L1_loss error while training SMOKE model (kitti dataset)

[Einhartd]

Dist: Pop OS 22.04
I have encountered error in l1_loss.py while training SMOKE.

pred.size(): torch.Size([0, 2])
target.size(): torch.Size([0, 2])
target.numel(): 0
Traceback (most recent call last):
  File "/home/einhart/MonoDetector/tools/train.py", line 63, in <module>
    main()
  File "/home/einhart/MonoDetector/tools/train.py", line 58, in main
    engine.train()
  File "/home/einhart/MonoDetector/tools/../engine/base_engine.py", line 115, in train
    avg_loss, elapsed_time = self.train_one_epoch()
  File "/home/einhart/MonoDetector/tools/../utils/decorators.py", line 7, in wrapper
    result = function(*args, **kwargs)
  File "/home/einhart/MonoDetector/tools/../engine/mono_engine.py", line 85, in train_one_epoch
    _, loss_dict = self.model(data_dict)
  File "/home/einhart/anaconda3/envs/MonoDetector/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl
    return self._call_impl(*args, **kwargs)
  File "/home/einhart/anaconda3/envs/MonoDetector/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl
    return forward_call(*args, **kwargs)
  File "/home/einhart/MonoDetector/tools/../model/detector/base_mono_detector.py", line 66, in forward
    pred_dict, loss_dict = self.head.forward_train(feat, data_dict)
  File "/home/einhart/MonoDetector/tools/../model/dense_heads/base_dense_heads.py", line 106, in forward_train
    loss_dict = self._get_losses(pred_dict, target_dict)
  File "/home/einhart/MonoDetector/tools/../model/dense_heads/smoke_heads.py", line 206, in _get_losses
    offset_loss = self.crit_offset(offset_pred, offset_target)
  File "/home/einhart/anaconda3/envs/MonoDetector/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl
    return self._call_impl(*args, **kwargs)
  File "/home/einhart/anaconda3/envs/MonoDetector/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl
    return forward_call(*args, **kwargs)
  File "/home/einhart/MonoDetector/tools/../losses/l1_loss.py", line 43, in forward
    return self.loss_weight * l1_loss(pred, target, weight, reduction=reduction, avg_factor=avg_factor)
  File "/home/einhart/MonoDetector/tools/../losses/utils.py", line 46, in wrapper
    loss = loss_func(pred, target, **kwargs)
  File "/home/einhart/MonoDetector/tools/../losses/l1_loss.py", line 19, in l1_loss
    assert pred.size() == target.size() and target.numel() > 0
AssertionError

I have added few lines in l1_loss to print pred.size(), target.size() and target.numel():

if pred.size() != target.size() or target.numel() <= 0:
print('pred.size():',pred.size())
print('target.size():',target.size())
print('target.numel():',target.numel())

Do you have any idea how could I fix this problem?