Open cxq1 opened 2 years ago
artest08
commented
2 years ago I am also following these kind of errors in detr architectures in recent weeks. I do not come across in this repo but one of my architectural modification to dab-deformable detr leads to the same error.
Are you directly trying to replicate the result of the repo or did you make some modifications. If there are some modifications, what are they?
facebookresearch/detr#101
FengLi-ust
commented
2 years ago Hey, this means your modification to the code fails to make the model work and produces illegal box predictions. For example, here the mistake means the value of the bottom left point coordinates of the box is larger than the top right point.
cxq1
commented
2 years ago I am also following these kind of errors in detr architectures in recent weeks. I do not come across in this repo but one of my architectural modification to dab-deformable detr leads to the same error.
Are you directly trying to replicate the result of the repo or did you make some modifications. If there are some modifications, what are they?
It is strange that this problem occurs in the middle of the training, and the results of the previous training are ok
cxq1
commented
2 years ago 嘿，这意味着您对代码的修改无法使模型工作并产生非法的框预测。例如，这里的错误表示框的左下点坐标值大于右上点。
It is strange that this problem occurs in the middle of the training, and the results of the previous training are ok
Traceback (most recent call last): File "main.py", line 427, in
main(args)
File "main.py", line 355, in main
args.clip_max_norm, wo_class_error=wo_class_error, lr_scheduler=lr_scheduler, args=args, logger=(logger if args.save_log else None))
File "/home/cxq/dp_work/objectdetection/DN-DETR/engine.py", line 50, in train_one_epoch
loss_dict = criterion(outputs, targets, mask_dict)
File "/home/cxq/.conda/envs/torch10/lib/python3.7/site-packages/torch/nn/modules/module.py", line 1102, in _call_impl
return forward_call(*input, kwargs)
File "/home/cxq/dp_work/objectdetection/DN-DETR/models/DN_DAB_DETR/DABDETR.py", line 371, in forward
indices = self.matcher(outputs_without_aux, targets)
File "/home/cxq/.conda/envs/torch10/lib/python3.7/site-packages/torch/nn/modules/module.py", line 1102, in _call_impl
return forward_call(*input, *kwargs)
File "/home/cxq/.conda/envs/torch10/lib/python3.7/site-packages/torch/autograd/grad_mode.py", line 28, in decorate_context
return func(args, kwargs)
File "/home/cxq/dp_work/objectdetection/DN-DETR/models/DN_DAB_DETR/matcher.py", line 83, in forward
cost_giou = -generalized_box_iou(box_cxcywh_to_xyxy(out_bbox), box_cxcywh_to_xyxy(tgt_bbox))
File "/home/cxq/dp_work/objectdetection/DN-DETR/util/box_ops.py", line 52, in generalized_box_iou
assert (boxes1[:, 2:] >= boxes1[:, :2]).all()
I'm not sure why I reported this error during training, I've trained Epoch: [38] [12440/14785] eta: 0:17:40