寒假PyTorch工具第十一天

课程记录

L1, L2正则化和Dropout正则化

---

课程代码

参考: https://blog.csdn.net/weixin_43673376/article/details/107527831, 谢谢~

import torch
import torch.nn as nn
import matplotlib.pyplot as plt
from tools import set_seed
from torch.utils.tensorboard import SummaryWriter

set_seed(1)  # 设置随机种子
n_hidden = 200
max_iter = 2000
disp_interval = 200
lr_init = 0.01


def gen_data(num_data=10, x_range=(-1, 1)):
    w = 1.5
    train_x = torch.linspace(*x_range, num_data).unsqueeze_(1)
    train_y = w*train_x + torch.normal(0, 0.5, size=train_x.size())
    test_x = torch.linspace(*x_range, num_data).unsqueeze_(1)
    test_y = w*test_x + torch.normal(0, 0.3, size=test_x.size())
    return train_x, train_y, test_x, test_y
train_x, train_y, test_x, test_y = gen_data(num_data=10, x_range=(-1, 1))

class MLP(nn.Module):
    def __init__(self, neural_num):
        super(MLP, self).__init__()
        self.linears = nn.Sequential(
            nn.Linear(1, neural_num),
            nn.ReLU(inplace=True),
            nn.Linear(neural_num, neural_num),
            nn.ReLU(inplace=True),
            nn.Linear(neural_num, neural_num),
            nn.ReLU(inplace=True),
            nn.Linear(neural_num, 1),
        )

    def forward(self, x):
        return self.linears(x)

net_n = MLP(neural_num=n_hidden)
net_weight_decay = MLP(neural_num=n_hidden)

optim_n = torch.optim.SGD(net_n.parameters(), lr=lr_init, momentum=0.9)
optim_wdecay = torch.optim.SGD(net_weight_decay.parameters(), lr=lr_init, momentum=0.9, weight_decay=1e-2)
loss_fun = torch.nn.MSELoss() #均方损失
writer = SummaryWriter(comment='test', filename_suffix='test')
for epoch in range(max_iter):
    pred_normal, pred_wdecay = net_n(train_x), net_weight_decay(train_x)
    loss_n, loss_wdecay = loss_fun(pred_normal, train_y), loss_fun(pred_wdecay, train_y)
    optim_n.zero_grad()
    optim_wdecay.zero_grad()
    loss_n.backward()
    loss_wdecay.backward()
    optim_n.step() #参数更新
    optim_wdecay.step()
    if (epoch + 1) % disp_interval == 0:
        for name, layer in net_n.named_parameters(): ##
            writer.add_histogram(name + '_grad_normal', layer.grad, epoch)
            writer.add_histogram(name + '_data_normal', layer, epoch)
        for name, layer in net_weight_decay.named_parameters():
            writer.add_histogram(name + '_grad_weight_decay', layer.grad, epoch)
            writer.add_histogram(name + '_data_weight_decay', layer, epoch)
        test_pred_normal, test_pred_wdecay = net_n(test_x), net_weight_decay(test_x)
        plt.scatter(train_x.data.numpy(), train_y.data.numpy(), c='blue', s=50, alpha=0.3, label='trainc')
        plt.scatter(test_x.data.numpy(), test_y.data.numpy(), c='red', s=50, alpha=0.3, label='test')
        plt.plot(test_x.data.numpy(), test_pred_normal.data.numpy(), 'r-', lw=3, label='no weight decay')
        plt.plot(test_x.data.numpy(), test_pred_wdecay.data.numpy(), 'b--', lw=3, label='weight decay')
        plt.text(-0.25, -1.5, 'no weight decay loss={:.6f}'.format(loss_n.item()),
                 fontdict={'size': 15, 'color': 'red'})
        plt.text(-0.25, -2, 'weight decay loss={:.6f}'.format(loss_wdecay.item()),
                 fontdict={'size': 15, 'color': 'red'})
        plt.ylim(-2.5, 2.5)
        plt.legend()
        plt.title('Epoch: {}'.format(epoch +  1))
        plt.show()
        plt.close()

---

作业

1.   weight decay在pytorch的SGD中实现代码是哪一行？它对应的数学公式为？

2.   PyTorch中，Dropout在训练的时候权值尺度会进行什么操作？

1. weight decay

optim_wdecay = torch.optim.SGD(net_weight_decay.parameters(), lr=lr_init,
										 momentum=0.9, weight_decay=1e-2)
optim_wdecay.step()	

2. dropout期望

Dropout随机失活，隐藏单元以一定概率被丢弃，以1-p的概率除以1-p做拉伸，即输出单元的计算不依赖于丢弃的隐藏层单元