, height);
if (width < 0 || height < 0) {
printf("Usage: darknet detector calc_anchors data/voc.data -num_of_clusters 9 -width 416 -height 416 \n");
printf("Error: set width and height \n");
return;
}
//float pointsdata[] = { 1,1, 2,2, 6,6, 5,5, 10,10 };
float* rel_width_height_array = (float*)calloc(1000, sizeof(float));
list *options = read_data_cfg(datacfg);
char *train_images = option_find_str(options, "train", "data/train.list");
list *plist = get_paths(train_images);
int number_of_images = plist->size;
char **paths = (char **)list_to_array(plist);
srand(time(0));
int number_of_boxes = 0;
printf(" read labels from %d images \n", number_of_images);
int i, j;
for (i = 0; i < number_of_images; ++i) {
char *path = paths[i];
char labelpath[4096];
replace_image_to_label(path, labelpath);
int num_labels = 0;
box_label *truth = read_boxes(labelpath, &num_labels);
//printf(" new path: %s \n", labelpath);
char buff[1024];
for (j = 0; j < num_labels; ++j)
{
if (truth[j].x > 1 || truth[j].x <= 0 || truth[j].y > 1 || truth[j].y <= 0 ||
truth[j].w > 1 || truth[j].w <= 0 || truth[j].h > 1 || truth[j].h <= 0)
{
printf("\n\nWrong label: %s - j = %d, x = %f, y = %f, width = %f, height = %f \n",
labelpath, j, truth[j].x, truth[j].y, truth[j].w, truth[j].h);
sprintf(buff, "echo \"Wrong label: %s - j = %d, x = %f, y = %f, width = %f, height = %f\" >> bad_label.list",
labelpath, j, truth[j].x, truth[j].y, truth[j].w, truth[j].h);
system(buff);
if (check_mistakes) getchar();
}
number_of_boxes++;
rel_width_height_array = (float*)realloc(rel_width_height_array, 2 * number_of_boxes * sizeof(float));
rel_width_height_array[number_of_boxes * 2 - 2] = truth[j].w * width;
rel_width_height_array[number_of_boxes * 2 - 1] = truth[j].h * height;
printf("\r loaded \t image: %d \t box: %d", i + 1, number_of_boxes);
}
}
printf("\n all loaded. \n");
printf("\n calculating k-means++ ...");
matrix boxes_data;
model anchors_data;
boxes_data = make_matrix(number_of_boxes, 2);
printf("\n");
for (i = 0; i < number_of_boxes; ++i) {
float w = boxes_data.vals[i][0] = rel_width_height_array[i * 2];
float h = boxes_data.vals[i][1] = rel_width_height_array[i * 2 + 1];
//if (w > 410 || h > 410) printf("i:%d, w = %f, h = %f \n", i, w, h);
}
// Is used: distance(box, centroid) = 1 - IoU(box, centroid)
// K-means
anchors_data = do_kmeans(boxes_data, num_of_clusters);
qsort((void*)anchors_data.centers.vals, num_of_clusters, 2 * sizeof(float), (__compar_fn_t)anchors_data_comparator);
//gen_anchors.py = 1.19, 1.99, 2.79, 4.60, 4.53, 8.92, 8.06, 5.29, 10.32, 10.66
//float orig_anch[] = { 1.19, 1.99, 2.79, 4.60, 4.53, 8.92, 8.06, 5.29, 10.32, 10.66 };
printf("\n");
float avg_iou = 0;
for (i = 0; i < number_of_boxes; ++i) {
float box_w = rel_width_height_array[i * 2]; //points->data.fl[i * 2];
float box_h = rel_width_height_array[i * 2 + 1]; //points->data.fl[i * 2 + 1];
//int cluster_idx = labels->data.i[i];
int cluster_idx = 0;
float min_dist = FLT_MAX;
float best_iou = 0;
for (j = 0; j < num_of_clusters; ++j) {
float anchor_w = anchors_data.centers.vals[j][0]; // centers->data.fl[j * 2];
float anchor_h = anchors_data.centers.vals[j][1]; // centers->data.fl[j * 2 + 1];
float min_w = (box_w < anchor_w) ? box_w : anchor_w;
float min_h = (box_h < anchor_h) ? box_h : anchor_h;
float box_intersect = min_w*min_h;
float box_union = box_w*box_h + anchor_w*anchor_h - box_intersect;
float iou = box_intersect / box_union;
