对AlexNet深度卷积神经网络的学习随记
自己用笔记本跑的Alexnet，笔记本配置：
只可惜自己的本子不支持gpu，跑的结果太吓人了。
trainning on cpu epoch:1, loss:0.0122, train_acc:0.417, test_acc:0.711, time:11123.8 sec--------3.08h epoch:2, loss:0.0056, train_acc:0.734, test_acc:0.772, time:10350.6 sec--------2.875h epoch:3, loss:0.0045, train_acc:0.782, test_acc:0.815, time:12629.7 sec--------3.508h epoch:4, loss:0.0040, train_acc:0.810, test_acc:0.828, time:32446.9 sec--------9.012h epoch:5, loss:0.0036, train_acc:0.831, test_acc:0.836, time:15733.0 sec--------4.370h
其中epoch4是晚上跑时，本子休眠了，早上起来接着跑的，所以用了9个小时，但其它几个epoch是正常情况下跑的，尤其是epoch5,尽然跑了4个半小时。太恐怖了。
没有对比，就没有伤害，看下在基于gpu跑的结果吧：
training on gpu(0) epoch： 1, loss：1.3018, train_acc: 0.511, test_acc: 0.745, time: 19.0 sec epoch： 2, loss：0.6527, train_acc: 0.756, test_acc: 0.809, time: 17.1 sec epoch： 3, loss：0.5350, train_acc: 0.800, test_acc: 0.837, time: 17.0 sec epoch： 4, loss：0.4712, train_acc: 0.826, test_acc: 0.855, time: 17.1 sec epoch： 5, loss：0.4267, train_acc: 0.844, test_acc: 0.866, time: 17.1 sec
本来还想调参多做几次实验呢，结果。。。。
代码及过程如下：
Conv2d output.shape: torch.Size([1, 96, 54, 54]) ReLU output.shape: torch.Size([1, 96, 54, 54]) MaxPool2d output.shape: torch.Size([1, 96, 26, 26]) Conv2d output.shape: torch.Size([1, 256, 26, 26]) ReLU output.shape: torch.Size([1, 256, 26, 26]) MaxPool2d output.shape: torch.Size([1, 256, 12, 12]) Conv2d output.shape: torch.Size([1, 384, 12, 12]) ReLU output.shape: torch.Size([1, 384, 12, 12]) Conv2d output.shape: torch.Size([1, 384, 12, 12]) ReLU output.shape: torch.Size([1, 384, 12, 12]) Conv2d output.shape: torch.Size([1, 256, 12, 12]) ReLU output.shape: torch.Size([1, 256, 12, 12]) MaxPool2d output.shape: torch.Size([1, 256, 5, 5]) Flatten output.shape: torch.Size([1, 6400]) Dropout output.shape: torch.Size([1, 6400]) Linear output.shape: torch.Size([1, 4096]) ReLU output.shape: torch.Size([1, 4096]) Dropout2d output.shape: torch.Size([1, 4096]) Linear output.shape: torch.Size([1, 4096]) ReLU output.shape: torch.Size([1, 4096]) Linear output.shape: torch.Size([1, 10])
