莫烦pytorch（15）——过拟合

１．构造数据集

import matplotlib.pyplot as plt

N_SAMPLES = 20
N_HIDDEN = 300

x=torch.unsqueeze(torch.linspace(-1,1,N_SAMPLES),1)
y=x+0.3*torch.normal(torch.zeros(N_SAMPLES,1),torch.ones(N_SAMPLES,1))
# test data
test_x = torch.unsqueeze(torch.linspace(-1, 1, N_SAMPLES), 1)
test_y = test_x + 0.3*torch.normal(torch.zeros(N_SAMPLES, 1), torch.ones(N_SAMPLES, 1))

plt.scatter(x.data.numpy(), y.data.numpy(), c='magenta', s=50, alpha=0.5, label='train')
plt.scatter(test_x.data.numpy(), test_y.data.numpy(), c='cyan', s=50, alpha=0.5, label='test')
plt.legend(loc='upper left')
plt.ylim((-2.5, 2.5))
plt.show()

２.构造dropout和一般的两个网络进行对比

net_overfitting = torch.nn.Sequential(
    torch.nn.Linear(1,N_HIDDEN),
    torch.nn.ReLU(),
    torch.nn.Linear(N_HIDDEN, N_HIDDEN),
    torch.nn.ReLU(),
    torch.nn.Linear(N_HIDDEN, 1),
)

net_dropped=torch.nn.Sequential(
    torch.nn.Linear(1,N_HIDDEN),
    torch.nn.Dropout(0.5),
    torch.nn.ReLU(),
    torch.nn.Linear(N_HIDDEN, N_HIDDEN),
    torch.nn.Dropout(0.5),
    torch.nn.ReLU(),
    torch.nn.Linear(N_HIDDEN, 1),
)

print(net_overfitting)  # net architecture
print(net_dropped)
optimizer_ofit = torch.optim.Adam(net_overfitting.parameters(), lr=0.01)
optimizer_drop = torch.optim.Adam(net_dropped.parameters(), lr=0.01)
loss_func = torch.nn.MSELoss()

plt.ion()   # something about plotting

dropout可以放在activation之前，也可以放在之后，大多数是放在之前，只需要torch.nn.drop(xxx),xxx代表的是百分之多少的units起作用。

３.训练

for t in range(500):
    pred_ofit = net_overfitting(x)
    pred_drop = net_dropped(x)
    loss_ofit = loss_func(pred_ofit, y)
    loss_drop = loss_func(pred_drop, y)

    optimizer_ofit.zero_grad()
    optimizer_drop.zero_grad()
    loss_ofit.backward()
    loss_drop.backward()
    optimizer_ofit.step()
    optimizer_drop.step()
    if t % 10 == 0:
        #因为在测试的过程中是不用dropout，所以在测试过程这个需要加eval()把dropout去掉
        net_overfitting.eval()
        net_dropped.eval()
        test_pred_ofit = net_overfitting(test_x)
        test_pred_drop = net_dropped(test_x)
        plt.plot(test_x.data.numpy(), test_pred_ofit.data.numpy(), 'r-', lw=3, label='overfitting')
        plt.plot(test_x.data.numpy(), test_pred_drop.data.numpy(), 'b--', lw=3, label='dropout(50%)')
        plt.text(0, -1.2, 'overfitting loss=%.4f' % loss_func(test_pred_ofit, test_y).data.numpy(), fontdict={'size': 20, 'color':  'red'})
        plt.text(0, -1.5, 'dropout loss=%.4f' % loss_func(test_pred_drop, test_y).data.numpy(),
                 fontdict={'size': 20, 'color': 'blue'})
        plt.legend(loc='upper left');
        plt.ylim((-2.5, 2.5));
        plt.pause(0.1)
        net_overfitting.train()
        net_overfitting.train()

plt.ioff()
plt.show()

明显看出用了dropout的loss更低，并且更为平滑

这里有一个注意点，就是训练集是需要dropout的，但是测试集是不用的，所以在测试过程中，需要用到net_overfitting.eval()和net_dropped.eval()（其实net_overfitting不需要eval,但是为了保险起见，都写一遍，在测试结束接着训练时，在加上net_overfitting.train()和net_overfitting.train()

————————————————————————————————————————————————