派生自 Algorithm/baseDetector
lib/detecter_tools/darknet/yolo_layer.c
@@ -1,909 +1,1229 @@ #include "yolo_layer.h" #include "activations.h" #include "blas.h" #include "box.h" #include "dark_cuda.h" #include "utils.h" #include <math.h> #include <stdio.h> #include <assert.h> #include <string.h> #include <stdlib.h> extern int check_mistakes; layer make_yolo_layer(int batch, int w, int h, int n, int total, int *mask, int classes, int max_boxes) { int i; layer l = { (LAYER_TYPE)0 }; l.type = YOLO; l.n = n; l.total = total; l.batch = batch; l.h = h; l.w = w; l.c = n*(classes + 4 + 1); l.out_w = l.w; l.out_h = l.h; l.out_c = l.c; l.classes = classes; l.cost = (float*)xcalloc(1, sizeof(float)); l.biases = (float*)xcalloc(total * 2, sizeof(float)); if(mask) l.mask = mask; else{ l.mask = (int*)xcalloc(n, sizeof(int)); for(i = 0; i < n; ++i){ l.mask[i] = i; } } l.bias_updates = (float*)xcalloc(n * 2, sizeof(float)); l.outputs = h*w*n*(classes + 4 + 1); l.inputs = l.outputs; l.max_boxes = max_boxes; l.truths = l.max_boxes*(4 + 1); // 90*(4 + 1); l.delta = (float*)xcalloc(batch * l.outputs, sizeof(float)); l.output = (float*)xcalloc(batch * l.outputs, sizeof(float)); for(i = 0; i < total*2; ++i){ l.biases[i] = .5; } l.forward = forward_yolo_layer; l.backward = backward_yolo_layer; #ifdef GPU l.forward_gpu = forward_yolo_layer_gpu; l.backward_gpu = backward_yolo_layer_gpu; l.output_gpu = cuda_make_array(l.output, batch*l.outputs); l.output_avg_gpu = cuda_make_array(l.output, batch*l.outputs); l.delta_gpu = cuda_make_array(l.delta, batch*l.outputs); free(l.output); if (cudaSuccess == cudaHostAlloc(&l.output, batch*l.outputs*sizeof(float), cudaHostRegisterMapped)) l.output_pinned = 1; else { cudaGetLastError(); // reset CUDA-error l.output = (float*)xcalloc(batch * l.outputs, sizeof(float)); } free(l.delta); if (cudaSuccess == cudaHostAlloc(&l.delta, batch*l.outputs*sizeof(float), cudaHostRegisterMapped)) l.delta_pinned = 1; else { cudaGetLastError(); // reset CUDA-error l.delta = (float*)xcalloc(batch * l.outputs, sizeof(float)); } #endif fprintf(stderr, "yolo\n"); srand(time(0)); return l; } void resize_yolo_layer(layer *l, int w, int h) { l->w = w; l->h = h; l->outputs = h*w*l->n*(l->classes + 4 + 1); l->inputs = l->outputs; if (!l->output_pinned) l->output = (float*)xrealloc(l->output, l->batch*l->outputs * sizeof(float)); if (!l->delta_pinned) l->delta = (float*)xrealloc(l->delta, l->batch*l->outputs*sizeof(float)); #ifdef GPU if (l->output_pinned) { CHECK_CUDA(cudaFreeHost(l->output)); if (cudaSuccess != cudaHostAlloc(&l->output, l->batch*l->outputs * sizeof(float), cudaHostRegisterMapped)) { cudaGetLastError(); // reset CUDA-error l->output = (float*)xcalloc(l->batch * l->outputs, sizeof(float)); l->output_pinned = 0; } } if (l->delta_pinned) { CHECK_CUDA(cudaFreeHost(l->delta)); if (cudaSuccess != cudaHostAlloc(&l->delta, l->batch*l->outputs * sizeof(float), cudaHostRegisterMapped)) { cudaGetLastError(); // reset CUDA-error l->delta = (float*)xcalloc(l->batch * l->outputs, sizeof(float)); l->delta_pinned = 0; } } cuda_free(l->delta_gpu); cuda_free(l->output_gpu); cuda_free(l->output_avg_gpu); l->delta_gpu = cuda_make_array(l->delta, l->batch*l->outputs); l->output_gpu = cuda_make_array(l->output, l->batch*l->outputs); l->output_avg_gpu = cuda_make_array(l->output, l->batch*l->outputs); #endif } box get_yolo_box(float *x, float *biases, int n, int index, int i, int j, int lw, int lh, int w, int h, int stride) { box b; // ln - natural logarithm (base = e) // x` = t.x * lw - i; // x = ln(x`/(1-x`)) // x - output of previous conv-layer // y` = t.y * lh - i; // y = ln(y`/(1-y`)) // y - output of previous conv-layer // w = ln(t.w * net.w / anchors_w); // w - output of previous conv-layer // h = ln(t.h * net.h / anchors_h); // h - output of previous conv-layer b.x = (i + x[index + 0*stride]) / lw; b.y = (j + x[index + 1*stride]) / lh; b.w = exp(x[index + 2*stride]) * biases[2*n] / w; b.h = exp(x[index + 3*stride]) * biases[2*n+1] / h; return b; } static inline float fix_nan_inf(float val) { if (isnan(val) || isinf(val)) val = 0; return val; } static inline float clip_value(float val, const float max_val) { if (val > max_val) { //printf("\n val = %f > max_val = %f \n", val, max_val); val = max_val; } else if (val < -max_val) { //printf("\n val = %f < -max_val = %f \n", val, -max_val); val = -max_val; } return val; } ious delta_yolo_box(box truth, float *x, float *biases, int n, int index, int i, int j, int lw, int lh, int w, int h, float *delta, float scale, int stride, float iou_normalizer, IOU_LOSS iou_loss, int accumulate, float max_delta) { ious all_ious = { 0 }; // i - step in layer width // j - step in layer height // Returns a box in absolute coordinates box pred = get_yolo_box(x, biases, n, index, i, j, lw, lh, w, h, stride); all_ious.iou = box_iou(pred, truth); all_ious.giou = box_giou(pred, truth); all_ious.diou = box_diou(pred, truth); all_ious.ciou = box_ciou(pred, truth); // avoid nan in dx_box_iou if (pred.w == 0) { pred.w = 1.0; } if (pred.h == 0) { pred.h = 1.0; } if (iou_loss == MSE) // old loss { float tx = (truth.x*lw - i); float ty = (truth.y*lh - j); float tw = log(truth.w*w / biases[2 * n]); float th = log(truth.h*h / biases[2 * n + 1]); //printf(" tx = %f, ty = %f, tw = %f, th = %f \n", tx, ty, tw, th); //printf(" x = %f, y = %f, w = %f, h = %f \n", x[index + 0 * stride], x[index + 1 * stride], x[index + 2 * stride], x[index + 3 * stride]); // accumulate delta delta[index + 0 * stride] += scale * (tx - x[index + 0 * stride]) * iou_normalizer; delta[index + 1 * stride] += scale * (ty - x[index + 1 * stride]) * iou_normalizer; delta[index + 2 * stride] += scale * (tw - x[index + 2 * stride]) * iou_normalizer; delta[index + 3 * stride] += scale * (th - x[index + 3 * stride]) * iou_normalizer; } else { // https://github.com/generalized-iou/g-darknet // https://arxiv.org/abs/1902.09630v2 // https://giou.stanford.edu/ all_ious.dx_iou = dx_box_iou(pred, truth, iou_loss); // jacobian^t (transpose) //float dx = (all_ious.dx_iou.dl + all_ious.dx_iou.dr); //float dy = (all_ious.dx_iou.dt + all_ious.dx_iou.db); //float dw = ((-0.5 * all_ious.dx_iou.dl) + (0.5 * all_ious.dx_iou.dr)); //float dh = ((-0.5 * all_ious.dx_iou.dt) + (0.5 * all_ious.dx_iou.db)); // jacobian^t (transpose) float dx = all_ious.dx_iou.dt; float dy = all_ious.dx_iou.db; float dw = all_ious.dx_iou.dl; float dh = all_ious.dx_iou.dr; // predict exponential, apply gradient of e^delta_t ONLY for w,h dw *= exp(x[index + 2 * stride]); dh *= exp(x[index + 3 * stride]); // normalize iou weight dx *= iou_normalizer; dy *= iou_normalizer; dw *= iou_normalizer; dh *= iou_normalizer; dx = fix_nan_inf(dx); dy = fix_nan_inf(dy); dw = fix_nan_inf(dw); dh = fix_nan_inf(dh); if (max_delta != FLT_MAX) { dx = clip_value(dx, max_delta); dy = clip_value(dy, max_delta); dw = clip_value(dw, max_delta); dh = clip_value(dh, max_delta); } if (!accumulate) { delta[index + 0 * stride] = 0; delta[index + 1 * stride] = 0; delta[index + 2 * stride] = 0; delta[index + 3 * stride] = 0; } // accumulate delta delta[index + 0 * stride] += dx; delta[index + 1 * stride] += dy; delta[index + 2 * stride] += dw; delta[index + 3 * stride] += dh; } return all_ious; } void averages_yolo_deltas(int class_index, int box_index, int stride, int classes, float *delta) { int classes_in_one_box = 0; int c; for (c = 0; c < classes; ++c) { if (delta[class_index + stride*c] > 0) classes_in_one_box++; } if (classes_in_one_box > 0) { delta[box_index + 0 * stride] /= classes_in_one_box; delta[box_index + 1 * stride] /= classes_in_one_box; delta[box_index + 2 * stride] /= classes_in_one_box; delta[box_index + 3 * stride] /= classes_in_one_box; } } void delta_yolo_class(float *output, float *delta, int index, int class_id, int classes, int stride, float *avg_cat, int focal_loss, float label_smooth_eps, float *classes_multipliers) { int n; if (delta[index + stride*class_id]){ float y_true = 1; if(label_smooth_eps) y_true = y_true * (1 - label_smooth_eps) + 0.5*label_smooth_eps; float result_delta = y_true - output[index + stride*class_id]; if(!isnan(result_delta) && !isinf(result_delta)) delta[index + stride*class_id] = result_delta; //delta[index + stride*class_id] = 1 - output[index + stride*class_id]; if (classes_multipliers) delta[index + stride*class_id] *= classes_multipliers[class_id]; if(avg_cat) *avg_cat += output[index + stride*class_id]; return; } // Focal loss if (focal_loss) { // Focal Loss float alpha = 0.5; // 0.25 or 0.5 //float gamma = 2; // hardcoded in many places of the grad-formula int ti = index + stride*class_id; float pt = output[ti] + 0.000000000000001F; // http://fooplot.com/#W3sidHlwZSI6MCwiZXEiOiItKDEteCkqKDIqeCpsb2coeCkreC0xKSIsImNvbG9yIjoiIzAwMDAwMCJ9LHsidHlwZSI6MTAwMH1d float grad = -(1 - pt) * (2 * pt*logf(pt) + pt - 1); // http://blog.csdn.net/linmingan/article/details/77885832 //float grad = (1 - pt) * (2 * pt*logf(pt) + pt - 1); // https://github.com/unsky/focal-loss for (n = 0; n < classes; ++n) { delta[index + stride*n] = (((n == class_id) ? 