深度学习_多层感知机

最新推荐文章于 2024-03-02 15:05:30 发布
最新推荐文章于 2024-03-02 15:05:30 发布
阅读量515 收藏 2
点赞数
文章标签： 深度学习 人工智能 pytorch
版权声明：本文为博主原创文章，遵循 CC 4.0 BY-SA 版权协议，转载请附上原文出处链接和本声明。
本文链接：https://blog.csdn.net/weixin_42754174/article/details/125265007
版权 
%matplotlib inline
import torch
from d2l import torch as d2l

x = torch.arange(-8.0, 8.0, 0.1, requires_grad=True) 
y = torch.relu(x)
d2l.plot(x.detach(), y.detach(), 'x', 'relu(x)', figsize=(5, 2.5))

y.backward(torch.ones_like(x), retain_graph=True)
d2l.plot(x.detach(), x.grad, 'x', 'grad of relu', figsize=(5, 2.5))

y = torch.sigmoid(x)
d2l.plot(x.detach(), y.detach(), 'x', 'sigmoid(x)', figsize=(5, 2.5))

# 清除以前的梯度
x.grad.data.zero_()
y.backward(torch.ones_like(x),retain_graph=True)
d2l.plot(x.detach(), x.grad, 'x', 'grad of sigmoid', figsize=(5, 2.5))

y = torch.tanh(x)
d2l.plot(x.detach(), y.detach(), 'x', 'tanh(x)', figsize=(5, 2.5))

# 清除以前的梯度
x.grad.data.zero_()
y.backward(torch.ones_like(x),retain_graph=True)
d2l.plot(x.detach(), x.grad, 'x', 'grad of tanh', figsize=(5, 2.5))

多层感知机的从零开始实现

import torch
from torch import nn
from d2l import torch as d2l

batch_size = 256
train_iter, test_iter = d2l.load_data_fashion_mnist(batch_size)

num_inputs, num_outputs, num_hiddens = 784, 10, 256
W1 = nn.Parameter(torch.randn(
num_inputs, num_hiddens, requires_grad=True) * 0.01)
b1 = nn.Parameter(torch.zeros(num_hiddens, requires_grad=True))
W2 = nn.Parameter(torch.randn(
num_hiddens, num_outputs, requires_grad=True) * 0.01)

b2 = nn.Parameter(torch.zeros(num_outputs, requires_grad=True))
params = [W1, b1, W2, b2]

def relu(X):
    a = torch.zeros_like(X)
    return torch.max(X, a)

def net(X):
    X = X.reshape((-1, num_inputs))
    H = relu(X@W1 + b1) # 这⾥“@”代表矩阵乘法
    return (H@W2 + b2)

loss = nn.CrossEntropyLoss(reduction='none')

num_epochs, lr = 10, 0.1
updater = torch.optim.SGD(params, lr=lr)
d2l.train_ch3(net, train_iter, test_iter, loss, num_epochs, updater)

d2l.predict_ch3(net, test_iter)

在所有其他参数保持不变的情况下，更改超参数num_hiddens的值，并查看此超参数的变化对结果有 何影响。确定此超参数的最佳值。

num_inputs, num_outputs, num_hiddens128 = 784, 10, 128
W1 = nn.Parameter(torch.randn(
num_inputs, num_hiddens128, requires_grad=True) * 0.01)
b1 = nn.Parameter(torch.zeros(num_hiddens128, requires_grad=True))
W2 = nn.Parameter(torch.randn(
num_hiddens128, num_outputs, requires_grad=True) * 0.01)

b2 = nn.Parameter(torch.zeros(num_outputs, requires_grad=True))
params = [W1, b1, W2, b2]

def relu(X):
    a = torch.zeros_like(X)
    return torch.max(X, a)

def net(X):
    X = X.reshape((-1, num_inputs))
    H = relu(X@W1 + b1) # 这⾥“@”代表矩阵乘法
    return (H@W2 + b2)

