Traceback (most recent call last):
  File "train_ammeter_twoclass.py", line 189, in <module>
    train(epoch)
  File "train_ammeter_twoclass.py", line 133, in train
    outputs = net(inputs)
  File "/home/iot/miniconda2/envs/pytorch3/lib/python3.6/site-packages/torch/nn/modules/module.py", line 477,
 
 in __call__
    result = self.forward(*input, **kwargs)
  File "/home/iot/chenjun/1_program/classifer/src/model.py", line 79, in forward
    x = self.net(x)             # inception會返回兩個矩陣
  File "/home/iot/miniconda2/envs/pytorch3/lib/python3.6/site-packages/torch/nn/modules/module.py", line 477, in __call__
    result =
 
 self.forward(*input, **kwargs)
  File "/home/iot/miniconda2/envs/pytorch3/lib/python3.6/site-packages/torchvision/models/inception.py", line 109, in forward
    aux = self.AuxLogits(x)
  File "/home/iot/miniconda2/envs/pytorch3/lib/python3.6/site-packages/torch/nn/modules/module.py", line 477, in __call__
    result = self.
 
forward(*input, **kwargs)
  File "/home/iot/miniconda2/envs/pytorch3/lib/python3.6/site-packages/torchvision/models/inception.py", line 308, in forward
    x = self.conv1(x)
  File "/home/iot/miniconda2/envs/pytorch3/lib/python3.6/site-packages/torch/nn/modules/module.py", line 477, in __call__
    result = self.forward(*input, **kwargs)
  File "/home/iot/miniconda2/envs/pytorch3/lib/python3.6/site-packages/torchvision/models/inception.py", line 326, in forward
    x = self.bn(x)
  File "/home/iot/miniconda2/envs/pytorch3/lib/python3.6/site-packages/torch/nn/modules/module.py", line 477, in __call__
    result = self.forward(*input, **kwargs)
  File "/home/iot/miniconda2/envs/pytorch3/lib/python3.6/site-packages/torch/nn/modules/batchnorm.py", line 66, in forward
    exponential_average_factor, self.eps)
  File "/home/iot/miniconda2/envs/pytorch3/lib/python3.6/site-packages/torch/nn/functional.py", line 1251, in batch_norm
    raise ValueError('Expected more than 1 value per channel when training, got input size {}'.format(size))
ValueError: Expected more than 1 value per channel when training, got input size [1, 768, 1, 1]

問題分析： 模型中用了batchnomolization，訓練中用batch訓練的時候，應該是有單數，比如dataset的總樣本數為17,你的batch_size為8,就會報這樣的錯誤。
解決方案： 從dataset中刪掉一個sample。