1.pytorch报错：ValueError: Expected more than 1 value per channel when training, got input size torch.Size([1,512,1,1])
![pytorch报错：ValueError: Expected more than 1 value per channel when training, got input size torch.Size([1,512,1,1])](https://image.shishitao.com:8440/aHR0cHM6Ly9iYnNtYXguaWthZmFuLmNvbS9zdGF0aWMvTDNCeWIzaDVMMmgwZEhCekwybHRaekl3TVRndVkyNWliRzluY3k1amIyMHZhUzFpWlhSaEx6RTJOVFl5TWpndk1qQXhPVEV4THpFMk5UWXlNamd0TWpBeE9URXhNVFF4TkRFNU5ESTJNekF0TWpNeU56YzVOVE0wTG5CdVp3PT0uanBn.jpg?w=700 "pytorch报错：ValueError: Expected more than 1 value per channel when training, got input size torch.Size([1,512,1,1])")
2.网上查找的原因为模型中用了batchnomolization，训练中用batch训练的时候当前batch恰好只含一个sample，而由于BatchNorm操作需要多于一个数据计算平均值，因此造成该错误。
3.解决方法：在torch.utils.data.DataLoader类中或自己创建的继承于DataLoader的类中设置参数drop_last=True，把不够一个batch_size的数据丢弃。
成功解决。