loss = nn.CrossEntropyLoss(reduction='none')

num_epochs, lr = 10, 0.1
updater = torch.optim.SGD(params, lr=lr)
d2l.train_ch3(net, train_iter, test_iter, loss, num_epochs, updater)

num_inputs, num_outputs, num_hiddens512 = 784, 10, 512
W1 = nn.Parameter(torch.randn(
num_inputs, num_hiddens512, requires_grad=True) * 0.01)
b1 = nn.Parameter(torch.zeros(num_hiddens512, requires_grad=True))
W2 = nn.Parameter(torch.randn(
num_hiddens512, num_outputs, requires_grad=True) * 0.01)

b2 = nn.Parameter(torch.zeros(num_outputs, requires_grad=True))
params = [W1, b1, W2, b2]

def relu(X):
    a = torch.zeros_like(X)
    return torch.max(X, a)

def net(X):
    X = X.reshape((-1, num_inputs))
    H = relu(X@W1 + b1) # 这⾥“@”代表矩阵乘法
    return (H@W2 + b2)

loss = nn.CrossEntropyLoss(reduction='none')

num_epochs, lr = 10, 0.1
updater = torch.optim.SGD(params, lr=lr)
d2l.train_ch3(net, train_iter, test_iter, loss, num_epochs, updater)

尝试添加更多的隐藏层，并查看它对结果有何影响。

num_inputs, num_outputs, num_hiddens512, num_hiddens= 784, 10, 512, 256
W1 = nn.Parameter(torch.randn(
num_inputs, num_hiddens512, requires_grad=True) * 0.01)
b1 = nn.Parameter(torch.zeros(num_hiddens512, requires_grad=True))

W2 = nn.Parameter(torch.randn(
num_hiddens512, num_hiddens, requires_grad=True) * 0.01)
b2 = nn.Parameter(torch.zeros(num_hiddens, requires_grad=True))

W3 = nn.Parameter(torch.randn(
num_hiddens, num_outputs, requires_grad=True) * 0.01)
b3 = nn.Parameter(torch.zeros(num_outputs, requires_grad=True))


params = [W1, b1, W2, b2, W3, b3]

def relu(X):
    a = torch.zeros_like(X)
    return torch.max(X, a)

def net(X):
    X = X.reshape((-1, num_inputs))
    H1 = relu(X@W1 + b1) # 这⾥“@”代表矩阵乘法
    H2 = relu(H1@W2 + b2)
    return (H2@W3 + b3)

loss = nn.CrossEntropyLoss(reduction='none')

num_epochs, lr = 10, 0.1
updater = torch.optim.SGD(params, lr=lr)
d2l.train_ch3(net, train_iter, test_iter, loss, num_epochs, updater)

改变学习速率会如何影响结果？保持模型架构和其他超参数（包括轮数）不变，学习率设置为多少会带 来最好的结果？

num_inputs, num_outputs, num_hiddens = 784, 10, 256
W1 = nn.Parameter(torch.randn(
num_inputs, num_hiddens, requires_grad=True) * 0.01)
b1 = nn.Parameter(torch.zeros(num_hiddens, requires_grad=True))
W2 = nn.Parameter(torch.randn(
num_hiddens, num_outputs, requires_grad=True) * 0.01)

b2 = nn.Parameter(torch.zeros(num_outputs, requires_grad=True))
params = [W1, b1, W2, b2]

def relu(X):
    a = torch.zeros_like(X)
    return torch.max(X, a)

def net(X):
    X = X.reshape((-1, num_inputs))
    H = relu(X@W1 + b1) # 这⾥“@”代表矩阵乘法
    return (H@W2 + b2)

loss = nn.CrossEntropyLoss(reduction='none')

num_epochs, lr = 10, 0.05
updater = torch.optim.SGD(params, lr=lr)
d2l.train_ch3(net, train_iter, test_iter, loss, num_epochs, updater)

