void forward_region_layer(const region_layer l, network_state state)
{
int i,j,b,t,n;
// 25
int size = l.coords + l.classes + 1;
memcpy(l.output, state.input, l.outputs*l.batch*sizeof(float));
#ifndef GPU
flatten(l.output, l.w*l.h, size*l.n, l.batch, 1);
#endif
// 每個預測框的置信度
for (b = 0; b < l.batch; ++b){
for(i = 0; i < l.h*l.w*l.n; ++i){
int index = size*i + b*l.outputs;
l.output[index + 4] = logistic_activate(l.output[index + 4]);
}
}
#ifndef GPU
if (l.softmax_tree){
for (b = 0; b < l.batch; ++b){
for(i = 0; i < l.h*l.w*l.n; ++i){
int index = size*i + b*l.outputs;
softmax_tree(l.output + index + 5, 1, 0, 1, l.softmax_tree, l.output + index + 5);
}
}
} else if (l.softmax){
for (b = 0; b < l.batch; ++b){
for(i = 0; i < l.h*l.w*l.n; ++i){
int index = size*i + b*l.outputs;
softmax(l.output + index + 5, l.classes, 1, l.output + index + 5, 1);
}
}
}
#endif
if(!state.train) return;
memset(l.delta, 0, l.outputs * l.batch * sizeof(float));
float avg_iou = 0;
float recall = 0;
float avg_cat = 0;
float avg_obj = 0;
float avg_anyobj = 0;
int count = 0;
int class_count = 0;
*(l.cost) = 0;
for (b = 0; b < l.batch; ++b) {
if(l.softmax_tree){
int onlyclass_id = 0;
// x,y,w,h,class_id
for(t = 0; t < l.max_boxes; ++t){
box truth = float_to_box(state.truth + t*5 + b*l.truths);
if(!truth.x) break; // continue;
int class_id = state.truth[t*5 + b*l.truths + 4];
float maxp = 0;
int maxi = 0;
// ???
if(truth.x > 100000 && truth.y > 100000){
for(n = 0; n < l.n*l.w*l.h; ++n){
int index = size*n + b*l.outputs + 5;
float scale = l.output[index-1];
float p = scale*get_hierarchy_probability(l.output + index, l.softmax_tree, class_id);
if(p > maxp){
maxp = p;
maxi = n;
}
}
int index = size*maxi + b*l.outputs + 5;
// 類別的誤差t-o,證明輸出到誤差經歷了mse
delta_region_class(l.output, l.delta, index, class_id, l.classes, l.softmax_tree, l.class_scale, &avg_cat, l.focal_loss);
++class_count;
onlyclass_id = 1;
break;
}
}
if(onlyclass_id) continue;
}
// output組織方式 : h(13)*w(13)*n(5)*size(25)
for (j = 0; j < l.h; ++j) {
for (i = 0; i < l.w; ++i) {
for (n = 0; n < l.n; ++n) {
int index = size*(j*l.w*l.n + i*l.n + n) + b*l.outputs;
// 25中的0,1,2,3:x,y,w,h,l.biases就是anchor
box pred = get_region_box(l.output, l.biases, n, index, i, j, l.w, l.h);
float best_iou = 0;
int best_class_id = -1;
// 每個預測框與所有ground truth計算IOU
for(t = 0; t < l.max_boxes; ++t){
box truth = float_to_box(state.truth + t*5 + b*l.truths);
int class_id = state.truth[t * 5 + b*l.truths + 4];
if (class_id >= l.classes) continue; // if label contains class_id more than number of classes in the cfg-file
if(!truth.x) break; // continue;
float iou = box_iou(pred, truth);
if (iou > best_iou) {
best_class_id = state.truth[t*5 + b*l.truths + 4];
best_iou = iou;
}
}
// 25中的4 : confidence
avg_anyobj += l.output[index + 4];
// logistic_gradient是sigmoid的導數,置信度誤差
// (t-o)*o*(1-o) 證明輸出到誤差經歷了sigmoid和mse
l.delta[index + 4] = l.noobject_scale * ((0 - l.output[index + 4]) * logistic_gradient(l.output[index + 4]));
if(l.classfix == -1) l.delta[index + 4] = l.noobject_scale * ((best_iou - l.output[index + 4]) * logistic_gradient(l.output[index + 4]));
else{
// 大於閾值不參與反向傳播
if (best_iou > l.thresh) {
l.delta[index + 4] = 0;
if(l.classfix > 0){
