yolov5 regress updates to yolov3 - build_targets()

utils/utils.py

@@ -408,49 +408,57 @@ def compute_loss(p, targets, model):  # predictions, targets, model
 
 def build_targets(p, targets, model):
     # Build targets for compute_loss(), input targets(image,class,x,y,w,h)
+
     nt = targets.shape[0]
     tcls, tbox, indices, anch = [], [], [], []
-    reject, use_all_anchors = True, True
     gain = torch.ones(6, device=targets.device)  # normalized to gridspace gain
+    off = torch.tensor([[1, 0], [0, 1], [-1, 0], [0, -1]], device=targets.device).float()  # overlap offsets
 
-
+    style = None
     multi_gpu = type(model) in (nn.parallel.DataParallel, nn.parallel.DistributedDataParallel)
     for i, j in enumerate(model.yolo_layers):
         # get number of grid points and anchor vec for this yolo layer
         anchors = model.module.module_list[j].anchor_vec if multi_gpu else model.module_list[j].anchor_vec
-
-        # iou of targets-anchors
         gain[2:] = torch.tensor(p[i].shape)[[3, 2, 3, 2]]  # xyxy gain
-        t, a = targets * gain, []
-        gwh = t[:, 4:6]
-        if nt:
-            iou = wh_iou(anchors, gwh)  # iou(3,n) = wh_iou(anchors(3,2), gwh(n,2))
-
-            if use_all_anchors:
         na = anchors.shape[0]  # number of anchors
-                a = torch.arange(na).view(-1, 1).repeat(1, nt).view(-1)
+        at = torch.arange(na).view(na, 1).repeat(1, nt)  # anchor tensor, same as .repeat_interleave(nt)
-                t = t.repeat(na, 1)
-            else:  # use best anchor only
-                iou, a = iou.max(0)  # best iou and anchor
 
-            # reject anchors below iou_thres (OPTIONAL, increases P, lowers R)
+        # Match targets to anchors
-            if reject:
+        a, t, offsets = [], targets * gain, 0
-                j = iou.view(-1) > model.hyp['iou_t']  # iou threshold hyperparameter
+        if nt:
-                t, a = t[j], a[j]
+            # r = t[None, :, 4:6] / anchors[:, None]  # wh ratio
+            # j = torch.max(r, 1. / r).max(2)[0] < model.hyp['anchor_t']  # compare
+            j = wh_iou(anchors, t[:, 4:6]) > model.hyp['iou_t']  # iou(3,n) = wh_iou(anchors(3,2), gwh(n,2))
+            a, t = at[j], t.repeat(na, 1, 1)[j]  # filter
 
-        # Indices
+            # overlaps
-        b, c = t[:, :2].long().t()  # image, class
+            gxy = t[:, 2:4]  # grid xy
+            z = torch.zeros_like(gxy)
+            if style == 'rect2':
+                g = 0.2  # offset
+                j, k = ((gxy % 1. < g) & (gxy > 1.)).T
+                a, t = torch.cat((a, a[j], a[k]), 0), torch.cat((t, t[j], t[k]), 0)
+                offsets = torch.cat((z, z[j] + off[0], z[k] + off[1]), 0) * g
+
+            elif style == 'rect4':
+                g = 0.5  # offset
+                j, k = ((gxy % 1. < g) & (gxy > 1.)).T
+                l, m = ((gxy % 1. > (1 - g)) & (gxy < (gain[[2, 3]] - 1.))).T
+                a, t = torch.cat((a, a[j], a[k], a[l], a[m]), 0), torch.cat((t, t[j], t[k], t[l], t[m]), 0)
+                offsets = torch.cat((z, z[j] + off[0], z[k] + off[1], z[l] + off[2], z[m] + off[3]), 0) * g
+
+        # Define
+        b, c = t[:, :2].long().T  # image, class
         gxy = t[:, 2:4]  # grid xy
         gwh = t[:, 4:6]  # grid wh
-        gi, gj = gxy.long().t()  # grid xy indices
+        gij = (gxy - offsets).long()
-        indices.append((b, a, gj, gi))
+        gi, gj = gij.T  # grid xy indices
 
-        # Box
+        # Append
-        tbox.append(torch.cat((gxy % 1., gwh), 1))  # xywh (grids)
+        indices.append((b, a, gj, gi))  # image, anchor, grid indices
-        anch.append(anchors[a])  # anchor vec
+        tbox.append(torch.cat((gxy - gij, gwh), 1))  # box
-
+        anch.append(anchors[a])  # anchors
-        # Class
+        tcls.append(c)  # class
-        tcls.append(c)
         if c.shape[0]:  # if any targets
             assert c.max() < model.nc, 'Model accepts %g classes labeled from 0-%g, however you labelled a class %g. ' \
                                        'See https://github.com/ultralytics/yolov3/wiki/Train-Custom-Data' % (