num_inputs, num_outputs, num_hiddens = 784, 10, 256
W1 = nn.Parameter(torch.randn(
num_inputs, num_hiddens, requires_grad=True) * 0.01)
b1 = nn.Parameter(torch.zeros(num_hiddens, requires_grad=True))
W2 = nn.Parameter(torch.randn(
num_hiddens, num_outputs, requires_grad=True) * 0.01)

b2 = nn.Parameter(torch.zeros(num_outputs, requires_grad=True))
params = [W1, b1, W2, b2]

def relu(X):
    a = torch.zeros_like(X)
    return torch.max(X, a)

def net(X):
    X = X.reshape((-1, num_inputs))
    H = relu(X@W1 + b1) # 这⾥“@”代表矩阵乘法
    return (H@W2 + b2)

loss = nn.CrossEntropyLoss(reduction='none')

num_epochs, lr = 10, 0.5
updater = torch.optim.SGD(params, lr=lr)
d2l.train_ch3(net, train_iter, test_iter, loss, num_epochs, updater)

通过对所有超参数（学习率、轮数、隐藏层数、每层的隐藏单元数）进⾏联合优化，可以得到的最佳结 果是什么？

num_inputs, num_outputs, num_hiddens512, num_hiddens= 784, 10, 512, 256
W1 = nn.Parameter(torch.randn(
num_inputs, num_hiddens512, requires_grad=True) * 0.01)
b1 = nn.Parameter(torch.zeros(num_hiddens512, requires_grad=True))

W2 = nn.Parameter(torch.randn(
num_hiddens512, num_hiddens, requires_grad=True) * 0.01)
b2 = nn.Parameter(torch.zeros(num_hiddens, requires_grad=True))

W3 = nn.Parameter(torch.randn(
num_hiddens, num_outputs, requires_grad=True) * 0.01)
b3 = nn.Parameter(torch.zeros(num_outputs, requires_grad=True))


params = [W1, b1, W2, b2, W3, b3]

def relu(X):
    a = torch.zeros_like(X)
    return torch.max(X, a)

def net(X):
    X = X.reshape((-1, num_inputs))
    H1 = relu(X@W1 + b1) # 这⾥“@”代表矩阵乘法
    H2 = relu(H1@W2 + b2)
    return (H2@W3 + b3)

loss = nn.CrossEntropyLoss(reduction='none')

num_epochs, lr = 10, 0.5
updater = torch.optim.SGD(params, lr=lr)
d2l.train_ch3(net, train_iter, test_iter, loss, num_epochs, updater)

多层感知机的简洁实现

import torch
from torch import nn
from d2l import torch as d2l

net = nn.Sequential(nn.Flatten(),
                    nn.Linear(784, 256),
                    nn.ReLU(),
                    nn.Linear(256, 10))
def init_weights(m):
    if type(m) == nn.Linear:
        nn.init.normal_(m.weight, std=0.01)
net.apply(init_weights)

batch_size, lr, num_epochs = 256, 0.1, 10
loss = nn.CrossEntropyLoss(reduction='none')
trainer = torch.optim.SGD(net.parameters(), lr=lr)

train_iter, test_iter = d2l.load_data_fashion_mnist(batch_size)
d2l.train_ch3(net, train_iter, test_iter, loss, num_epochs, trainer)