1 : 0) - output[index + stride*n]); delta[index + stride*n] *= alpha*grad; if (n == class_id && avg_cat) *avg_cat += output[index + stride*n]; } } else { // default for (n = 0; n < classes; ++n) { float y_true = ((n == class_id) ? 1 : 0); if (label_smooth_eps) y_true = y_true * (1 - label_smooth_eps) + 0.5*label_smooth_eps; float result_delta = y_true - output[index + stride*n]; if (!isnan(result_delta) && !isinf(result_delta)) delta[index + stride*n] = result_delta; if (classes_multipliers && n == class_id) delta[index + stride*class_id] *= classes_multipliers[class_id]; if (n == class_id && avg_cat) *avg_cat += output[index + stride*n]; } } } int compare_yolo_class(float *output, int classes, int class_index, int stride, float objectness, int class_id, float conf_thresh) { int j; for (j = 0; j < classes; ++j) { //float prob = objectness * output[class_index + stride*j]; float prob = output[class_index + stride*j]; if (prob > conf_thresh) { return 1; } } return 0; } static int entry_index(layer l, int batch, int location, int entry) { int n = location / (l.w*l.h); int loc = location % (l.w*l.h); return batch*l.outputs + n*l.w*l.h*(4+l.classes+1) + entry*l.w*l.h + loc; } void forward_yolo_layer(const layer l, network_state state) { int i, j, b, t, n; memcpy(l.output, state.input, l.outputs*l.batch * sizeof(float)); #ifndef GPU for (b = 0; b < l.batch; ++b) { for (n = 0; n < l.n; ++n) { int index = entry_index(l, b, n*l.w*l.h, 0); activate_array(l.output + index, 2 * l.w*l.h, LOGISTIC); // x,y, scal_add_cpu(2 * l.w*l.h, l.scale_x_y, -0.5*(l.scale_x_y - 1), l.output + index, 1); // scale x,y index = entry_index(l, b, n*l.w*l.h, 4); activate_array(l.output + index, (1 + l.classes)*l.w*l.h, LOGISTIC); } } #endif // delta is zeroed memset(l.delta, 0, l.outputs * l.batch * sizeof(float)); if (!state.train) return; //float avg_iou = 0; float tot_iou = 0; float tot_giou = 0; float tot_diou = 0; float tot_ciou = 0; float tot_iou_loss = 0; float tot_giou_loss = 0; float tot_diou_loss = 0; float tot_ciou_loss = 0; float recall = 0; float recall75 = 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) { for (j = 0; j < l.h; ++j) { for (i = 0; i < l.w; ++i) { for (n = 0; n < l.n; ++n) { const int class_index = entry_index(l, b, n*l.w*l.h + j*l.w + i, 4 + 1); const int obj_index = entry_index(l, b, n*l.w*l.h + j*l.w + i, 4); const int box_index = entry_index(l, b, n*l.w*l.h + j*l.w + i, 0); const int stride = l.w*l.h; box pred = get_yolo_box(l.output, l.biases, l.mask[n], box_index, i, j, l.w, l.h, state.net.w, state.net.h, l.w*l.h); float best_match_iou = 0; int best_match_t = 0; float best_iou = 0; int best_t = 0; for (t = 0; t < l.max_boxes; ++t) { box truth = float_to_box_stride(state.truth + t*(4 + 1) + b*l.truths, 1); int class_id = state.truth[t*(4 + 1) + b*l.truths + 4]; if (class_id >= l.classes || class_id < 0) { 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); printf("\n truth.x = %f, truth.y = %f, truth.w = %f, truth.h = %f, class_id = %d \n", truth.x, truth.y, truth.w, truth.h, class_id); if (check_mistakes) getchar(); continue; // if label contains class_id more than number of classes in the cfg-file and class_id check garbage value } if (!truth.x) break; // continue; float objectness = l.output[obj_index]; if (isnan(objectness) || isinf(objectness)) l.output[obj_index] = 0; int class_id_match = compare_yolo_class(l.output, l.classes, class_index, l.w*l.h, objectness, class_id, 0.25f); float iou = box_iou(pred, truth); if (iou > best_match_iou && class_id_match == 1) { best_match_iou = iou; best_match_t = t; } if (iou > best_iou) { best_iou = iou; best_t = t; } } avg_anyobj += l.output[obj_index]; l.delta[obj_index] = l.cls_normalizer * (0 - l.output[obj_index]); if (best_match_iou > l.ignore_thresh) { const float iou_multiplier = best_match_iou*best_match_iou;// (best_match_iou - l.ignore_thresh) / (1.0 - l.ignore_thresh); if (l.objectness_smooth) { l.delta[obj_index] = l.cls_normalizer * (iou_multiplier - l.output[obj_index]); int class_id = state.truth[best_match_t*(4 + 1) + b*l.truths + 4]; if (l.map) class_id = l.map[class_id]; const float class_multiplier = (l.classes_multipliers) ? l.classes_multipliers[class_id] : 1.0f; l.delta[class_index + stride*class_id] = class_multiplier * (iou_multiplier - l.output[class_index + stride*class_id]); } else l.delta[obj_index] = 0; } else if (state.net.adversarial) { int stride = l.w*l.h; float scale = pred.w * pred.h; if (scale > 0) scale = sqrt(scale); l.delta[obj_index] = scale * l.cls_normalizer * (0 - l.output[obj_index]); int cl_id; for (cl_id = 0; cl_id < l.classes; ++cl_id) { if(l.output[class_index + stride*cl_id] * l.output[obj_index] > 0.25) l.delta[class_index + stride*cl_id] = scale * (0 - l.output[class_index + stride*cl_id]); } } if (best_iou > l.truth_thresh) { const float iou_multiplier = best_iou*best_iou;// (best_iou - l.truth_thresh) / (1.0 - l.truth_thresh); if (l.objectness_smooth) l.delta[obj_index] = l.cls_normalizer * (iou_multiplier - l.output[obj_index]); else l.delta[obj_index] = l.cls_normalizer * (1 - l.output[obj_index]); //l.delta[obj_index] = l.cls_normalizer * (1 - l.output[obj_index]); int class_id = state.truth[best_t*(4 + 1) + b*l.truths + 4]; if (l.map) class_id = l.map[class_id]; delta_yolo_class(l.output, l.delta, class_index, class_id, l.classes, l.w*l.h, 0, l.focal_loss, l.label_smooth_eps, l.classes_multipliers); const float class_multiplier = (l.classes_multipliers) ? l.classes_multipliers[class_id] : 1.0f; if (l.objectness_smooth) l.delta[class_index + stride*class_id] = class_multiplier * (iou_multiplier - l.output[class_index + stride*class_id]); box truth = float_to_box_stride(state.truth + best_t*(4 + 1) + b*l.truths, 1); delta_yolo_box(truth, l.output, l.biases, l.mask[n], box_index, i, j, l.w, l.h, state.net.w, state.net.h, l.delta, (2 - truth.w*truth.h), l.w*l.h, l.iou_normalizer * class_multiplier, l.iou_loss, 1, l.max_delta); } } } } for (t = 0; t < l.max_boxes; ++t) { box truth = float_to_box_stride(state.truth + t*(4 + 1) + b*l.truths, 1); if (truth.x < 0 || truth.y < 0 || truth.x > 1 || truth.y > 1 || truth.w < 0 || truth.h < 0) { char buff[256]; printf(" Wrong label: truth.x = %f, truth.y = %f, truth.w = %f, truth.h = %f \n", truth.x, truth.y, truth.w, truth.h); sprintf(buff, "echo \"Wrong label: truth.x = %f, truth.y = %f, truth.w = %f, truth.h = %f\" >> bad_label.list", truth.x, truth.y, truth.w, truth.h); system(buff); } int class_id = state.truth[t*(4 + 1) + b*l.truths + 4]; if (class_id >= l.classes || class_id < 0) continue; // if label contains class_id more than number of classes in the cfg-file and class_id check garbage value if (!truth.x) break; // continue; float best_iou = 0; int best_n = 0; i = (truth.x * l.w); j = (truth.y * l.h); box truth_shift = truth; truth_shift.x = truth_shift.y = 0; for (n = 0; n < l.total; ++n) { box pred = { 0 }; pred.w = l.biases[2 * n] / state.net.w; pred.h = l.biases[2 * n + 1] / state.net.h; float iou = box_iou(pred, truth_shift); if (iou > best_iou) { best_iou = iou; best_n = n; } } int mask_n = int_index(l.mask, best_n, l.n); if (mask_n >= 0) { int class_id = state.truth[t*(4 + 1) + b*l.truths + 4]; if (l.map) class_id = l.map[class_id]; int box_index = entry_index(l, b, mask_n*l.w*l.h + j*l.w + i, 0); const float class_multiplier = (l.classes_multipliers) ? l.classes_multipliers[class_id] : 1.0f; ious all_ious = delta_yolo_box(truth, l.output, l.biases, best_n, box_index, i, j, l.w, l.h, state.net.w, state.net.h, l.delta, (2 - truth.w*truth.h), l.w*l.h, l.iou_normalizer * class_multiplier, l.iou_loss, 1, l.max_delta); // range is 0 <= 1 tot_iou += all_ious.iou; tot_iou_loss += 1 - all_ious.iou; // range is -1 <= giou <= 1 tot_giou += all_ious.giou; tot_giou_loss += 1 - all_ious.giou; tot_diou += all_ious.diou; tot_diou_loss += 1 - all_ious.diou; tot_ciou += all_ious.ciou; tot_ciou_loss += 1 - all_ious.ciou; int obj_index = entry_index(l, b, mask_n*l.w*l.h + j*l.w + i, 4); avg_obj += l.output[obj_index]; l.delta[obj_index] = class_multiplier * l.cls_normalizer * (1 - l.output[obj_index]); int class_index = entry_index(l, b, mask_n*l.w*l.h + j*l.w + i, 4 + 1); delta_yolo_class(l.output, l.delta, class_index, class_id, l.classes, l.w*l.h, &avg_cat, l.focal_loss, l.label_smooth_eps, l.classes_multipliers); //printf(" label: class_id = %d, truth.x = %f, truth.y = %f, truth.w = %f, truth.h = %f \n", class_id, truth.x, truth.y, truth.w, truth.h); //printf(" mask_n = %d, l.output[obj_index] = %f, l.output[class_index + class_id] = %f \n\n", mask_n, l.output[obj_index], l.output[class_index + class_id]); ++count; ++class_count; if (all_ious.iou > .5) recall += 1; if (all_ious.iou > .75) recall75 += 1; } // iou_thresh for (n = 0; n < l.total; ++n) { int mask_n = int_index(l.mask, n, l.n); if (mask_n >= 0 && n != best_n && l.iou_thresh < 1.0f) { box pred = { 0 }; pred.w = l.biases[2 * n] / state.net.w; pred.h = l.biases[2 * n + 1] / state.net.h; float iou = box_iou_kind(pred, truth_shift, l.iou_thresh_kind); // IOU, GIOU, MSE, DIOU, CIOU // iou, n if (iou > l.iou_thresh) { int class_id = state.truth[t*(4 + 1) + b*l.truths + 4]; if (l.map) class_id = l.map[class_id]; int box_index = entry_index(l, b, mask_n*l.w*l.h + j*l.w + i, 0); const float class_multiplier = (l.classes_multipliers) ? l.classes_multipliers[class_id] : 1.0f; ious all_ious = delta_yolo_box(truth, l.output, l.biases, n, box_index, i, j, l.w, l.h, state.net.w, state.net.h, l.delta, (2 - truth.w*truth.h), l.w*l.h, l.iou_normalizer * class_multiplier, l.iou_loss, 1, l.max_delta); // range is 0 <= 1 tot_iou += all_ious.iou; tot_iou_loss += 1 - all_ious.iou; // range is -1 <= giou <= 1 tot_giou += all_ious.giou; tot_giou_loss += 1 - all_ious.giou; tot_diou += all_ious.diou; tot_diou_loss += 1 - all_ious.diou; tot_ciou += all_ious.ciou; tot_ciou_loss += 1 - all_ious.ciou; int obj_index = entry_index(l, b, mask_n*l.w*l.h + j*l.w + i, 4); avg_obj += l.output[obj_index]; l.delta[obj_index] = class_multiplier * l.cls_normalizer * (1 - l.output[obj_index]); int class_index = entry_index(l, b, mask_n*l.w*l.h + j*l.w + i, 4 + 1); delta_yolo_class(l.output, l.delta, class_index, class_id, l.classes, l.w*l.h, &avg_cat, l.focal_loss, l.label_smooth_eps, l.classes_multipliers); ++count; ++class_count; if (all_ious.iou > .5) recall += 1; if (all_ious.iou > .75) recall75 += 1; } } } } // averages the deltas obtained by the function: delta_yolo_box()_accumulate for (j = 0; j < l.h; ++j) { for (i = 0; i < l.w; ++i) { for (n = 0; n < l.n; ++n) { int box_index = entry_index(l, b, n*l.w*l.h + j*l.w + i, 0); int class_index = entry_index(l, b, n*l.w*l.h + j*l.w + i, 4 + 1); const int stride = l.w*l.h; averages_yolo_deltas(class_index, box_index, stride, l.classes, l.delta); } } } } if (count == 0) count = 1; if (class_count == 0) class_count = 1; //*(l.cost) = pow(mag_array(l.delta, l.outputs * l.batch), 2); //printf("Region %d Avg IOU: %f, Class: %f, Obj: %f, No Obj: %f, .5R: %f, .75R: %f, count: %d\n", state.index, avg_iou / count, avg_cat / class_count, avg_obj / count, avg_anyobj / (l.w*l.h*l.n*l.batch), recall / count, recall75 / count, count); int stride = l.w*l.h; float* no_iou_loss_delta = (float *)calloc(l.batch * l.outputs, sizeof(float)); memcpy(no_iou_loss_delta, l.delta, l.batch * l.outputs * sizeof(float)); for (b = 0; b < l.batch; ++b) { for (j = 0; j < l.h; ++j) { for (i = 0; i < l.w; ++i) { for (n = 0; n < l.n; ++n) { int index = entry_index(l, b, n*l.w*l.h + j*l.w + i, 0); no_iou_loss_delta[index + 0 * stride] = 0; no_iou_loss_delta[index + 1 * stride] = 0; no_iou_loss_delta[index + 2 * stride] = 0; no_iou_loss_delta[index + 3 * stride] = 0; } } } } float classification_loss = l.cls_normalizer * pow(mag_array(no_iou_loss_delta, l.outputs * l.batch), 2); free(no_iou_loss_delta); float loss = pow(mag_array(l.delta, l.outputs * l.batch), 2); float iou_loss = loss - classification_loss; float avg_iou_loss = 0; // gIOU loss + MSE (objectness) loss if (l.iou_loss == MSE) { *(l.cost) = pow(mag_array(l.delta, l.outputs * l.batch), 2); } else { // Always compute classification loss both for iou + cls loss and for logging with mse loss // TODO: remove IOU loss fields before computing MSE on class // probably split into two arrays if (l.iou_loss == GIOU) { avg_iou_loss = count > 0 ? l.iou_normalizer * (tot_giou_loss / count) : 0; } else { avg_iou_loss = count > 0 ? l.iou_normalizer * (tot_iou_loss / count) : 0; } *(l.cost) = avg_iou_loss + classification_loss; } loss /= l.batch; classification_loss /= l.batch; iou_loss /= l.batch; fprintf(stderr, "v3 (%s loss, Normalizer: (iou: %.2f, cls: %.2f) Region %d Avg (IOU: %f, GIOU: %f), Class: %f, Obj: %f, No Obj: %f, .5R: %f, .75R: %f, count: %d, class_loss = %f, iou_loss = %f, total_loss = %f \n", (l.iou_loss == MSE ? "mse" : (l.iou_loss == GIOU ? "giou" : "iou")), l.iou_normalizer, l.cls_normalizer, state.index, tot_iou / count, tot_giou / count, avg_cat / class_count, avg_obj / count, avg_anyobj / (l.w*l.h*l.n*l.batch), recall / count, recall75 / count, count, classification_loss, iou_loss, loss); } void backward_yolo_layer(const layer l, network_state state) { axpy_cpu(l.batch*l.inputs, 1, l.delta, 1, state.delta, 1); } // Converts output of the network to detection boxes // w,h: image width,height // netw,neth: network width,height // relative: 1 (all callers seems to pass TRUE) void correct_yolo_boxes(detection *dets, int n, int w, int h, int netw, int neth, int relative, int letter) { int i; // network height (or width) int new_w = 0; // network height (or width) int new_h = 0; // Compute scale given image w,h vs network w,h // I think this "rotates" the image to match network to input image w/h ratio // new_h and new_w are really just network width and height if (letter) { if (((float)netw / w) < ((float)neth / h)) { new_w = netw; new_h = (h * netw) / w; } else { new_h = neth; new_w = (w * neth) / h; } } else { new_w = netw; new_h = neth; } // difference between network width and "rotated" width float deltaw = netw - new_w; // difference between network height and "rotated" height float deltah = neth - new_h; // ratio between rotated network width and network width float ratiow = (float)new_w / netw; // ratio between rotated network width and network width float ratioh = (float)new_h / neth; for (i = 0; i < n; ++i) { box b = dets[i].bbox; // x = ( x - (deltaw/2)/netw ) / ratiow; // x - [(1/2 the difference of the network width and rotated width) / (network width)] b.x = (b.x - deltaw / 2. / netw) / ratiow; b.y = (b.y - deltah / 2. / neth) / ratioh; // scale to match rotation of incoming image b.w *= 1 / ratiow; b.h *= 1 / ratioh; // relative seems to always be == 1, I don't think we hit this condition, ever. if (!relative) { b.x *= w; b.w *= w; b.y *= h; b.h *= h; } dets[i].bbox = b; } } /* void correct_yolo_boxes(detection *dets, int n, int w, int h, int netw, int neth, int relative, int letter) { int i; int new_w=0; int new_h=0; if (letter) { if (((float)netw / w) < ((float)neth / h)) { new_w = netw; new_h = (h * netw) / w; } else { new_h = neth; new_w = (w * neth) / h; } } else { new_w = netw; new_h = neth; } for (i = 0; i < n; ++i){ box b = dets[i].bbox; b.x = (b.x - (netw - new_w)/2./netw) / ((float)new_w/netw); b.y = (b.y - (neth - new_h)/2./neth) / ((float)new_h/neth); b.w *= (float)netw/new_w; b.h *= (float)neth/new_h; if(!relative){ b.x *= w; b.w *= w; b.y *= h; b.h *= h; } dets[i].bbox = b; } } */ int yolo_num_detections(layer l, float thresh) { int i, n; int count = 0; for(n = 0; n < l.n; ++n){ for (i = 0; i < l.w*l.h; ++i) { int obj_index = entry_index(l, 0, n*l.w*l.h + i, 4); if(l.output[obj_index] > thresh){ ++count; } } } return count; } int yolo_num_detections_batch(layer l, float thresh, int batch) { int i, n; int count = 0; for (i = 0; i < l.w*l.h; ++i){ for(n = 0; n < l.n; ++n){ int obj_index = entry_index(l, batch, n*l.w*l.h + i, 4); if(l.output[obj_index] > thresh){ ++count; } } } return count; } void avg_flipped_yolo(layer l) { int i,j,n,z; float *flip = l.output + l.outputs; for (j = 0; j < l.h; ++j) { for (i = 0; i < l.w/2; ++i) { for (n = 0; n < l.n; ++n) { for(z = 0; z < l.classes + 4 + 1; ++z){ int i1 = z*l.w*l.h*l.n + n*l.w*l.h + j*l.w + i; int i2 = z*l.w*l.h*l.n + n*l.w*l.h + j*l.w + (l.w - i - 1); float swap = flip[i1]; flip[i1] = flip[i2]; flip[i2] = swap; if(z == 0){ flip[i1] = -flip[i1]; flip[i2] = -flip[i2]; } } } } } for(i = 0; i < l.outputs; ++i){ l.output[i] = (l.output[i] + flip[i])/2.; } } int get_yolo_detections(layer l, int w, int h, int netw, int neth, float thresh, int *map, int relative, detection *dets, int letter) { //printf("\n l.batch = %d, l.w = %d, l.h = %d, l.n = %d \n", l.batch, l.w, l.h, l.n); int i,j,n; float *predictions = l.output; // This snippet below is not necessary // Need to comment it in order to batch processing >= 2 images //if (l.batch == 2) avg_flipped_yolo(l); int count = 0; for (i = 0; i < l.w*l.h; ++i){ int row = i / l.w; int col = i % l.w; for(n = 0; n < l.n; ++n){ int obj_index = entry_index(l, 0, n*l.w*l.h + i, 4); float objectness = predictions[obj_index]; //if(objectness <= thresh) continue; // incorrect behavior for Nan values if (objectness > thresh) { //printf("\n objectness = %f, thresh = %f, i = %d, n = %d \n", objectness, thresh, i, n); int box_index = entry_index(l, 0, n*l.w*l.h + i, 0); dets[count].bbox = get_yolo_box(predictions, l.biases, l.mask[n], box_index, col, row, l.w, l.h, netw, neth, l.w*l.h); dets[count].objectness = objectness; dets[count].classes = l.classes; for (j = 0; j < l.classes; ++j) { int