RuntimeError: 1only batches of spatial targets supported (3D tensors) but got targets of size: : [1, 1, 187, 621]

[haixiongli]

hello, I use the new version to train with multi classes dataset of KITTI semantics. but I meet the error like this
`
INFO: Using device cuda
INFO: Network:
3 input channels
20 output channels (classes)
Bilinear upscaling
INFO: Creating dataset with 200 examples
INFO: Starting training:
Epochs: 5
Batch size: 1
Learning rate: 0.0001
Training size: 180
Validation size: 20
Checkpoints: True
Device: cuda
Images scaling: 0.5

Epoch 1/5: 0%| | 0/180 [00:00<?, ?img/s]
Traceback (most recent call last):
File "/home/lab2/work/lhx/Unet/train.py", line 179, in
val_percent=args.val / 100)
File "/home/lab2/work/lhx/Unet/train.py", line 78, in train_net
loss = criterion(masks_pred, true_masks)
File "/home/lab2/anaconda3/envs/torch1.6/lib/python3.6/site-packages/torch/nn/modules/module.py", line 722, in _call_impl
result = self.forward(*input, **kwargs)
File "/home/lab2/anaconda3/envs/torch1.6/lib/python3.6/site-packages/torch/nn/modules/loss.py", line 948, in forward
ignore_index=self.ignore_index, reduction=self.reduction)
File "/home/lab2/anaconda3/envs/torch1.6/lib/python3.6/site-packages/torch/nn/functional.py", line 2422, in cross_entropy
return nll_loss(log_softmax(input, 1), target, weight, None, ignore_index, None, reduction)
File "/home/lab2/anaconda3/envs/torch1.6/lib/python3.6/site-packages/torch/nn/functional.py", line 2220, in nll_loss
ret = torch._C._nn.nll_loss2d(input, target, weight, _Reduction.get_enum(reduction), ignore_index)
RuntimeError: 1only batches of spatial targets supported (3D tensors) but got targets of size: : [1, 1, 187, 621]

`

[milesial]

Duplicate of #123