float distance = 1 - iou;
if (distance < min_dist) min_dist = distance, cluster_idx = j, best_iou = iou;
}
float anchor_w = anchors_data.centers.vals[cluster_idx][0]; //centers->data.fl[cluster_idx * 2];
float anchor_h = anchors_data.centers.vals[cluster_idx][1]; //centers->data.fl[cluster_idx * 2 + 1];
if (best_iou > 1 || best_iou < 0) { // || box_w > width || box_h > height) {
printf(" Wrong label: i = %d, box_w = %d, box_h = %d, anchor_w = %d, anchor_h = %d, iou = %f \n",
i, box_w, box_h, anchor_w, anchor_h, best_iou);
}
else avg_iou += best_iou;
}
avg_iou = 100 * avg_iou / number_of_boxes;
printf("\n avg IoU = %2.2f %% \n", avg_iou);
char buff[1024];
FILE* fw = fopen("anchors.txt", "wb");
if (fw) {
printf("\nSaving anchors to the file: anchors.txt \n");
printf("anchors = ");
for (i = 0; i < num_of_clusters; ++i) {
float anchor_w = anchors_data.centers.vals[i][0]; //centers->data.fl[i * 2];
float anchor_h = anchors_data.centers.vals[i][1]; //centers->data.fl[i * 2 + 1];
if (width > 32) sprintf(buff, "%3.0f,%3.0f", anchor_w, anchor_h);
else sprintf(buff, "%2.4f,%2.4f", anchor_w, anchor_h);
printf("%s", buff);
fwrite(buff, sizeof(char), strlen(buff), fw);
if (i + 1 < num_of_clusters) {
fwrite(", ", sizeof(char), 2, fw);
printf(", ");
}
}
printf("\n");
fclose(fw);
}
else {
printf(" Error: file anchors.txt can't be open \n");
}
if (show) {
#ifdef OPENCV
CvMat* labels = cvCreateMat(number_of_boxes, 1, CV_32SC1);
CvMat* points = cvCreateMat(number_of_boxes, 2, CV_32FC1);
CvMat* centers = cvCreateMat(num_of_clusters, 2, CV_32FC1);
for (i = 0; i < number_of_boxes; ++i) {
points->data.fl[i * 2] = rel_width_height_array[i * 2];
points->data.fl[i * 2 + 1] = rel_width_height_array[i * 2 + 1];
//cvSet1D(points, i * 2, cvScalar(rel_width_height_array[i * 2], 0, 0, 0));
//cvSet1D(points, i * 2 + 1, cvScalar(rel_width_height_array[i * 2 + 1], 0, 0, 0));
}
for (i = 0; i < num_of_clusters; ++i) {
centers->data.fl[i * 2] = anchors_data.centers.vals[i][0];
centers->data.fl[i * 2 + 1] = anchors_data.centers.vals[i][1];
}
for (i = 0; i < number_of_boxes; ++i) {
labels->data.i[i] = anchors_data.assignments[i];
}
size_t img_size = 700;
IplImage* img = cvCreateImage(cvSize(img_size, img_size), 8, 3);
cvZero(img);
for (j = 0; j < num_of_clusters; ++j) {
CvPoint pt1, pt2;
pt1.x = pt1.y = 0;
pt2.x = centers->data.fl[j * 2] * img_size / width;
pt2.y = centers->data.fl[j * 2 + 1] * img_size / height;
cvRectangle(img, pt1, pt2, CV_RGB(255, 255, 255), 1, 8, 0);
}
for (i = 0; i < number_of_boxes; ++i) {
CvPoint pt;
pt.x = points->data.fl[i * 2] * img_size / width;
pt.y = points->data.fl[i * 2 + 1] * img_size / height;
int cluster_idx = labels->data.i[i];
int red_id = (cluster_idx * (uint64_t)123 + 55) % 255;
int green_id = (cluster_idx * (uint64_t)321 + 33) % 255;
int blue_id = (cluster_idx * (uint64_t)11 + 99) % 255;
cvCircle(img, pt, 1, CV_RGB(red_id, green_id, blue_id), CV_FILLED, 8, 0);
//if(pt.x > img_size || pt.y > img_size) printf("\n pt.x = %d, pt.y = %d \n", pt.x, pt.y);
}
cvShowImage("clusters", img);
cvWaitKey(0);
cvReleaseImage(&img);
cvDestroyAllWindows();
cvReleaseMat(&labels);
cvReleaseMat(&points);
cvReleaseMat(¢ers);
#endif // OPENCV
}
free(rel_width_height_array);
getchar();
}
//#else
//void calc_anchors(char *datacfg, int num_of_clusters, int width, int height, int show) {
// printf(" k-means++ can't be used without OpenCV, because there is used cvKMeans2 implementation \n");
//}
//#endif // OPENCV