尝试添加不同数量的隐藏层（也可以修改学习率），怎么样设置效果最好？

net = nn.Sequential(nn.Flatten(),
                    nn.Linear(784, 512),
                    nn.ReLU(),
                    nn.Linear(512,256),
                    nn.ReLU(),
                    nn.Linear(256,128),
                    nn.ReLU(),
                    nn.Linear(128, 10))
def init_weights(m):
    if type(m) == nn.Linear:
        nn.init.normal_(m.weight, std=0.01)
net.apply(init_weights)

batch_size, lr, num_epochs = 256, 0.1, 10
loss = nn.CrossEntropyLoss(reduction='none')
trainer = torch.optim.SGD(net.parameters(), lr=lr)

train_iter, test_iter = d2l.load_data_fashion_mnist(batch_size)
d2l.train_ch3(net, train_iter, test_iter, loss, num_epochs, trainer)

net = nn.Sequential(nn.Flatten(),
                    nn.Linear(784, 256),
                    nn.Tanh(),
                    nn.Linear(256, 10))
def init_weights(m):
    if type(m) == nn.Linear:
        nn.init.normal_(m.weight, std=0.01)
net.apply(init_weights)

batch_size, lr, num_epochs = 256, 0.1, 10
loss = nn.CrossEntropyLoss(reduction='none')
trainer = torch.optim.SGD(net.parameters(), lr=lr)

train_iter, test_iter = d2l.load_data_fashion_mnist(batch_size)
d2l.train_ch3(net, train_iter, test_iter, loss, num_epochs, trainer)

net = nn.Sequential(nn.Flatten(),
                    nn.Linear(784, 256),
                    nn.Sigmoid(),
                    nn.Linear(256, 10))
def init_weights(m):
    if type(m) == nn.Linear:
        nn.init.normal_(m.weight, std=0.01)
net.apply(init_weights)

batch_size, lr, num_epochs = 256, 0.1, 10
loss = nn.CrossEntropyLoss(reduction='none')
trainer = torch.optim.SGD(net.parameters(), lr=lr)

train_iter, test_iter = d2l.load_data_fashion_mnist(batch_size)
d2l.train_ch3(net, train_iter, test_iter, loss, num_epochs, trainer)

尝试不同的⽅案来初始化权重，什么⽅法效果最好？

net = nn.Sequential(nn.Flatten(),
                    nn.Linear(784, 256),
                    nn.ReLU(),
                    nn.Linear(256, 10))
def init_weights(m):
    if type(m) == nn.Linear:
#         nn.init.normal_(m.weight, std=0.01)
          nn.init.zeros_(m.weight)
net.apply(init_weights)

batch_size, lr, num_epochs = 256, 0.1, 10
loss = nn.CrossEntropyLoss(reduction='none')
trainer = torch.optim.SGD(net.parameters(), lr=lr)

train_iter, test_iter = d2l.load_data_fashion_mnist(batch_size)
d2l.train_ch3(net, train_iter, test_iter, loss, num_epochs, trainer)

net = nn.Sequential(nn.Flatten(),
                    nn.Linear(784, 256),
                    nn.ReLU(),
                    nn.Linear(256, 10))
def init_weights(m):
    if type(m) == nn.Linear:
#         nn.init.normal_(m.weight, std=0.01)
#           nn.init.zeros_(m.weight)
            nn.init.ones_(m.weight)
    
net.apply(init_weights)

batch_size, lr, num_epochs = 256, 0.1, 10
loss = nn.CrossEntropyLoss(reduction='none')
trainer = torch.optim.SGD(net.parameters(), lr=lr)

train_iter, test_iter = d2l.load_data_fashion_mnist(batch_size)
d2l.train_ch3(net, train_iter, test_iter, loss, num_epochs, trainer)

模型选择、欠拟合和过拟合

import math
import numpy as np
import torch
from torch import nn
from d2l import torch as d2l

max_degree = 20 # 多项式的最⼤阶数
n_train, n_test = 100, 100 # 训练和测试数据集⼤⼩
true_w = np.zeros(max_degree) # 分配⼤量的空间
true_w[0:4] = np.array([5, 1.2, -3.4, 5.6])

features = np.random.normal(size=(n_train + n_test, 1))
np.random.shuffle(features)
poly_features = np.power(features, np.arange(max_degree).reshape(1, -1))
for i in range(max_degree):
    poly_features[:, i] /= math.gamma(i + 1) # gamma(n)=(n-1)!
# labels的维度:(n_train+n_test,)
labels = np.dot(poly_features, true_w)
labels += np.random.normal(scale=0.1, size=labels.shape)

