Expected object of scalar type Long but got scalar type Float for argument #3 'index'

[yunxingluoyun]

Encountered an error problem：

File "train_R50.py", line 285, in <module>
    main()
  File "train_R50.py", line 147, in main
    main_loss = criterion(outputs, labels)
  File "/data/wdh/.conda/envs/AI_studywdh/lib/python3.6/site-packages/torch/nn/modules/module.py", line 547, in __call__
    result = self.forward(*input, **kwargs)
  File "/data/wdh/MSCG-Net-master/MSCG-Net-master/lib/loss/acw_loss.py", line 23, in forward
    one_hot_label, mask = self.encode_one_hot_label(pred, target)
  File "/data/wdh/MSCG-Net-master/MSCG-Net-master/lib/loss/acw_loss.py", line 70, in encode_one_hot_label
    one_hot_label.scatter_(0, target.unsqueeze(1),1)
RuntimeError: Expected object of scalar type Long but got scalar type Float for argument #3 'index'

def encode_one_hot_label(self, pred, target):
    print(pred.size(),target.size())
    one_hot_label = pred.detach() * 0
    # print(one_hot_label)
    if self.ignore_index is not None:
        mask = (target == self.ignore_index)
        target = target.clone()
        target[mask] = 0
        # print(tptarget.unsqueeze(1))
        one_hot_label.scatter_(0, target.unsqueeze(1),1) #error
        mask = mask.unsqueeze(1).expand_as(one_hot_label)
        one_hot_label[mask] = 0
        return one_hot_label, mask
    else:
        one_hot_label.scatter_(1, target.unsqueeze(1), 1)
        return one_hot_label, None

pred：torch.Size([10, 7, 512, 512])

target：torch.Size([10, 512, 512])

This problem has bothered me for several days. At first I thought it was a label problem, so I changed the rgb label to a single-channel label. But still this error still exists. May I ask what causes this?