Excuse me, can this model be used to run Chinese data sets
lalala1020 opened this issue · 3 comments
I converted the dataset to Chinese dataset but there was an error.
Traceback (most recent call last):
File "E:/PycharmProjects/LSTM_LSTM_Bias/train.py", line 276, in
train()
File "E:/PycharmProjects/LSTM_LSTM_Bias/train.py", line 168, in train
loss = criterion(output, targets)
File "D:\Program Files (x86)\anaconda3\envs\bias\lib\site-packages\torch\nn\modules\module.py", line 889, in _call_impl
result = self.forward(*input, **kwargs)
File "D:\Program Files (x86)\anaconda3\envs\bias\lib\site-packages\torch\nn\modules\loss.py", line 216, in forward
return F.nll_loss(input, target, weight=self.weight, ignore_index=self.ignore_index, reduction=self.reduction)
File "D:\Program Files (x86)\anaconda3\envs\bias\lib\site-packages\torch\nn\functional.py", line 2385, in nll_loss
ret = torch._C._nn.nll_loss(input, target, weight, _Reduction.get_enum(reduction), ignore_index)
RuntimeError: weight tensor should be defined either for all 193 classes or no classes but got weight tensor of shape: [121] at C:/w/b/windows/pytorch/aten/src\THCUNN/generic/ClassNLLCriterion.cu:44
I want to add one thing.
that the number 193 in the number of tags in the model's built-in data, the number 121 is the number of tags after I changed the dataset. Finally, I hope you can help me solve this problem.
Sorry，this repository is no longer maintained.
I really want to help you, but now I am busy with other research and I haven’t updated this repository for several years, and I almost forgot its contents.
Sorry，this repository is no longer maintained. I really want to help you, but now I am busy with other research and I haven’t updated this repository for several years, and I almost forgot its contents.
OK, thank you