# NumPy ndarray转换为tensor
true_w, features, poly_features, labels = [torch.tensor(x, dtype=
    torch.float32) for x in [true_w, features, poly_features, labels]]

features[:2], poly_features[:2, :], labels[:2]

def evaluate_loss(net, data_iter, loss): #@save
    """评估给定数据集上模型的损失"""
    metric = d2l.Accumulator(2) # 损失的总和,样本数量
    for X, y in data_iter:
        out = net(X)
        y = y.reshape(out.shape)
        l = loss(out, y)
        metric.add(l.sum(), l.numel())
    return metric[0] / metric[1]

def train(train_features, test_features, train_labels, test_labels,
            num_epochs=400):
    loss = nn.MSELoss(reduction='none')
    input_shape = train_features.shape[-1] # 不设置偏置，因为我们已经在多项式中实现了它
    net = nn.Sequential(nn.Linear(input_shape, 1, bias=False))
    batch_size = min(10, train_labels.shape[0])
    train_iter = d2l.load_array((train_features, train_labels.reshape(-1,1)),
                                    batch_size)
    test_iter = d2l.load_array((test_features, test_labels.reshape(-1,1)),
                                    batch_size, is_train=False)
    trainer = torch.optim.SGD(net.parameters(), lr=0.01)
    animator = d2l.Animator(xlabel='epoch', ylabel='loss', yscale='log',
                            xlim=[1, num_epochs], ylim=[1e-3, 1e2],
                            legend=['train', 'test'])
    for epoch in range(num_epochs):
        d2l.train_epoch_ch3(net, train_iter, loss, trainer)
        if epoch == 0 or (epoch + 1) % 20 == 0:
            animator.add(epoch + 1, (evaluate_loss(net, train_iter, loss),
                                    evaluate_loss(net, test_iter, loss)))
    print('weight:', net[0].weight.data.numpy())

# 从多项式特征中选择前4个维度，即1,x,x^2/2!,x^3/3!
train(poly_features[:n_train, :4], poly_features[n_train:, :4],
        labels[:n_train], labels[n_train:])

# 从多项式特征中选择前2个维度，即1和x
train(poly_features[:n_train, :2], poly_features[n_train:, :2],
        labels[:n_train], labels[n_train:])

# 从多项式特征中选取所有维度
train(poly_features[:n_train, :], poly_features[n_train:, :],
        labels[:n_train], labels[n_train:], num_epochs=1500)

权重衰减

%matplotlib inline
import torch
from torch import nn
from d2l import torch as d2l

n_train, n_test, num_inputs, batch_size = 20, 100, 200, 5
true_w, true_b = torch.ones((num_inputs, 1)) * 0.01, 0.05
train_data = d2l.synthetic_data(true_w, true_b, n_train)
train_iter = d2l.load_array(train_data, batch_size)
test_data = d2l.synthetic_data(true_w, true_b, n_test)
test_iter = d2l.load_array(test_data, batch_size, is_train=False)

def init_params():
    w = torch.normal(0, 1, size=(num_inputs, 1), requires_grad=True) 
    b = torch.zeros(1, requires_grad=True)
    return [w, b]

def l2_penalty(w):
    return torch.sum(w.pow(2)) / 2

def train(lambd):
    w, b = init_params()
    net, loss = lambda X: d2l.linreg(X, w, b), d2l.squared_loss
    num_epochs, lr = 100, 0.003
    animator = d2l.Animator(xlabel='epochs', ylabel='loss', yscale='log',
                            xlim=[5, num_epochs], legend=['train', 'test'])
    for epoch in range(num_epochs):
        for X, y in train_iter:
            # 增加了L2范数惩罚项，
            # ⼴播机制使l2_penalty(w)成为⼀个⻓度为batch_size的向量
            l = loss(net(X), y) + lambd * l2_penalty(w)
            l.sum().backward()
            d2l.sgd([w, b], lr, batch_size)
        if (epoch + 1) % 5 == 0:
            animator.add(epoch + 1, (d2l.evaluate_loss(net, train_iter, loss),
                                        d2l.evaluate_loss(net, test_iter, loss)))
    print('w的L2范数是：', torch.norm(w).item())

train(lambd=0)

train(lambd=3)