void test_detector(char *datacfg, char *cfgfile, char *weightfile, char *filename, float thresh,
float hier_thresh, int dont_show, int ext_output, int save_labels, char *outfile)
{
list *options = read_data_cfg(datacfg);
char *name_list = option_find_str(options, "names", "data/names.list");
int names_size = 0;
char **names = get_labels_custom(name_list, &names_size); //get_labels(name_list);
image **alphabet = load_alphabet();
network net = parse_network_cfg_custom(cfgfile, 1, 1); // set batch=1
if (weightfile) {
load_weights(&net, weightfile);
}
//set_batch_network(&net, 1);
fuse_conv_batchnorm(net);
calculate_binary_weights(net);
if (net.layers[net.n - 1].classes != names_size) {
printf(" Error: in the file %s number of names %d that isn't equal to classes=%d in the file %s \n",
name_list, names_size, net.layers[net.n - 1].classes, cfgfile);
if (net.layers[net.n - 1].classes > names_size) getchar();
}
srand(2222222);
double time;
char buff[256];
char *input = buff;
char *json_buf = NULL;
int json_image_id = 0;
FILE* json_file = NULL;
if (outfile) {
json_file = fopen(outfile, "wb");
char *tmp = "[\n";
fwrite(tmp, sizeof(char), strlen(tmp), json_file);
}
int j;
float nms = .45; // 0.4F
while (1) {
if (filename) {
strncpy(input, filename, 256);
if (strlen(input) > 0)
if (input[strlen(input) - 1] == 0x0d) input[strlen(input) - 1] = 0;
}
else {
printf("Enter Image Path: ");
fflush(stdout);
input = fgets(input, 256, stdin);
if (!input) break;
strtok(input, "\n");
}
//image im;
//image sized = load_image_resize(input, net.w, net.h, net.c, &im);
image im = load_image(input, 0, 0, net.c);
image sized = resize_image(im, net.w, net.h);
int letterbox = 0;
//image sized = letterbox_image(im, net.w, net.h); letterbox = 1;
layer l = net.layers[net.n - 1];
//box *boxes = calloc(l.w*l.h*l.n, sizeof(box));
//float **probs = calloc(l.w*l.h*l.n, sizeof(float*));
//for(j = 0; j < l.w*l.h*l.n; ++j) probs[j] = (float*)calloc(l.classes, sizeof(float));
float *X = sized.data;
//time= what_time_is_it_now();
double time = get_time_point();
network_predict(net, X);
//network_predict_image(&net, im); letterbox = 1;
printf("%s: Predicted in %lf milli-seconds.\n", input, ((double)get_time_point() - time) / 1000);
//printf("%s: Predicted in %f seconds.\n", input, (what_time_is_it_now()-time));
int nboxes = 0;
detection *dets = get_network_boxes(&net, im.w, im.h, thresh, hier_thresh, 0, 1, &nboxes, letterbox);
if (nms) do_nms_sort(dets, nboxes, l.classes, nms);
draw_detections_v3(im, dets, nboxes, thresh, names, alphabet, l.classes, ext_output);
save_image(im, "predictions");
if (!dont_show) {
show_image(im, "predictions");
}
if (outfile) {
if (json_buf) {
char *tmp = ", \n";
fwrite(tmp, sizeof(char), strlen(tmp), json_file);
}
++json_image_id;
json_buf = detection_to_json(dets, nboxes, l.classes, names, json_image_id, input);
fwrite(json_buf, sizeof(char), strlen(json_buf), json_file);
free(json_buf);
}
// pseudo labeling concept - fast.ai
if (save_labels)
{
char labelpath[4096];
replace_image_to_label(input, labelpath);
FILE* fw = fopen(labelpath, "wb");
int i;
for (i = 0; i < nboxes; ++i) {
char buff[1024];
int class_id = -1;
float prob = 0;
for (j = 0; j < l.classes; ++j) {
if (dets[i].prob[j] > thresh && dets[i].prob[j] > prob) {
prob = dets[i].prob[j];
class_id = j;
}
}
if (class_id >= 0) {
sprintf(buff, "%d %2.4f %2.4f %2.4f %2.4f\n", class_id, dets[i].bbox.x, dets[i].bbox.y, dets[i].bbox.w, dets[i].bbox.h);