delta_region_class(l.output, l.delta, index + 5, best_class_id, l.classes, l.softmax_tree, l.class_scale*(l.classfix == 2 ? l.output[index + 4] : 1), &avg_cat, l.focal_loss);
++class_count;
}
}
}
if(*(state.net.seen) < 12800){
box truth = {0};
truth.x = (i + .5)/l.w;
truth.y = (j + .5)/l.h;
truth.w = l.biases[2*n];
truth.h = l.biases[2*n+1];
if(DOABS){
truth.w = l.biases[2*n]/l.w;
truth.h = l.biases[2*n+1]/l.h;
}
// 同時計算IOU和delta,座標的誤差
// x,y的誤差經歷了sigmoid和mse; w,h的誤差經歷了mse
delta_region_box(truth, l.output, l.biases, n, index, i, j, l.w, l.h, l.delta, .01);
}
}
}
}
for(t = 0; t < l.max_boxes; ++t){
box truth = float_to_box(state.truth + t*5 + b*l.truths);
int class_id = state.truth[t * 5 + b*l.truths + 4];
if (class_id >= l.classes) {
printf("\n Warning: in txt-labels class_id=%d >= classes=%d in cfg-file. In txt-labels class_id should be [from 0 to %d] \n", class_id, l.classes, l.classes-1);
getchar();
continue; // if label contains class_id more than number of classes in the cfg-file
}
if(!truth.x) break; // continue;
float best_iou = 0;
int best_index = 0;
int best_n = 0;
// cell的座標
i = (truth.x * l.w);
j = (truth.y * l.h);
//printf("%d %f %d %f\n", i, truth.x*l.w, j, truth.y*l.h);
box truth_shift = truth;
// 移動到原點,爲了計算IOU,預測框同理
truth_shift.x = 0;
truth_shift.y = 0;
//printf("index %d %d\n",i, j);
// 遍歷ground truth所在cell的5個預測框或先驗框anchor,找出IOU最大的
for(n = 0; n < l.n; ++n){
int index = size*(j*l.w*l.n + i*l.n + n) + b*l.outputs;
box pred = get_region_box(l.output, l.biases, n, index, i, j, l.w, l.h);
if(l.bias_match){
pred.w = l.biases[2*n];
pred.h = l.biases[2*n+1];
if(DOABS){
pred.w = l.biases[2*n]/l.w;
pred.h = l.biases[2*n+1]/l.h;
}
}
//printf("pred: (%f, %f) %f x %f\n", pred.x, pred.y, pred.w, pred.h);
// 移動到原點,爲了計算IOU,對應上面
pred.x = 0;
pred.y = 0;
float iou = box_iou(pred, truth_shift);
if (iou > best_iou){
best_index = index;
best_iou = iou;
best_n = n;
}
}
//printf("%d %f (%f, %f) %f x %f\n", best_n, best_iou, truth.x, truth.y, truth.w, truth.h);
// 同時計算IOU和delta
float iou = delta_region_box(truth, l.output, l.biases, best_n, best_index, i, j, l.w, l.h, l.delta, l.coord_scale);
if(iou > .5) recall += 1;
avg_iou += iou;
//l.delta[best_index + 4] = iou - l.output[best_index + 4];
avg_obj += l.output[best_index + 4];
// (t-o)*o*(1-o) 證明輸出到誤差經歷了sigmoid和mse
l.delta[best_index + 4] = l.object_scale * (1 - l.output[best_index + 4]) * logistic_gradient(l.output[best_index + 4]);
if (l.rescore) {
l.delta[best_index + 4] = l.object_scale * (iou - l.output[best_index + 4]) * logistic_gradient(l.output[best_index + 4]);
}
if (l.map) class_id = l.map[class_id];
// 一般只計算delta
delta_region_class(l.output, l.delta, best_index + 5, class_id, l.classes, l.softmax_tree, l.class_scale, &avg_cat, l.focal_loss);
++count;
++class_count;
}
}
//printf("\n");
#ifndef GPU
flatten(l.delta, l.w*l.h, size*l.n, l.batch, 0);
#endif
*(l.cost) = pow(mag_array(l.delta, l.outputs * l.batch), 2);
printf("Region Avg IOU: %f, Class: %f, Obj: %f, No Obj: %f, Avg Recall: %f, count: %d\n", avg_iou/count, avg_cat/class_count, avg_obj/count, avg_anyobj/(l.w*l.h*l.n*l.batch), recall/count, count);
}
darknet之region_layer源碼分析
發表評論
所有評論
還沒有人評論,想成為第一個評論的人麼? 請在上方評論欄輸入並且點擊發布.