updates

utils/utils.py

@@ -334,10 +334,11 @@ def compute_loss(p, targets, model):  # predictions, targets, model
     tcls, tbox, indices, anchor_vec = build_targets(model, targets)
     h = model.hyp  # hyperparameters
     arc = model.arc  # # (default, uCE, uBCE) detection architectures
+    red = 'mean'  # Loss reduction (sum or mean)
 
     # Define criteria
-    BCEcls = nn.BCEWithLogitsLoss(pos_weight=ft([h['cls_pw']]))
-    BCEobj = nn.BCEWithLogitsLoss(pos_weight=ft([h['obj_pw']]))
+    BCEcls = nn.BCEWithLogitsLoss(pos_weight=ft([h['cls_pw']]), reduction=red)
+    BCEobj = nn.BCEWithLogitsLoss(pos_weight=ft([h['obj_pw']]), reduction=red)
     BCE = nn.BCEWithLogitsLoss()
     CE = nn.CrossEntropyLoss()  # weight=model.class_weights
 
@@ -346,13 +347,16 @@ def compute_loss(p, targets, model):  # predictions, targets, model
         BCEcls, BCEobj, BCE, CE = FocalLoss(BCEcls, g), FocalLoss(BCEobj, g), FocalLoss(BCE, g), FocalLoss(CE, g)
 
     # Compute losses
+    np, ng = 0, 0  # number grid points, targets
     for i, pi in enumerate(p):  # layer index, layer predictions
         b, a, gj, gi = indices[i]  # image, anchor, gridy, gridx
         tobj = torch.zeros_like(pi[..., 0])  # target obj
+        np += tobj.numel()
 
         # Compute losses
         nb = len(b)
         if nb:  # number of targets
+            ng += nb
             ps = pi[b, a, gj, gi]  # prediction subset corresponding to targets
             tobj[b, a, gj, gi] = 1.0  # obj
             # ps[:, 2:4] = torch.sigmoid(ps[:, 2:4])  # wh power loss (uncomment)
@@ -360,8 +364,8 @@ def compute_loss(p, targets, model):  # predictions, targets, model
             # GIoU
             pxy = torch.sigmoid(ps[:, 0:2])  # pxy = pxy * s - (s - 1) / 2,  s = 1.5  (scale_xy)
             pbox = torch.cat((pxy, torch.exp(ps[:, 2:4]).clamp(max=1E3) * anchor_vec[i]), 1)  # predicted box
-            giou = bbox_iou(pbox.t(), tbox[i], x1y1x2y2=False, GIoU=True)  # giou computation
-            lbox += (1.0 - giou).mean()  # giou loss
+            giou = 1.0 - bbox_iou(pbox.t(), tbox[i], x1y1x2y2=False, GIoU=True)  # giou computation
+            lbox += giou.sum() if red == 'sum' else giou.mean()  # giou loss
 
             if 'default' in arc and model.nc > 1:  # cls loss (only if multiple classes)
                 t = torch.zeros_like(ps[:, 5:])  # targets
@@ -396,6 +400,11 @@ def compute_loss(p, targets, model):  # predictions, targets, model
     lbox *= h['giou']
     lobj *= h['obj']
     lcls *= h['cls']
+    if red == 'sum':
+        lbox *= 3 / ng
+        lobj *= 3 / np
+        lcls *= 3 / ng / model.nc
+
     loss = lbox + lobj + lcls
     return loss, torch.cat((lbox, lobj, lcls, loss)).detach()