fwrite(buff, sizeof(char), strlen(buff), fw);
}
}
fclose(fw);
}
free_detections(dets, nboxes);
free_image(im);
free_image(sized);
//free(boxes);
//free_ptrs((void **)probs, l.w*l.h*l.n);
#ifdef OPENCV
if (!dont_show) {
cvWaitKey(0);
cvDestroyAllWindows();
}
#endif
if (filename) break;
}
if (outfile) {
char *tmp = "\n]";
fwrite(tmp, sizeof(char), strlen(tmp), json_file);
fclose(json_file);
}
// free memory
free_ptrs((void**)names, net.layers[net.n - 1].classes);
free_list_contents_kvp(options);
free_list(options);
int i;
const int nsize = 8;
for (j = 0; j < nsize; ++j) {
for (i = 32; i < 127; ++i) {
free_image(alphabet[j][i]);
}
free(alphabet[j]);
}
free(alphabet);
free_network(net);
}
void run_detector(int argc, char **argv)
{
int dont_show = find_arg(argc, argv, "-dont_show");
int show = find_arg(argc, argv, "-show");
int calc_map = find_arg(argc, argv, "-map");
check_mistakes = find_arg(argc, argv, "-check_mistakes");
int mjpeg_port = find_int_arg(argc, argv, "-mjpeg_port", -1);
int json_port = find_int_arg(argc, argv, "-json_port", -1);
char *out_filename = find_char_arg(argc, argv, "-out_filename", 0);
char *outfile = find_char_arg(argc, argv, "-out", 0);
char *prefix = find_char_arg(argc, argv, "-prefix", 0);
float thresh = find_float_arg(argc, argv, "-thresh", .25); // 0.24
float iou_thresh = find_float_arg(argc, argv, "-iou_thresh", .5); // 0.5 for mAP
float hier_thresh = find_float_arg(argc, argv, "-hier", .5);
int cam_index = find_int_arg(argc, argv, "-c", 0);
int frame_skip = find_int_arg(argc, argv, "-s", 0);
int num_of_clusters = find_int_arg(argc, argv, "-num_of_clusters", 5);
int width = find_int_arg(argc, argv, "-width", -1);
int height = find_int_arg(argc, argv, "-height", -1);
// extended output in test mode (output of rect bound coords)
// and for recall mode (extended output table-like format with results for best_class fit)
int ext_output = find_arg(argc, argv, "-ext_output");
int save_labels = find_arg(argc, argv, "-save_labels");
if (argc < 4) {
fprintf(stderr, "usage: %s %s [train/test/valid/demo/map] [data] [cfg] [weights (optional)]\n", argv[0], argv[1]);
return;
}
char *gpu_list = find_char_arg(argc, argv, "-gpus", 0);
int *gpus = 0;
int gpu = 0;
int ngpus = 0;
if (gpu_list) {
printf("%s\n", gpu_list);
int len = strlen(gpu_list);
ngpus = 1;
int i;
for (i = 0; i < len; ++i) {
if (gpu_list[i] == ',') ++ngpus;
}
gpus = (int*)calloc(ngpus, sizeof(int));
for (i = 0; i < ngpus; ++i) {
gpus[i] = atoi(gpu_list);
gpu_list = strchr(gpu_list, ',') + 1;
}
}
else {
gpu = gpu_index;
gpus = &gpu;
ngpus = 1;
}
int clear = find_arg(argc, argv, "-clear");
char *datacfg = argv[3];
char *cfg = argv[4];
char *weights = (argc > 5) ? argv[5] : 0;
if (weights)
if (strlen(weights) > 0)
if (weights[strlen(weights) - 1] == 0x0d) weights[strlen(weights) - 1] = 0;
char *filename = (argc > 6) ? argv[6] : 0;
if (0 == strcmp(argv[2], "test")) test_detector(datacfg, cfg, weights, filename, thresh, hier_thresh, dont_show, ext_output, save_labels, outfile);
else if (0 == strcmp(argv[2], "train")) train_detector(datacfg, cfg, weights, gpus, ngpus, clear, dont_show, calc_map, mjpeg_port);
else if (0 == strcmp(argv[2], "valid")) validate_detector(datacfg, cfg, weights, outfile);
else if (0 == strcmp(argv[2], "recall")) validate_detector_recall(datacfg, cfg, weights);
else if (0 == strcmp(argv[2], "map")) validate_detector_map(datacfg, cfg, weights, thresh, iou_thresh, NULL);