简洁实现

def train_concise(wd):
    net = nn.Sequential(nn.Linear(num_inputs, 1))
    for param in net.parameters():
        param.data.normal_()
    loss = nn.MSELoss(reduction='none')
    num_epochs, lr = 100, 0.003
    # 偏置参数没有衰减
    trainer = torch.optim.SGD([
        {"params":net[0].weight,'weight_decay': wd},
        {"params":net[0].bias}], lr=lr)
    animator = d2l.Animator(xlabel='epochs', ylabel='loss', yscale='log',
                            xlim=[5, num_epochs], legend=['train', 'test'])
    for epoch in range(num_epochs):
        for X, y in train_iter:
            trainer.zero_grad()
            l = loss(net(X), y)
            l.mean().backward()
            trainer.step()
        if (epoch + 1) % 5 == 0:
            animator.add(epoch + 1,
                            (d2l.evaluate_loss(net, train_iter, loss),
                            d2l.evaluate_loss(net, test_iter, loss)))
    print('w的L2范数：', net[0].weight.norm().item())

train_concise(0)

train_concise(3)

在本节的估计问题中使⽤λ的值进⾏实验。绘制训练和测试精度关于λ的函数。你观察到了什么？

def train_concise(wd):
    net = nn.Sequential(nn.Linear(num_inputs, 1))
    
    for param in net.parameters():
        for wd in range(0,20):
            param.data.normal_()
        loss = nn.MSELoss(reduction='none')
        num_epochs, lr = 20, 0.003

        # 偏置参数没有衰减
        trainer = torch.optim.SGD([
            {"params":net[0].weight,'weight_decay': wd},
            {"params":net[0].bias}], lr=lr)
        animator_lamda = d2l.Animator(xlabel='lamdas', ylabel='loss', yscale='log',
                                xlim=[5, num_epochs], legend=['train', 'test'])
        for lamda in range(0,20):
            for X, y in train_iter:
                trainer.zero_grad()
                l = loss(net(X), y)
                l.mean().backward()
                trainer.step()
            if (lamda + 1) % 5 == 0:
                animator_lamda.add(lamda + 1,
                                (d2l.evaluate_loss(net, train_iter, loss),
                                d2l.evaluate_loss(net, test_iter, loss)))
        print('w的L2范数：', net[0].weight.norm().item())

    

train_concise(0)

如果我们使⽤∑i |wi|作为我们选择的惩罚（L1正则化），那么更新⽅程会是什么样⼦？

def train_conciseL1(wd):
    net = nn.Sequential(nn.Linear(num_inputs, 1))
    for param in net.parameters():
        param.data.normal_()
    loss = nn.L1Loss(reduction='none')
    num_epochs, lr = 100, 0.003
    # 偏置参数没有衰减
    trainer = torch.optim.SGD([
        {"params":net[0].weight,'weight_decay': wd},
        {"params":net[0].bias}], lr=lr)
    animator = d2l.Animator(xlabel='epochs', ylabel='loss', yscale='log',
                            xlim=[5, num_epochs], legend=['train', 'test'])
    for epoch in range(num_epochs):
        for X, y in train_iter:
            trainer.zero_grad()
            l = loss(net(X), y)
            l.mean().backward()
            trainer.step()
        if (epoch + 1) % 5 == 0:
            animator.add(epoch + 1,
                            (d2l.evaluate_loss(net, train_iter, loss),
                            d2l.evaluate_loss(net, test_iter, loss)))
    print('w的L1范数：', net[0].weight.norm().item())

train_conciseL1(0)