class_index = entry_index(l, 0, n*l.w*l.h + i, 4 + 1 + j); float prob = objectness*predictions[class_index]; dets[count].prob[j] = (prob > thresh) ? prob : 0; } ++count; } } } correct_yolo_boxes(dets, count, w, h, netw, neth, relative, letter); return count; } int get_yolo_detections_batch(layer l, int w, int h, int netw, int neth, float thresh, int *map, int relative, detection *dets, int letter, int batch) { int i,j,n; float *predictions = l.output; //if (l.batch == 2) avg_flipped_yolo(l); int count = 0; for (i = 0; i < l.w*l.h; ++i){ int row = i / l.w; int col = i % l.w; for(n = 0; n < l.n; ++n){ int obj_index = entry_index(l, batch, n*l.w*l.h + i, 4); float objectness = predictions[obj_index]; //if(objectness <= thresh) continue; // incorrect behavior for Nan values if (objectness > thresh) { //printf("\n objectness = %f, thresh = %f, i = %d, n = %d \n", objectness, thresh, i, n); int box_index = entry_index(l, batch, n*l.w*l.h + i, 0); dets[count].bbox = get_yolo_box(predictions, l.biases, l.mask[n], box_index, col, row, l.w, l.h, netw, neth, l.w*l.h); dets[count].objectness = objectness; dets[count].classes = l.classes; for (j = 0; j < l.classes; ++j) { int class_index = entry_index(l, batch, n*l.w*l.h + i, 4 + 1 + j); float prob = objectness*predictions[class_index]; dets[count].prob[j] = (prob > thresh) ? prob : 0; } ++count; } } } correct_yolo_boxes(dets, count, w, h, netw, neth, relative, letter); return count; } #ifdef GPU void forward_yolo_layer_gpu(const layer l, network_state state) { //copy_ongpu(l.batch*l.inputs, state.input, 1, l.output_gpu, 1); simple_copy_ongpu(l.batch*l.inputs, state.input, l.output_gpu); int b, n; for (b = 0; b < l.batch; ++b){ for(n = 0; n < l.n; ++n){ int index = entry_index(l, b, n*l.w*l.h, 0); // y = 1./(1. + exp(-x)) // x = ln(y/(1-y)) // ln - natural logarithm (base = e) // if(y->1) x -> inf // if(y->0) x -> -inf activate_array_ongpu(l.output_gpu + index, 2*l.w*l.h, LOGISTIC); // x,y if (l.scale_x_y != 1) scal_add_ongpu(2 * l.w*l.h, l.scale_x_y, -0.5*(l.scale_x_y - 1), l.output_gpu + index, 1); // scale x,y index = entry_index(l, b, n*l.w*l.h, 4); activate_array_ongpu(l.output_gpu + index, (1+l.classes)*l.w*l.h, LOGISTIC); // classes and objectness } } if(!state.train || l.onlyforward){ //cuda_pull_array(l.output_gpu, l.output, l.batch*l.outputs); if (l.mean_alpha && l.output_avg_gpu) mean_array_gpu(l.output_gpu, l.batch*l.outputs, l.mean_alpha, l.output_avg_gpu); cuda_pull_array_async(l.output_gpu, l.output, l.batch*l.outputs); CHECK_CUDA(cudaPeekAtLastError()); return; } float *in_cpu = (float *)xcalloc(l.batch*l.inputs, sizeof(float)); cuda_pull_array(l.output_gpu, l.output, l.batch*l.outputs); memcpy(in_cpu, l.output, l.batch*l.outputs*sizeof(float)); float *truth_cpu = 0; if (state.truth) { int num_truth = l.batch*l.truths; truth_cpu = (float *)xcalloc(num_truth, sizeof(float)); cuda_pull_array(state.truth, truth_cpu, num_truth); } network_state cpu_state = state; cpu_state.net = state.net; cpu_state.index = state.index; cpu_state.train = state.train; cpu_state.truth = truth_cpu; cpu_state.input = in_cpu; forward_yolo_layer(l, cpu_state); //forward_yolo_layer(l, state); cuda_push_array(l.delta_gpu, l.delta, l.batch*l.outputs); free(in_cpu); if (cpu_state.truth) free(cpu_state.truth); } void backward_yolo_layer_gpu(const layer l, network_state state) { axpy_ongpu(l.batch*l.inputs, state.net.loss_scale, l.delta_gpu, 1, state.delta, 1); } #endif #include "yolo_layer.h" #include "activations.h" #include "blas.h" #include "box.h" #include "dark_cuda.h" #include "utils.h" #include <math.h> #include <stdio.h> #include <assert.h> #include <string.h> #include <stdlib.h> extern int check_mistakes; layer make_yolo_layer(int batch, int w, int h, int n, int total, int *mask, int classes, int max_boxes) { int i; layer l = { (LAYER_TYPE)0 }; l.type = YOLO; l.n = n; l.total = total; l.batch = batch; l.h = h; l.w = w; l.c = n*(classes + 4 + 1); l.out_w = l.w; l.out_h = l.h; l.out_c = l.c; l.classes = classes; l.cost = (float*)xcalloc(1, sizeof(float)); l.biases = (float*)xcalloc(total * 2, sizeof(float)); if(mask) l.mask = mask; else{ l.mask = (int*)xcalloc(n, sizeof(int)); for(i = 0; i < n; ++i){ l.mask[i] = i; } } l.bias_updates = (float*)xcalloc(n * 2, sizeof(float)); l.outputs = h*w*n*(classes + 4 + 1); l.inputs = l.outputs; l.max_boxes = max_boxes; l.truth_size = 4 + 2; l.truths = l.max_boxes*l.truth_size; // 90*(4 + 1); l.labels = (int*)xcalloc(batch * l.w*l.h*l.n, sizeof(int)); for (i = 0; i < batch * l.w*l.h*l.n; ++i) l.labels[i] = -1; l.class_ids = (int*)xcalloc(batch * l.w*l.h*l.n, sizeof(int)); for (i = 0; i < batch * l.w*l.h*l.n; ++i) l.class_ids[i] = -1; l.delta = (float*)xcalloc(batch * l.outputs, sizeof(float)); l.output = (float*)xcalloc(batch * l.outputs, sizeof(float)); for(i = 0; i < total*2; ++i){ l.biases[i] = .5; } l.forward = forward_yolo_layer; l.backward = backward_yolo_layer; #ifdef GPU l.forward_gpu = forward_yolo_layer_gpu; l.backward_gpu = backward_yolo_layer_gpu; l.output_gpu = cuda_make_array(l.output, batch*l.outputs); l.output_avg_gpu = cuda_make_array(l.output, batch*l.outputs); l.delta_gpu = cuda_make_array(l.delta, batch*l.outputs); free(l.output); if (cudaSuccess == cudaHostAlloc(&l.output, batch*l.outputs*sizeof(float), cudaHostRegisterMapped)) l.output_pinned = 1; else { cudaGetLastError(); // reset CUDA-error l.output = (float*)xcalloc(batch * l.outputs, sizeof(float)); } free(l.delta); if (cudaSuccess == cudaHostAlloc(&l.delta, batch*l.outputs*sizeof(float), cudaHostRegisterMapped)) l.delta_pinned = 1; else { cudaGetLastError(); // reset CUDA-error l.delta = (float*)xcalloc(batch * l.outputs, sizeof(float)); } #endif fprintf(stderr, "yolo\n"); srand(time(0)); return l; } void resize_yolo_layer(layer *l, int w, int h) { l->w = w; l->h = h; l->outputs = h*w*l->n*(l->classes + 4 + 1); l->inputs = l->outputs; if (l->embedding_output) l->embedding_output = (float*)xrealloc(l->output, l->batch * l->embedding_size * l->n * l->h * l->w * sizeof(float)); if (l->labels) l->labels = (int*)xrealloc(l->labels, l->batch * l->n * l->h * l->w * sizeof(int)); if (l->class_ids) l->class_ids = (int*)xrealloc(l->class_ids, l->batch * l->n * l->h * l->w * sizeof(int)); if (!l->output_pinned) l->output = (float*)xrealloc(l->output, l->batch*l->outputs * sizeof(float)); if (!l->delta_pinned) l->delta = (float*)xrealloc(l->delta, l->batch*l->outputs*sizeof(float)); #ifdef GPU if (l->output_pinned) { CHECK_CUDA(cudaFreeHost(l->output)); if (cudaSuccess != cudaHostAlloc(&l->output, l->batch*l->outputs * sizeof(float), cudaHostRegisterMapped)) { cudaGetLastError(); // reset CUDA-error l->output = (float*)xcalloc(l->batch * l->outputs, sizeof(float)); l->output_pinned = 0; } } if (l->delta_pinned) { CHECK_CUDA(cudaFreeHost(l->delta)); if (cudaSuccess != cudaHostAlloc(&l->delta, l->batch*l->outputs * sizeof(float), cudaHostRegisterMapped)) { cudaGetLastError(); // reset CUDA-error l->delta = (float*)xcalloc(l->batch * l->outputs, sizeof(float)); l->delta_pinned = 0; } } cuda_free(l->delta_gpu); cuda_free(l->output_gpu); cuda_free(l->output_avg_gpu); l->delta_gpu = cuda_make_array(l->delta, l->batch*l->outputs); l->output_gpu = cuda_make_array(l->output, l->batch*l->outputs); l->output_avg_gpu = cuda_make_array(l->output, l->batch*l->outputs); #endif } box get_yolo_box(float *x, float *biases, int n, int index, int i, int j, int lw, int lh, int w, int h, int stride, int new_coords) { box b; // ln - natural logarithm (base = e) // x` = t.x * lw - i; // x = ln(x`/(1-x`)) // x - output of previous conv-layer // y` = t.y * lh - i; // y = ln(y`/(1-y`)) // y - output of previous conv-layer // w = ln(t.w * net.w / anchors_w); // w - output of previous conv-layer // h = ln(t.h * net.h / anchors_h); // h - output of previous conv-layer if (new_coords) { b.x = (i + x[index + 0 * stride]) / lw; b.y = (j + x[index + 1 * stride]) / lh; b.w = x[index + 2 * stride] * x[index + 2 * stride] * 4 * biases[2 * n] / w; b.h = x[index + 3 * stride] * x[index + 3 * stride] * 4 * biases[2 * n + 1] / h; } else { b.x = (i + x[index + 0 * stride]) / lw; b.y = (j + x[index + 1 * stride]) / lh; b.w = exp(x[index + 2 * stride]) * biases[2 * n] / w; b.h = exp(x[index + 3 * stride]) * biases[2 * n + 1] / h; } return b; } static inline float fix_nan_inf(float val) { if (isnan(val) || isinf(val)) val = 0; return val; } static inline float clip_value(float val, const float max_val) { if (val > max_val) { //printf("\n val = %f > max_val = %f \n", val, max_val); val = max_val; } else if (val < -max_val) { //printf("\n val = %f < -max_val = %f \n", val, -max_val); val = -max_val; } return val; } ious delta_yolo_box(box truth, float *x, float *biases, int n, int index, int i, int j, int lw, int lh, int w, int h, float *delta, float scale, int stride, float iou_normalizer, IOU_LOSS iou_loss, int