else if (0 == strcmp(argv[2], "calc_anchors")) calc_anchors(datacfg, num_of_clusters, width, height, show);
else if (0 == strcmp(argv[2], "demo")) {
list *options = read_data_cfg(datacfg);
int classes = option_find_int(options, "classes", 20);
char *name_list = option_find_str(options, "names", "data/names.list");
char **names = get_labels(name_list);
if (filename)
if (strlen(filename) > 0)
if (filename[strlen(filename) - 1] == 0x0d) filename[strlen(filename) - 1] = 0;
demo(cfg, weights, thresh, hier_thresh, cam_index, filename, names, classes, frame_skip, prefix, out_filename,
mjpeg_port, json_port, dont_show, ext_output);
free_list_contents_kvp(options);
free_list(options);
}
else printf(" There isn't such command: %s", argv[2]);
}
(3) demo.c
此文件主要是实现darknet的实时监测功能(即调用摄像头检测,需编译opencv)。
#include "network.h"
#include "detection_layer.h"
#include "region_layer.h"
#include "cost_layer.h"
#include "utils.h"
#include "parser.h"
#include "box.h"
#include "image.h"
#include "demo.h"
#ifdef WIN32
#include <time.h>
#include "gettimeofday.h"
#else
#include <sys/time.h>
#endif
#ifdef OPENCV
#include <opencv2/highgui/highgui_c.h>
#include <opencv2/imgproc/imgproc_c.h>
#include <opencv2/core/version.hpp>
#ifndef CV_VERSION_EPOCH
#include <opencv2/videoio/videoio_c.h>
#endif
#include "http_stream.h"
image get_image_from_stream(CvCapture *cap);
static char **demo_names;
static image **demo_alphabet;
static int demo_classes;
static int nboxes = 0;
static detection *dets = NULL;
static network net;
static image in_s ;
static image det_s;
static CvCapture * cap;
static int cpp_video_capture = 0;
static float fps = 0;
static float demo_thresh = 0;
static int demo_ext_output = 0;
static long long int frame_id = 0;
static int demo_json_port = -1;
static float* predictions[NFRAMES];
static int demo_index = 0;
static image images[NFRAMES];
static IplImage* ipl_images[NFRAMES];
static float *avg;
void draw_detections_cv_v3(IplImage* show_img, detection *dets, int num, float thresh, char **names, image **alphabet, int classes, int ext_output);
void show_image_cv_ipl(IplImage *disp, const char *name);
void save_cv_png(IplImage *img, const char *name);
void save_cv_jpg(IplImage *img, const char *name);
image get_image_from_stream_resize(CvCapture *cap, int w, int h, int c, IplImage** in_img, int cpp_video_capture, int dont_close);
image get_image_from_stream_letterbox(CvCapture *cap, int w, int h, int c, IplImage** in_img, int cpp_video_capture, int dont_close);
int get_stream_fps(CvCapture *cap, int cpp_video_capture);
IplImage* in_img;
IplImage* det_img;
IplImage* show_img;
static int flag_exit;
static int letter_box = 0;
void *fetch_in_thread(void *ptr)
{
//in = get_image_from_stream(cap);
int dont_close_stream = 0; // set 1 if your IP-camera periodically turns off and turns on video-stream
if(letter_box)
in_s = get_image_from_stream_letterbox(cap, net.w, net.h, net.c, &in_img, cpp_video_capture, dont_close_stream);
else
in_s = get_image_from_stream_resize(cap, net.w, net.h, net.c, &in_img, cpp_video_capture, dont_close_stream);
if(!in_s.data){
//error("Stream closed.");
printf("Stream closed.\n");
flag_exit = 1;
exit(EXIT_FAILURE);
}
//in_s = resize_image(in, net.w, net.h);
return 0;
}
void *detect_in_thread(void *ptr)
{
layer l = net.layers[net.n-1];
float *X = det_s.data;
float *prediction = network_predict(net, X);
memcpy(predictions[demo_index], prediction, l.outputs*sizeof(float));
