HongChuang/ultralytics-ascend
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions 3

ultralytics 8.0.55 unified YOLOv8 model YAMLs (#1475)

Browse source
This commit is contained in:
Glenn Jocher 2023-03-20 13:54:20 +01:00 • committed by GitHub
parent 701fba4770
commit 25cc07401f
45 changed files with 203 additions and 896 deletions
Download patch file Download diff file Expand all files Collapse all files

16
ultralytics/yolo/v8/segment/train.py
View file

@ -122,13 +122,15 @@ class SegLoss(Loss):
xyxyn = target_bboxes[i][fg_mask[i]] / imgsz[[1, 0, 1, 0]]
marea = xyxy2xywh(xyxyn)[:, 2:].prod(1)
mxyxy = xyxyn * torch.tensor([mask_w, mask_h, mask_w, mask_h], device=self.device)
loss[1] += self.single_mask_loss(gt_mask, pred_masks[i][fg_mask[i]], proto[i], mxyxy,
marea) # seg loss
# WARNING: Uncomment lines below in case of Multi-GPU DDP unused gradient errors
# else:
# loss[1] += proto.sum() * 0 + pred_masks.sum() * 0
# else:
# loss[1] += proto.sum() * 0 + pred_masks.sum() * 0
loss[1] += self.single_mask_loss(gt_mask, pred_masks[i][fg_mask[i]], proto[i], mxyxy, marea) # seg
# WARNING: lines below prevents Multi-GPU DDP 'unused gradient' PyTorch errors, do not remove
else:
loss[1] += proto.sum() * 0 + pred_masks.sum() * 0
# WARNING: lines below prevent Multi-GPU DDP 'unused gradient' PyTorch errors, do not remove
else:
loss[1] += proto.sum() * 0 + pred_masks.sum() * 0
loss[0] *= self.hyp.box # box gain
loss[1] *= self.hyp.box / batch_size # seg gain