accumulate, float max_delta, int *rewritten_bbox, int new_coords) { if (delta[index + 0 * stride] || delta[index + 1 * stride] || delta[index + 2 * stride] || delta[index + 3 * stride]) { (*rewritten_bbox)++; } ious all_ious = { 0 }; // i - step in layer width // j - step in layer height // Returns a box in absolute coordinates box pred = get_yolo_box(x, biases, n, index, i, j, lw, lh, w, h, stride, new_coords); all_ious.iou = box_iou(pred, truth); all_ious.giou = box_giou(pred, truth); all_ious.diou = box_diou(pred, truth); all_ious.ciou = box_ciou(pred, truth); // avoid nan in dx_box_iou if (pred.w == 0) { pred.w = 1.0; } if (pred.h == 0) { pred.h = 1.0; } if (iou_loss == MSE) // old loss { float tx = (truth.x*lw - i); float ty = (truth.y*lh - j); float tw = log(truth.w*w / biases[2 * n]); float th = log(truth.h*h / biases[2 * n + 1]); if (new_coords) { //tx = (truth.x*lw - i + 0.5) / 2; //ty = (truth.y*lh - j + 0.5) / 2; tw = sqrt(truth.w*w / (4 * biases[2 * n])); th = sqrt(truth.h*h / (4 * biases[2 * n + 1])); } //printf(" tx = %f, ty = %f, tw = %f, th = %f \n", tx, ty, tw, th); //printf(" x = %f, y = %f, w = %f, h = %f \n", x[index + 0 * stride], x[index + 1 * stride], x[index + 2 * stride], x[index + 3 * stride]); // accumulate delta delta[index + 0 * stride] += scale * (tx - x[index + 0 * stride]) * iou_normalizer; delta[index + 1 * stride] += scale * (ty - x[index + 1 * stride]) * iou_normalizer; delta[index + 2 * stride] += scale * (tw - x[index + 2 * stride]) * iou_normalizer; delta[index + 3 * stride] += scale * (th - x[index + 3 * stride]) * iou_normalizer; } else { // https://github.com/generalized-iou/g-darknet // https://arxiv.org/abs/1902.09630v2 // https://giou.stanford.edu/ all_ious.dx_iou = dx_box_iou(pred, truth, iou_loss); // jacobian^t (transpose) //float dx = (all_ious.dx_iou.dl + all_ious.dx_iou.dr); //float dy = (all_ious.dx_iou.dt + all_ious.dx_iou.db); //float dw = ((-0.5 * all_ious.dx_iou.dl) + (0.5 * all_ious.dx_iou.dr)); //float dh = ((-0.5 * all_ious.dx_iou.dt) + (0.5 * all_ious.dx_iou.db)); // jacobian^t (transpose) float dx = all_ious.dx_iou.dt; float dy = all_ious.dx_iou.db; float dw = all_ious.dx_iou.dl; float dh = all_ious.dx_iou.dr; // predict exponential, apply gradient of e^delta_t ONLY for w,h if (new_coords) { //dw *= 8 * x[index + 2 * stride]; //dh *= 8 * x[index + 3 * stride]; //dw *= 8 * x[index + 2 * stride] * biases[2 * n] / w; //dh *= 8 * x[index + 3 * stride] * biases[2 * n + 1] / h; //float grad_w = 8 * exp(-x[index + 2 * stride]) / pow(exp(-x[index + 2 * stride]) + 1, 3); //float grad_h = 8 * exp(-x[index + 3 * stride]) / pow(exp(-x[index + 3 * stride]) + 1, 3); //dw *= grad_w; //dh *= grad_h; } else { dw *= exp(x[index + 2 * stride]); dh *= exp(x[index + 3 * stride]); } //dw *= exp(x[index + 2 * stride]); //dh *= exp(x[index + 3 * stride]); // normalize iou weight dx *= iou_normalizer; dy *= iou_normalizer; dw *= iou_normalizer; dh *= iou_normalizer; dx = fix_nan_inf(dx); dy = fix_nan_inf(dy); dw = fix_nan_inf(dw); dh = fix_nan_inf(dh); if (max_delta != FLT_MAX) { dx = clip_value(dx, max_delta); dy = clip_value(dy, max_delta); dw = clip_value(dw, max_delta); dh = clip_value(dh, max_delta); } if (!accumulate) { delta[index + 0 * stride] = 0; delta[index + 1 * stride] = 0; delta[index + 2 * stride] = 0; delta[index + 3 * stride] = 0; } // accumulate delta delta[index + 0 * stride] += dx; delta[index + 1 * stride] += dy; delta[index + 2 * stride] += dw; delta[index + 3 * stride] += dh; } return all_ious; } void averages_yolo_deltas(int class_index, int box_index, int stride, int classes, float *delta) { int classes_in_one_box = 0; int c; for (c = 0; c < classes; ++c) { if (delta[class_index + stride*c] > 0) classes_in_one_box++; } if (classes_in_one_box > 0) { delta[box_index + 0 * stride] /= classes_in_one_box; delta[box_index + 1 * stride] /= classes_in_one_box; delta[box_index + 2 * stride] /= classes_in_one_box; delta[box_index + 3 * stride] /= classes_in_one_box; } } void delta_yolo_class(float *output, float *delta, int index, int class_id, int classes, int stride, float *avg_cat, int focal_loss, float label_smooth_eps, float *classes_multipliers, float cls_normalizer) { int n; if (delta[index + stride*class_id]){ float y_true = 1; if(label_smooth_eps) y_true = y_true * (1 - label_smooth_eps) + 0.5*label_smooth_eps; float result_delta = y_true - output[index + stride*class_id]; if(!isnan(result_delta) && !isinf(result_delta)) delta[index + stride*class_id] = result_delta; //delta[index + stride*class_id] = 1 - output[index + stride*class_id]; if (classes_multipliers) delta[index + stride*class_id] *= classes_multipliers[class_id]; if(avg_cat) *avg_cat += output[index + stride*class_id]; return; } // Focal loss if (focal_loss) { // Focal Loss float alpha = 0.5; // 0.25 or 0.5 //float gamma = 2; // hardcoded in many places of the grad-formula int ti = index + stride*class_id; float pt = output[ti] + 0.000000000000001F; // http://fooplot.com/#W3sidHlwZSI6MCwiZXEiOiItKDEteCkqKDIqeCpsb2coeCkreC0xKSIsImNvbG9yIjoiIzAwMDAwMCJ9LHsidHlwZSI6MTAwMH1d float grad = -(1 - pt) * (2 * pt*logf(pt) + pt - 1); // http://blog.csdn.net/linmingan/article/details/77885832 //float grad = (1 - pt) * (2 * pt*logf(pt) + pt - 1); // https://github.com/unsky/focal-loss for (n = 0; n < classes; ++n) { delta[index + stride*n] = (((n == class_id) ? 1 : 0) - output[index + stride*n]); delta[index + stride*n] *= alpha*grad; if (n == class_id && avg_cat) *avg_cat += output[index + stride*n]; } } else { // default for (n = 0; n < classes; ++n) { float y_true = ((n == class_id) ? 1 : 0); if (label_smooth_eps) y_true = y_true * (1 - label_smooth_eps) + 0.5*label_smooth_eps; float result_delta = y_true - output[index + stride*n]; if (!isnan(result_delta) && !isinf(result_delta)) delta[index + stride*n] = result_delta; if (classes_multipliers && n == class_id) delta[index + stride*class_id] *= classes_multipliers[class_id] * cls_normalizer; if (n == class_id && avg_cat) *avg_cat += output[index + stride*n]; } } } int compare_yolo_class(float *output, int classes, int class_index, int stride, float objectness, int class_id, float conf_thresh) { int j; for (j = 0; j < classes; ++j) { //float prob = objectness * output[class_index + stride*j]; float prob = output[class_index + stride*j]; if (prob > conf_thresh) { return 1; } } return 0; } static int entry_index(layer l, int batch, int location, int entry) { int n = location / (l.w*l.h); int loc = location % (l.w*l.h); return batch*l.outputs + n*l.w*l.h*(4+l.classes+1) + entry*l.w*l.h + loc; } typedef struct train_yolo_args { layer l; network_state state; int b; float tot_iou; float tot_giou_loss; float tot_iou_loss; int count; int class_count; } train_yolo_args; void *process_batch(void* ptr) { { train_yolo_args *args = (train_yolo_args*)ptr; const layer l = args->l; network_state state = args->state; int b = args->b; int i, j, t, n; //printf(" b = %d \n", b, b); //float tot_iou = 0; float tot_giou = 0; float tot_diou = 0; float tot_ciou = 0; //float tot_iou_loss = 0; //float tot_giou_loss = 0; float tot_diou_loss = 0; float tot_ciou_loss = 0; float recall = 0; float recall75 = 0; float avg_cat = 0; float avg_obj = 0; float avg_anyobj = 0; //int count = 0; //int class_count = 0; for (j = 0; j < l.h; ++j) { for (i = 0; i < l.w; ++i) { for (n = 0; n < l.n; ++n) { const int class_index = entry_index(l, b, n * l.w * l.h + j * l.w + i, 4 + 1); const int obj_index = entry_index(l, b, n * l.w * l.h + j * l.w + i, 4); const int box_index = entry_index(l, b, n * l.w * l.h + j * l.w + i, 0); const int stride = l.w * l.h; box pred = get_yolo_box(l.output, l.biases, l.mask[n], box_index, i, j, l.w, l.h, state.net.w, state.net.h, l.w * l.h, l.new_coords); float best_match_iou = 0; int best_match_t = 0; float best_iou = 0; int best_t = 0; for (t = 0; t < l.max_boxes; ++t) { box truth = float_to_box_stride(state.truth + t * l.truth_size + b * l.truths, 1); if (!truth.x) break; // continue; int class_id = state.truth[t * l.truth_size + b * l.truths + 4]; if (class_id >= l.classes || class_id < 0) { 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); printf("\n truth.x = %f, truth.y = %f, truth.w = %f, truth.h = %f, class_id = %d \n", truth.x, truth.y, truth.w, truth.h, class_id); if (check_mistakes) getchar(); continue; // if label contains class_id more than number of classes in the cfg-file and class_id check garbage value } float objectness = l.output[obj_index]; if (isnan(objectness) || isinf(objectness)) l.output[obj_index] = 0; int class_id_match = compare_yolo_class(l.output, l.classes, class_index, l.w * l.h, objectness, class_id, 0.25f); float iou = box_iou(pred, truth); if (iou > best_match_iou && class_id_match == 1) { best_match_iou = iou; best_match_t = t; } if (iou > best_iou) { best_iou = iou; best_t = t; } } avg_anyobj += l.output[obj_index]; l.delta[obj_index] = l.obj_normalizer * (0 - l.output[obj_index]); if (best_match_iou > l.ignore_thresh) { if (l.objectness_smooth) { const float delta_obj = l.obj_normalizer * (best_match_iou - l.output[obj_index]); if (delta_obj > l.delta[obj_index]) l.delta[obj_index] = delta_obj; } else l.delta[obj_index] = 0; } else if (state.net.adversarial) { int stride = l.w * l.h; float scale = pred.w * pred.h; if (scale > 0) scale = sqrt(scale); l.delta[obj_index] = scale * l.obj_normalizer * (0 - l.output[obj_index]); int cl_id; int found_object = 0; for (cl_id = 0; cl_id < l.classes; ++cl_id) { if (l.output[class_index + stride * cl_id] * l.output[obj_index] > 0.25) { l.delta[class_index + stride * cl_id] = scale * (0 - l.output[class_index + stride * cl_id]); found_object = 1; } } if (found_object) { // don't use this loop for adversarial attack drawing for (cl_id = 0; cl_id < l.classes; ++cl_id) if (l.output[class_index + stride * cl_id] * l.output[obj_index] < 0.25) l.delta[class_index + stride * cl_id] = scale * (1 - l.output[class_index + stride * cl_id]); l.delta[box_index + 0 * stride] += scale * (0 - l.output[box_index + 0 * stride]); l.delta[box_index + 1 * stride] += scale * (0 - l.output[box_index + 1 * stride]); l.delta[box_index + 2 * stride] += scale * (0 - l.output[box_index + 2 * stride]); l.delta[box_index + 3 * stride] += scale * (0 - l.output[box_index + 3 * stride]); } } if (best_iou > l.truth_thresh) { const float iou_multiplier = best_iou * best_iou;// (best_iou - l.truth_thresh) / (1.0 - l.truth_thresh); if (l.objectness_smooth) l.delta[obj_index] = l.obj_normalizer * (iou_multiplier - l.output[obj_index]); else l.delta[obj_index] = l.obj_normalizer * (1 - l.output[obj_index]); //l.delta[obj_index] = l.obj_normalizer * (1 - l.output[obj_index]); int class_id = state.truth[best_t * l.truth_size + b * l.truths + 4]; if (l.map) class_id = l.map[class_id]; delta_yolo_class(l.output, l.delta, class_index, class_id, l.classes, l.w * l.h, 0, l.focal_loss, l.label_smooth_eps, l.classes_multipliers, l.cls_normalizer); const float class_multiplier = (l.classes_multipliers) ? l.classes_multipliers[class_id] : 1.0f; if (l.objectness_smooth) l.delta[class_index + stride * class_id] = class_multiplier * (iou_multiplier - l.output[class_index + stride * class_id]); box truth = float_to_box_stride(state.truth + best_t * l.truth_size + b * l.truths, 1); delta_yolo_box(truth, l.output, l.biases, l.mask[n], box_index, i, j, l.w, l.h, state.net.w, state.net.h, l.delta, (2 - truth.w * truth.h), l.w * l.h, l.iou_normalizer * class_multiplier, l.iou_loss, 1, l.max_delta, state.net.rewritten_bbox, l.new_coords); (*state.net.total_bbox)++; } } } } for (t = 0; t < l.max_boxes; ++t) { box truth = float_to_box_stride(state.truth + t * l.truth_size + b * l.truths, 1); if (!truth.x) break; // continue; if (truth.x < 0 || truth.y < 0 || truth.x > 1 || truth.y > 1 || truth.w < 0 || truth.h < 0) { char buff[256]; printf(" Wrong label: truth.x = %f, truth.y = %f, truth.w = %f, truth.h = %f \n", truth.x, truth.y, truth.w, truth.h); sprintf(buff, "echo \"Wrong label: truth.x = %f, truth.y = %f, truth.w = %f, truth.h = %f\" >> bad_label.list", truth.x, truth.y, truth.w, truth.h); system(buff); } int class_id = state.truth[t * l.truth_size + b * l.truths + 4]; if (class_id >= l.classes || class_id < 0) continue; // if label contains class_id more than number of classes in the cfg-file and class_id check garbage value float best_iou = 0; int best_n = 0; i = (truth.x * l.w); j = (truth.y * l.h); box truth_shift = truth; truth_shift.x = truth_shift.y = 0; for (n = 0; n < l.total; ++n) { box pred = { 0 }; pred.w = l.biases[2 * n] / state.net.w; pred.h = l.biases[2 * n + 1] / state.net.h; float iou = box_iou(pred, truth_shift); if (iou > best_iou) { best_iou = iou; best_n = n; } } int mask_n = int_index(l.mask, best_n, l.n); if (mask_n >= 0) { int class_id = state.truth[t * l.truth_size + b * l.truths + 4]; if (l.map) class_id = l.map[class_id]; int box_index = entry_index(l, b, mask_n * l.w * l.h + j * l.w + i, 0); const float class_multiplier = (l.classes_multipliers) ? l.classes_multipliers[class_id] : 1.0f; ious all_ious = delta_yolo_box(truth, l.output, l.biases, best_n, box_index, i, j, l.w, l.h, state.net.w, state.net.h, l.delta, (2 - truth.w * truth.h), l.w * l.h, l.iou_normalizer * class_multiplier, l.iou_loss, 1, l.max_delta, state.net.rewritten_bbox, l.new_coords); (*state.net.total_bbox)++; const int truth_in_index = t * l.truth_size + b * l.truths + 5; const int track_id = state.truth[truth_in_index]; const int truth_out_index = b * l.n * l.w * l.h + mask_n * l.w * l.h + j * l.w + i; l.labels[truth_out_index] = track_id; l.class_ids[truth_out_index] = class_id; //printf(" track_id = %d, t = %d, b = %d, truth_in_index = %d, truth_out_index = %d \n", track_id, t, b, truth_in_index, truth_out_index); // range is 0 <= 1 args->tot_iou += all_ious.iou; args->tot_iou_loss += 1 - all_ious.iou; // range is -1 <= giou <= 1 tot_giou += all_ious.giou; args->tot_giou_loss += 1 - all_ious.giou; tot_diou += all_ious.diou; tot_diou_loss += 1 - all_ious.diou; tot_ciou += all_ious.ciou; tot_ciou_loss += 1 - all_ious.ciou; int obj_index = entry_index(l, b, mask_n * l.w * l.h + j * l.w + i, 4); avg_obj += l.output[obj_index]; if (l.objectness_smooth) { float delta_obj = class_multiplier * l.obj_normalizer * (1 - l.output[obj_index]); if (l.delta[obj_index] == 0) l.delta[obj_index] = delta_obj; } else l.delta[obj_index] = class_multiplier * l.obj_normalizer * (1 - l.output[obj_index]); int class_index = entry_index(l, b, mask_n * l.w * l.h + j * l.w + i, 4 + 1); delta_yolo_class(l.output, l.delta, class_index, class_id, l.classes, l.w * l.h, &avg_cat, l.focal_loss, l.label_smooth_eps, l.classes_multipliers, l.cls_normalizer); //printf(" label: class_id = %d, truth.x = %f, truth.y = %f, truth.w = %f, truth.h = %f \n", class_id, truth.x, truth.y, truth.w, truth.h); //printf(" mask_n = %d, l.output[obj_index] = %f, l.output[class_index + class_id] = %f \n\n", mask_n, l.output[obj_index], l.output[class_index + class_id]); ++(args->count); ++(args->class_count); if (all_ious.iou > .5) recall += 1; if (all_ious.iou > .75) recall75 += 1; } // iou_thresh for (n = 0; n < l.total; ++n) { int mask_n = int_index(l.mask, n, l.n); if (mask_n >= 0 && n != best_n && l.iou_thresh < 1.0f) { box pred = { 0 }; pred.w = l.biases[2 * n] / state.net.w; pred.h = l.biases[2 * n + 1] / state.net.h; float iou = box_iou_kind(pred, truth_shift, l.iou_thresh_kind); // IOU, GIOU, MSE, DIOU, CIOU // iou, n if (iou > l.iou_thresh) { int class_id = state.truth[t * l.truth_size + b * l.truths + 4]; if (l.map) class_id = l.map[class_id]; int box_index = entry_index(l, b, mask_n * l.w * l.h + j * l.w + i, 0); const float class_multiplier = (l.classes_multipliers) ? l.classes_multipliers[class_id] : 1.0f; ious all_ious = delta_yolo_box(truth, l.output, l.biases, n, box_index, i, j, l.w, l.h, state.net.w, state.net.h, l.delta, (2 - truth.w * truth.h), l.w * l.h, l.iou_normalizer * class_multiplier, l.iou_loss, 1, l.max_delta, state.net.rewritten_bbox, l.new_coords); (*state.net.total_bbox)++; // range is 0 <= 1 args->tot_iou += all_ious.iou; args->tot_iou_loss += 1 - all_ious.iou; // range is -1 <= giou <= 1 tot_giou += all_ious.giou; args->tot_giou_loss += 1 - all_ious.giou; tot_diou += all_ious.diou; tot_diou_loss += 1 - all_ious.diou; tot_ciou += all_ious.ciou; tot_ciou_loss += 1 - all_ious.ciou; int obj_index = entry_index(l, b, mask_n * l.w * l.h + j * l.w + i, 4); avg_obj += l.output[obj_index]; if (l.objectness_smooth) { float delta_obj = class_multiplier * l.obj_normalizer * (1 - l.output[obj_index]); if (l.delta[obj_index] == 0) l.delta[obj_index] = delta_obj; } else l.delta[obj_index] = class_multiplier * l.obj_normalizer * (1 - l.output[obj_index]); int class_index = entry_index(l, b, mask_n * l.w * l.h + j * l.w + i, 4 + 1); delta_yolo_class(l.output, l.delta, class_index, class_id, l.classes, l.w * l.h, &avg_cat, l.focal_loss, l.label_smooth_eps, l.classes_multipliers, l.cls_normalizer); ++(args->count); ++(args->class_count); if (all_ious.iou > .5) recall += 1; if (all_ious.iou > .75) recall75 += 1; } } } } if (l.iou_thresh < 1.0f) { // averages the deltas obtained by the function: delta_yolo_box()_accumulate for (j = 0; j < l.h; ++j) { for (i = 0; i < l.w; ++i) { for (n = 0; n < l.n; ++n) { int obj_index = entry_index(l, b, n*l.w*l.h + j*l.w + i, 4); int box_index = entry_index(l, b, n*l.w*l.h + j*l.w + i, 0); int class_index = entry_index(l, b, n*l.w*l.h + j*l.w + i, 4 + 1); const int stride = l.w*l.h; if (l.delta[obj_index] != 0) averages_yolo_deltas(class_index, box_index, stride, l.classes, l.delta); } } } } } return 0; } void forward_yolo_layer(const layer l, network_state state) { //int i, j, b, t, n; memcpy(l.output, state.input, l.outputs*l.batch * sizeof(float)); int b, n; #ifndef GPU for (b = 0; b < l.batch; ++b) { for (n = 0; n < l.n; ++n) { int bbox_index = entry_index(l, b, n*l.w*l.h, 0); if (l.new_coords) { //activate_array(l.output + bbox_index, 4 * l.w*l.h, LOGISTIC); // x,y,w,h } else { activate_array(l.output + bbox_index, 2 * l.w*l.h, LOGISTIC); // x,y, int obj_index = entry_index(l, b, n*l.w*l.h, 4); activate_array(l.output + obj_index, (1 + l.classes)*l.w*l.h, LOGISTIC); } scal_add_cpu(2 * l.w*l.h, l.scale_x_y, -0.5*(l.scale_x_y - 1), l.output + bbox_index, 1); // scale x,y } } #endif // delta is zeroed memset(l.delta, 0, l.outputs * l.batch * sizeof(float)); if (!state.train) return; int i; for (i = 0; i < l.batch * l.w*l.h*l.n; ++i) l.labels[i] = -1; for (i = 0; i < l.batch * l.w*l.h*l.n; ++i) l.class_ids[i] = -1; //float avg_iou = 0; float tot_iou = 0; float tot_giou = 0; float tot_diou = 0; float tot_ciou = 0; float tot_iou_loss = 0; float tot_giou_loss = 0; float tot_diou_loss = 0; float tot_ciou_loss = 0; float recall = 0; float recall75 = 0; float avg_cat = 0; float avg_obj = 0; float avg_anyobj = 0; int count = 0; int class_count = 0; *(l.cost) = 0; int num_threads = l.batch; pthread_t* threads = (pthread_t*)calloc(num_threads, sizeof(pthread_t)); struct train_yolo_args* yolo_args = (train_yolo_args*)xcalloc(l.batch, sizeof(struct train_yolo_args)); for (b = 0; b < l.batch; b++) { yolo_args[b].l = l; yolo_args[b].state = state; yolo_args[b].b = b; yolo_args[b].tot_iou = 0; yolo_args[b].tot_iou_loss = 0; yolo_args[b].tot_giou_loss = 0; yolo_args[b].count = 0; yolo_args[b].class_count = 0; if (pthread_create(&threads[b], 0, process_batch, &(yolo_args[b]))) error("Thread creation failed"); } for (b = 0; b < l.batch; b++) { pthread_join(threads[b], 0); tot_iou += yolo_args[b].tot_iou; tot_iou_loss += yolo_args[b].tot_iou_loss; tot_giou_loss += yolo_args[b].tot_giou_loss; count += yolo_args[b].count; class_count += yolo_args[b].class_count; } free(yolo_args); free(threads); // Search for an equidistant point from the distant boundaries of the local minimum int iteration_num = get_current_iteration(state.net); const int start_point = state.net.max_batches * 3 / 4; //printf(" equidistant_point ep = %d, it = %d \n", state.net.equidistant_point, iteration_num); if ((state.net.badlabels_rejection_percentage && start_point < iteration_num) || (state.net.num_sigmas_reject_badlabels && start_point < iteration_num) || (state.net.equidistant_point && state.net.equidistant_point < iteration_num)) { const float progress_it = iteration_num - state.net.equidistant_point; const float progress = progress_it / (state.net.max_batches - state.net.equidistant_point); float ep_loss_threshold = (*state.net.delta_rolling_avg) * progress * 1.4; float cur_max = 0; float cur_avg = 0; float counter = 0; for (i = 0; i < l.batch * l.outputs; ++i) { if (l.delta[i] != 0) { counter++; cur_avg += fabs(l.delta[i]); if (cur_max < fabs(l.delta[i])) cur_max = fabs(l.delta[i]); } } cur_avg = cur_avg / counter; if (*state.net.delta_rolling_max == 0) *state.net.delta_rolling_max = cur_max; *state.net.delta_rolling_max = *state.net.delta_rolling_max * 0.99 + cur_max * 0.01; *state.net.delta_rolling_avg = *state.net.delta_rolling_avg * 0.99 + cur_avg * 0.01; // reject high loss to filter bad labels if (state.net.num_sigmas_reject_badlabels && start_point < iteration_num) { const float rolling_std = (*state.net.delta_rolling_std); const float rolling_max = (*state.net.delta_rolling_max); const float rolling_avg = (*state.net.delta_rolling_avg); const float progress_badlabels = (float)(iteration_num - start_point) / (start_point); float cur_std = 0; float counter = 0; for (i = 0; i < l.batch * l.outputs; ++i) { if (l.delta[i] != 0) { counter++; cur_std += pow(l.delta[i] - rolling_avg, 2); } } cur_std = sqrt(cur_std / counter); *state.net.delta_rolling_std = *state.net.delta_rolling_std * 0.99 + cur_std * 0.01; float final_badlebels_threshold = rolling_avg + rolling_std * state.net.num_sigmas_reject_badlabels; float badlabels_threshold = rolling_max - progress_badlabels * fabs(rolling_max - final_badlebels_threshold); badlabels_threshold = max_val_cmp(final_badlebels_threshold, badlabels_threshold); for (i = 0; i < l.batch * l.outputs; ++i) { if (fabs(l.delta[i]) > badlabels_threshold) l.delta[i] = 0; } printf(" rolling_std = %f, rolling_max = %f, rolling_avg = %f \n", rolling_std, rolling_max, rolling_avg); printf(" badlabels loss_threshold = %f, start_it = %d, progress = %f \n", badlabels_threshold, start_point, progress_badlabels *100); ep_loss_threshold = min_val_cmp(final_badlebels_threshold, rolling_avg) * progress; } // reject some percent of the highest deltas to filter bad labels if (state.net.badlabels_rejection_percentage && start_point < iteration_num) { if (*state.net.badlabels_reject_threshold == 0) *state.net.badlabels_reject_threshold = *state.net.delta_rolling_max; printf(" badlabels_reject_threshold = %f \n", *state.net.badlabels_reject_threshold); const float num_deltas_per_anchor = (l.classes + 4 + 1); float counter_reject = 0; float counter_all = 0; for (i = 0; i < l.batch * l.outputs; ++i) { if (l.delta[i] != 0) { counter_all++; if (fabs(l.delta[i]) > (*state.net.badlabels_reject_threshold)) { counter_reject++; l.delta[i] = 0; } } } float cur_percent = 100 * (counter_reject*num_deltas_per_anchor / counter_all); if (cur_percent > state.net.badlabels_rejection_percentage) { *state.net.badlabels_reject_threshold += 0.01; printf(" increase!!! \n"); } else if (*state.net.badlabels_reject_threshold > 0.01) { *state.net.badlabels_reject_threshold -= 0.01; printf(" decrease!!! \n"); } printf(" badlabels_reject_threshold = %f, cur_percent = %f, badlabels_rejection_percentage = %f, delta_rolling_max = %f \n", *state.net.badlabels_reject_threshold, cur_percent, state.net.badlabels_rejection_percentage, *state.net.delta_rolling_max); } // reject low loss to find equidistant point if (state.net.equidistant_point && state.net.equidistant_point < iteration_num) { printf(" equidistant_point loss_threshold = %f, start_it = %d, progress = %3.1f %% \n", ep_loss_threshold, state.net.equidistant_point, progress * 100); for (i = 0; i < l.batch * l.outputs; ++i) { if (fabs(l.delta[i]) < ep_loss_threshold) l.delta[i] = 0; } } } if (count == 0) count = 1; if (class_count == 0) class_count = 1; if (l.show_details == 0) { float loss = pow(mag_array(l.delta, l.outputs * l.batch), 2); *(l.cost) = loss; loss /= l.batch; fprintf(stderr, "v3 (%s loss, Normalizer: (iou: %.2f, obj: %.2f, cls: %.2f) Region %d Avg (IOU: %f), count: %d, total_loss = %f \n", (l.iou_loss == MSE ? "mse" : (l.iou_loss == GIOU ? "giou" : "iou")), l.iou_normalizer, l.obj_normalizer, l.cls_normalizer, state.index, tot_iou / count, count, loss); } else { // show detailed output int stride = l.w*l.h; float* no_iou_loss_delta = (float *)calloc(l.batch * l.outputs, sizeof(float)); memcpy(no_iou_loss_delta, l.delta, l.batch * l.outputs * sizeof(float)); int j, n; for (b = 0; b < l.batch; ++b) { for (j = 0; j < l.h; ++j) { for (i = 0; i < l.w; ++i) { for (n = 0; n < l.n; ++n) { int index = entry_index(l, b, n*l.w*l.h + j*l.w + i, 0); no_iou_loss_delta[index + 0 * stride] = 0; no_iou_loss_delta[index + 1 * stride] = 0; no_iou_loss_delta[index + 2 * stride] = 0; no_iou_loss_delta[index + 3 * stride] = 0; } } } } float classification_loss = l.obj_normalizer * pow(mag_array(no_iou_loss_delta, l.outputs * l.batch), 2); free(no_iou_loss_delta); float loss = pow(mag_array(l.delta, l.outputs * l.batch), 2); float iou_loss = loss - classification_loss; float avg_iou_loss = 0; *(l.cost) = loss; // gIOU loss + MSE (objectness) loss if (l.iou_loss == MSE) { *(l.cost) = pow(mag_array(l.delta, l.outputs * l.batch), 2); } else { // Always compute classification loss both for iou + cls loss and for logging with mse loss // TODO: remove IOU loss fields before computing MSE on class // probably split into two arrays if (l.iou_loss == GIOU) { avg_iou_loss = count > 0 ? l.iou_normalizer * (tot_giou_loss / count) : 0; } else { avg_iou_loss = count > 0 ? l.iou_normalizer * (tot_iou_loss / count) : 0; } *(l.cost) = avg_iou_loss + classification_loss; } loss /= l.batch; classification_loss /= l.batch; iou_loss /= l.batch; fprintf(stderr, "v3 (%s loss, Normalizer: (iou: %.2f, obj: %.2f, cls: %.2f) Region %d Avg (IOU: %f), count: %d, class_loss = %f, iou_loss = %f, total_loss = %f \n", (l.iou_loss == MSE ? "mse" : (l.iou_loss == GIOU ? "giou" : "iou")), l.iou_normalizer, l.obj_normalizer, l.cls_normalizer, state.index, tot_iou / count, count, classification_loss, iou_loss, loss); //fprintf(stderr, "v3 (%s loss, Normalizer: (iou: %.2f, cls: %.2f) Region %d Avg (IOU: %f, GIOU: %f), Class: %f, Obj: %f, No Obj: %f, .5R: %f, .75R: %f, count: %d, class_loss = %f, iou_loss = %f, total_loss = %f \n", // (l.iou_loss == MSE ? "mse" : (l.iou_loss == GIOU ? "giou" : "iou")), l.iou_normalizer, l.obj_normalizer, state.index, tot_iou / count, tot_giou / count, avg_cat / class_count, avg_obj / count, avg_anyobj / (l.w*l.h*l.n*l.batch), recall / count, recall75 / count, count, // classification_loss, iou_loss, loss); } } void backward_yolo_layer(const layer l, network_state state) { axpy_cpu(l.batch*l.inputs, 1, l.delta, 1, state.delta, 1); } // Converts output of the network to detection boxes // w,h: image width,height // netw,neth: network width,height // relative: 1 (all callers seems to pass TRUE) void correct_yolo_boxes(detection *dets, int n, int w, int h, int netw, int neth, int relative, int letter) { int i; // network height (or width) int new_w = 0; // network height (or width) int new_h = 0; // Compute scale given image w,h vs network w,h // I think this "rotates" the image to match network to input image w/h ratio // new_h and new_w are really just network width and height if (letter) { if (((float)netw / w) < ((float)neth / h)) { new_w = netw; new_h = (h * netw) / w; } else { new_h = neth; new_w = (w * neth) / h; } } else { new_w = netw; new_h = neth; } // difference between network width and "rotated" width float deltaw = netw - new_w; // difference between network height and "rotated" height float deltah = neth - new_h; // ratio between rotated network width and network width float ratiow = (float)new_w / netw; // ratio between rotated network width and network width float ratioh = (float)new_h / neth; for (i = 0; i < n; ++i) { box b = dets[i].bbox; // x = ( x - (deltaw/2)/netw ) / ratiow; // x - [(1/2 the difference of the network width and rotated width) / (network width)] b.x = (b.x - deltaw / 2. / netw) / ratiow; b.y = (b.y - deltah / 2. / neth) / ratioh; // scale to match rotation of incoming image b.w *= 1 / ratiow; b.h *= 1 / ratioh; // relative seems to always be == 1, I don't think we hit this condition, ever. if (!relative) { b.x *= w; b.w *= w; b.y *= h; b.h *= h; } dets[i].bbox = b; } } /* void correct_yolo_boxes(detection *dets, int n, int w, int h, int netw, int neth, int relative, int letter) { int i; int new_w=0; int new_h=0; if (letter) { if (((float)netw / w) < ((float)neth / h)) { new_w = netw; new_h = (h * netw) / w; } else { new_h = neth; new_w = (w * neth) / h; } } else { new_w = netw; new_h = neth; } for (i = 0; i < n; ++i){ box b = dets[i].bbox; b.x = (b.x - (netw - new_w)/2./netw) / ((float)new_w/netw); b.y = (b.y - (neth - new_h)/2./neth) / ((float)new_h/neth); b.w *= (float)netw/new_w; b.h *= (float)neth/new_h; if(!relative){ b.x *= w; b.w *= w; b.y *= h; b.h *= h; } dets[i].bbox = b; } } */ int yolo_num_detections(layer l, float thresh) { int i, n; int count = 0; for(n = 0; n < l.n; ++n){ for (i = 0; i < l.w*l.h; ++i) { int obj_index = entry_index(l, 0, n*l.w*l.h + i, 4); if(l.output[obj_index] > thresh){ ++count; } } } return count; } int yolo_num_detections_batch(layer l, float thresh, int batch) { int i, n; int count = 0; for (i = 0; i < l.w*l.h; ++i){ for(n = 0; n < l.n; ++n){ int obj_index = entry_index(l, batch, n*l.w*l.h + i, 4); if(l.output[obj_index] > thresh){ ++count; } } } return count; } void avg_flipped_yolo(layer l) { int i,j,n,z; float *flip = l.output + l.outputs; for (j = 0; j < l.h; ++j) { for (i = 0; i < l.w/2; ++i) { for (n = 0; n < l.n; ++n) { for(z = 0; z < l.classes + 4 + 1; ++z){ int i1 = z*l.w*l.h*l.n + n*l.w*l.h + j*l.w + i; int i2 = z*l.w*l.h*l.n + n*l.w*l.h + j*l.w + (l.w - i - 1); float swap = flip[i1]; flip[i1] = flip[i2]; flip[i2] = swap; if(z == 0){ flip[i1] = -flip[i1]; flip[i2] = -flip[i2]; } } } } } for(i = 0; i < l.outputs; ++i){ l.output[i] = (l.output[i] + flip[i])/2.; } } int get_yolo_detections(layer l, int w, int h, int netw, int neth, float thresh, int *map, int relative, detection *dets, int letter) { //printf("\n l.batch = %d, l.w = %d, l.h = %d, l.n = %d \n", l.batch, l.w, l.h, l.n); int i,j,n; float *predictions = l.output; // This snippet below is not necessary // Need to comment it in order to batch processing >= 2 images //if (l.batch == 2) avg_flipped_yolo(l); int count = 0; for (i = 0; i < l.w*l.h; ++i){ int row = i / l.w; int col = i % l.w; for(n = 0; n < l.n; ++n){ int obj_index = entry_index(l, 0, n*l.w*l.h + i, 4); float objectness = predictions[obj_index]; //if(objectness <= thresh) continue; // incorrect behavior for Nan values if (objectness > thresh) { //printf("\n objectness = %f, thresh = %f, i = %d, n = %d \n", objectness, thresh, i, n); int box_index = entry_index(l, 0, n*l.w*l.h + i, 0); dets[count].bbox = get_yolo_box(predictions, l.biases, l.mask[n], box_index, col, row, l.w, l.h, netw, neth, l.w*l.h, l.new_coords); dets[count].objectness = objectness; dets[count].classes = l.classes; if (l.embedding_output) { get_embedding(l.embedding_output, l.w, l.h, l.n*l.embedding_size, l.embedding_size, col, row, n, 0, dets[count].embeddings); } for (j = 0; j < l.classes; ++j) { int class_index = entry_index(l, 0, n*l.w*l.h + i, 4 + 1 + j); float prob = objectness*predictions[class_index]; dets[count].prob[j] = (prob > thresh) ? prob : 0; } ++count; } } } correct_yolo_boxes(dets, count, w, h, netw, neth, relative, letter); return count; } int get_yolo_detections_batch(layer l, int w, int h, int netw, int neth, float thresh, int *map, int relative, detection *dets, int letter, int batch) { int i,j,n; float *predictions = l.output; //if (l.batch == 2) avg_flipped_yolo(l); int count = 0; for (i = 0; i < l.w*l.h; ++i){ int row = i / l.w; int col = i % l.w; for(n = 0; n < l.n; ++n){ int obj_index = entry_index(l, batch, n*l.w*l.h + i, 4); float objectness = predictions[obj_index]; //if(objectness <= thresh) continue; // incorrect behavior for Nan values if (objectness > thresh) { //printf("\n objectness = %f, thresh = %f, i = %d, n = %d \n", objectness, thresh, i, n); int box_index = entry_index(l, batch, n*l.w*l.h + i, 0); dets[count].bbox = get_yolo_box(predictions, l.biases, l.mask[n], box_index, col, row, l.w, l.h, netw, neth, l.w*l.h, l.new_coords); dets[count].objectness = objectness; dets[count].classes = l.classes; if (l.embedding_output) { get_embedding(l.embedding_output, l.w, l.h, l.n*l.embedding_size, l.embedding_size, col, row, n, batch, dets[count].embeddings); } for (j = 0; j < l.classes; ++j) { int class_index = entry_index(l, batch, n*l.w*l.h + i, 4 + 1 + j); float prob = objectness*predictions[class_index]; dets[count].prob[j] = (prob > thresh) ? prob : 0; } ++count; } } } correct_yolo_boxes(dets, count, w, h, netw, neth, relative, letter); return count; } #ifdef GPU void forward_yolo_layer_gpu(const layer l, network_state state) { if (l.embedding_output) { layer le = state.net.layers[l.embedding_layer_id]; cuda_pull_array_async(le.output_gpu, l.embedding_output, le.batch*le.outputs); } //copy_ongpu(l.batch*l.inputs, state.input, 1, l.output_gpu, 1); simple_copy_ongpu(l.batch*l.inputs, state.input, l.output_gpu); int b, n; for (b = 0; b < l.batch; ++b){ for(n = 0; n < l.n; ++n){ int bbox_index = entry_index(l, b, n*l.w*l.h, 0); // y = 1./(1. + exp(-x)) // x = ln(y/(1-y)) // ln - natural logarithm (base = e) // if(y->1) x -> inf // if(y->0) x -> -inf if (l.new_coords) { //activate_array_ongpu(l.output_gpu + bbox_index, 4 * l.w*l.h, LOGISTIC); // x,y,w,h } else { activate_array_ongpu(l.output_gpu + bbox_index, 2 * l.w*l.h, LOGISTIC); // x,y int obj_index = entry_index(l, b, n*l.w*l.h, 4); activate_array_ongpu(l.output_gpu + obj_index, (1 + l.classes)*l.w*l.h, LOGISTIC); // classes and objectness } if (l.scale_x_y != 1) scal_add_ongpu(2 * l.w*l.h, l.scale_x_y, -0.5*(l.scale_x_y - 1), l.output_gpu + bbox_index, 1); // scale x,y } } if(!state.train || l.onlyforward){ //cuda_pull_array(l.output_gpu, l.output, l.batch*l.outputs); if (l.mean_alpha && l.output_avg_gpu) mean_array_gpu(l.output_gpu, l.batch*l.outputs, l.mean_alpha, l.output_avg_gpu); cuda_pull_array_async(l.output_gpu, l.output, l.batch*l.outputs); CHECK_CUDA(cudaPeekAtLastError()); return; } float *in_cpu = (float *)xcalloc(l.batch*l.inputs, sizeof(float)); cuda_pull_array(l.output_gpu, l.output, l.batch*l.outputs); memcpy(in_cpu, l.output, l.batch*l.outputs*sizeof(float)); float *truth_cpu = 0; if (state.truth) { int num_truth = l.batch*l.truths; truth_cpu = (float *)xcalloc(num_truth, sizeof(float)); cuda_pull_array(state.truth, truth_cpu, num_truth); } network_state cpu_state = state; cpu_state.net = state.net; cpu_state.index = state.index; cpu_state.train = state.train; cpu_state.truth = truth_cpu; cpu_state.input = in_cpu; forward_yolo_layer(l, cpu_state); //forward_yolo_layer(l, state); cuda_push_array(l.delta_gpu, l.delta, l.batch*l.outputs); free(in_cpu); if (cpu_state.truth) free(cpu_state.truth); } void backward_yolo_layer_gpu(const layer l, network_state state) { axpy_ongpu(l.batch*l.inputs, state.net.loss_scale * l.delta_normalizer, l.delta_gpu, 1, state.delta, 1); } #endif