mean_arrays(predictions, NFRAMES, l.outputs, avg);
l.output = avg;
free_image(det_s);
ipl_images[demo_index] = det_img;
det_img = ipl_images[(demo_index + NFRAMES / 2 + 1) % NFRAMES];
demo_index = (demo_index + 1) % NFRAMES;
if (letter_box)
dets = get_network_boxes(&net, in_img->width, in_img->height, demo_thresh, demo_thresh, 0, 1, &nboxes, 1); // letter box
else
dets = get_network_boxes(&net, net.w, net.h, demo_thresh, demo_thresh, 0, 1, &nboxes, 0); // resized
return 0;
}
double get_wall_time()
{
struct timeval walltime;
if (gettimeofday(&walltime, NULL)) {
return 0;
}
return (double)walltime.tv_sec + (double)walltime.tv_usec * .000001;
}
void demo(char *cfgfile, char *weightfile, float thresh, float hier_thresh, int cam_index, const char *filename, char **names, int classes,
int frame_skip, char *prefix, char *out_filename, int mjpeg_port, int json_port, int dont_show, int ext_output)
{
in_img = det_img = show_img = NULL;
//skip = frame_skip;
image **alphabet = load_alphabet();
int delay = frame_skip;
demo_names = names;
demo_alphabet = alphabet;
demo_classes = classes;
demo_thresh = thresh;
demo_ext_output = ext_output;
demo_json_port = json_port;
printf("Demo\n");
net = parse_network_cfg_custom(cfgfile, 1, 1); // set batch=1
if(weightfile){
load_weights(&net, weightfile);
}
//set_batch_network(&net, 1);
fuse_conv_batchnorm(net);
calculate_binary_weights(net);
srand(2222222);
if(filename){
printf("video file: %s\n", filename);
cpp_video_capture = 1;
cap = get_capture_video_stream(filename);
}else{
printf("Webcam index: %d\n", cam_index);
cpp_video_capture = 1;
cap = get_capture_webcam(cam_index);
}
if (!cap) {
#ifdef WIN32
printf("Check that you have copied file opencv_ffmpeg340_64.dll to the same directory where is darknet.exe \n");
#endif
error("Couldn't connect to webcam.\n");
}
layer l = net.layers[net.n-1];
int j;
avg = (float *) calloc(l.outputs, sizeof(float));
for(j = 0; j < NFRAMES; ++j) predictions[j] = (float *) calloc(l.outputs, sizeof(float));
for(j = 0; j < NFRAMES; ++j) images[j] = make_image(1,1,3);
if (l.classes != demo_classes) {
printf("Parameters don't match: in cfg-file classes=%d, in data-file classes=%d \n", l.classes, demo_classes);
getchar();
exit(0);
}
flag_exit = 0;
pthread_t fetch_thread;
pthread_t detect_thread;
fetch_in_thread(0);
det_img = in_img;
det_s = in_s;
fetch_in_thread(0);
detect_in_thread(0);
det_img = in_img;
det_s = in_s;
for (j = 0; j < NFRAMES / 2; ++j) {
fetch_in_thread(0);
detect_in_thread(0);
det_img = in_img;
det_s = in_s;
}
int count = 0;
if(!prefix && !dont_show){
cvNamedWindow("Demo", CV_WINDOW_NORMAL);
cvMoveWindow("Demo", 0, 0);
cvResizeWindow("Demo", 1352, 1013);
}
CvVideoWriter* output_video_writer = NULL; // cv::VideoWriter output_video;
if (out_filename && !flag_exit)
{
CvSize size;
size.width = det_img->width, size.height = det_img->height;
int src_fps = 25;
src_fps = get_stream_fps(cap, cpp_video_capture);
//const char* output_name = "test_dnn_out.avi";
//output_video_writer = cvCreateVideoWriter(out_filename, CV_FOURCC('H', '2', '6', '4'), src_fps, size, 1);
output_video_writer = cvCreateVideoWriter(out_filename, CV_FOURCC('D', 'I', 'V', 'X'), src_fps, size, 1);
//output_video_writer = cvCreateVideoWriter(out_filename, CV_FOURCC('M', 'J', 'P', 'G'), src_fps, size, 1);
//output_video_writer = cvCreateVideoWriter(out_filename, CV_FOURCC('M', 'P', '4', 'V'), src_fps, size, 1);
