The torch using summary

使用tensorflowI didn't think about it at the time,Read the code today and found outtorchThere is such an important statement,记录学习一下：

1. model.train()
   启用 BatchNormalization 和 Dropout
2. model.eval()
   不启用 BatchNormalization 和 Dropout
   训练完 train 样本后,生成的模型 model 要用来测试样本.在 model(test) 之前,需要加上model.eval(),否则只要有输入数据,即使不训练,model will also change the weight.这是model中含有的 batch normalization 层所带来的的性质.
   转载自：https://zhuanlan.zhihu.com/p/208233193