//output_video_writer = cvCreateVideoWriter(out_filename, CV_FOURCC('M', 'P', '4', '2'), src_fps, size, 1);
//output_video_writer = cvCreateVideoWriter(out_filename, CV_FOURCC('X', 'V', 'I', 'D'), src_fps, size, 1);
//output_video_writer = cvCreateVideoWriter(out_filename, CV_FOURCC('W', 'M', 'V', '2'), src_fps, size, 1);
}
double before = get_wall_time();
while(1){
++count;
{
if(pthread_create(&fetch_thread, 0, fetch_in_thread, 0)) error("Thread creation failed");
if(pthread_create(&detect_thread, 0, detect_in_thread, 0)) error("Thread creation failed");
float nms = .45; // 0.4F
int local_nboxes = nboxes;
detection *local_dets = dets;
//if (nms) do_nms_obj(local_dets, local_nboxes, l.classes, nms); // bad results
if (nms) do_nms_sort(local_dets, local_nboxes, l.classes, nms);
printf("\033[2J");
printf("\033[1;1H");
printf("\nFPS:%.1f\n", fps);
printf("Objects:\n\n");
++frame_id;
if (demo_json_port > 0) {
int timeout = 400000;
send_json(local_dets, local_nboxes, l.classes, demo_names, frame_id, demo_json_port, timeout);
}
draw_detections_cv_v3(show_img, local_dets, local_nboxes, demo_thresh, demo_names, demo_alphabet, demo_classes, demo_ext_output);
free_detections(local_dets, local_nboxes);
if(!prefix){
if (!dont_show) {
show_image_cv_ipl(show_img, "Demo");
int c = cvWaitKey(1);
if (c == 10) {
if (frame_skip == 0) frame_skip = 60;
else if (frame_skip == 4) frame_skip = 0;
else if (frame_skip == 60) frame_skip = 4;
else frame_skip = 0;
}
else if (c == 27 || c == 1048603) // ESC - exit (OpenCV 2.x / 3.x)
{
flag_exit = 1;
}
}
}else{
char buff[256];
sprintf(buff, "%s_%08d.jpg", prefix, count);
if(show_img) save_cv_jpg(show_img, buff);
}
// if you run it with param -mjpeg_port 8090 then open URL in your web-browser: http://localhost:8090
if (mjpeg_port > 0 && show_img) {
int port = mjpeg_port;
int timeout = 400000;
int jpeg_quality = 40; // 1 - 100
send_mjpeg(show_img, port, timeout, jpeg_quality);
}
// save video file
if (output_video_writer && show_img) {
cvWriteFrame(output_video_writer, show_img);
printf("\n cvWriteFrame \n");
}
cvReleaseImage(&show_img);
pthread_join(fetch_thread, 0);
pthread_join(detect_thread, 0);
if (flag_exit == 1) break;
if(delay == 0){
show_img = det_img;
}
det_img = in_img;
det_s = in_s;
}
--delay;
if(delay < 0){
delay = frame_skip;
//double after = get_wall_time();
//float curr = 1./(after - before);
double after = get_time_point(); // more accurate time measurements
float curr = 1000000. / (after - before);
fps = curr;
before = after;
}
}
printf("input video stream closed. \n");
if (output_video_writer) {
cvReleaseVideoWriter(&output_video_writer);
printf("output_video_writer closed. \n");
}
// free memory
cvReleaseImage(&show_img);
cvReleaseImage(&in_img);
free_image(in_s);
free(avg);
for (j = 0; j < NFRAMES; ++j) free(predictions[j]);
for (j = 0; j < NFRAMES; ++j) free_image(images[j]);
free_ptrs((void **)names, net.layers[net.n - 1].classes);
int i;
const int nsize = 8;
for (j = 0; j < nsize; ++j) {
for (i = 32; i < 127; ++i) {
free_image(alphabet[j][i]);
}
free(alphabet[j]);
}
free(alphabet);
free_network(net);
//cudaProfilerStop();
}
#else
void demo(char *cfgfile, char *weightfile, float thresh, float hier_thresh, int cam_index, const char *filename, char **names, int classes,
int frame_skip, char *prefix, char *out_filename, int mjpeg_port, int json_port, int dont_show, int ext_output)
{
fprintf(stderr, "Demo needs OpenCV for webcam images.\n");
}
#endif