4.6 暂退法（Dropout）

import torch
from torch import nn
from d2l import torch as d2l

def dropout_layer(X, dropout):
    assert 0 <= dropout <= 1 # 在本情况中，所有元素都被丢弃
    if dropout == 1:
        return torch.zeros_like(X)
    # 在本情况中，所有元素都被保留
    if dropout == 0:
        return X
    mask = (torch.rand(X.shape) > dropout).float()
    return mask * X / (1.0 - dropout)

X= torch.arange(16, dtype = torch.float32).reshape((2, 8))
print(X)
print(dropout_layer(X, 0.))
print(dropout_layer(X, 0.5))
print(dropout_layer(X, 1.))

num_inputs, num_outputs, num_hiddens1, num_hiddens2 = 784, 10, 256, 256

dropout1, dropout2 = 0.2, 0.5
class Net(nn.Module):
    def __init__(self, num_inputs, num_outputs, num_hiddens1, num_hiddens2,
                    is_training = True):
        super(Net, self).__init__()
        self.num_inputs = num_inputs
        self.training = is_training
        self.lin1 = nn.Linear(num_inputs, num_hiddens1)
        self.lin2 = nn.Linear(num_hiddens1, num_hiddens2)
        self.lin3 = nn.Linear(num_hiddens2, num_outputs)
        self.relu = nn.ReLU()
        
    def forward(self, X):
        H1 = self.relu(self.lin1(X.reshape((-1, self.num_inputs))))
        # 只有在训练模型时才使⽤dropout
        if self.training == True: # 在第⼀个全连接层之后添加⼀个dropout层
            H1 = dropout_layer(H1, dropout1)
        H2 = self.relu(self.lin2(H1))
        if self.training == True: # 在第⼆个全连接层之后添加⼀个dropout层
            H2 = dropout_layer(H2, dropout2)
        out = self.lin3(H2)
        return out
    
net = Net(num_inputs, num_outputs, num_hiddens1, num_hiddens2)

num_epochs, lr, batch_size = 10, 0.5, 256
loss = nn.CrossEntropyLoss(reduction='none')
train_iter, test_iter = d2l.load_data_fashion_mnist(batch_size)
trainer = torch.optim.SGD(net.parameters(), lr=lr)
d2l.train_ch3(net, train_iter, test_iter, loss, num_epochs, trainer)

net = nn.Sequential(nn.Flatten(),
        nn.Linear(784, 256),
        nn.ReLU(),
        # 在第⼀个全连接层之后添加⼀个dropout层
        nn.Dropout(dropout1),
        nn.Linear(256, 256),
        nn.ReLU(),
        # 在第⼆个全连接层之后添加⼀个dropout层
        nn.Dropout(dropout2),
        nn.Linear(256, 10))
def init_weights(m):
    if type(m) == nn.Linear:
        nn.init.normal_(m.weight, std=0.01)

net.apply(init_weights)
trainer = torch.optim.SGD(net.parameters(), lr=lr)
d2l.train_ch3(net, train_iter, test_iter, loss, num_epochs, trainer)

如果更改第⼀层和第⼆层的暂退法概率，会发⽣什么情况？具体地说，如果交换这两个层，会发⽣什么 情况？设计⼀个实验来回答这些问题，定量描述你的结果，并总结定性的结论。

dropout1, dropout2 = 0.2, 0.5
dropoutX1, dropoutX2 = 0.4, 0.8

net = nn.Sequential(nn.Flatten(),
        nn.Linear(784, 256),
        nn.ReLU(),
        # 在第⼀个全连接层之后添加⼀个dropout层
        nn.Dropout(dropoutX1),
        nn.Linear(256, 256),
        nn.ReLU(),
        # 在第⼆个全连接层之后添加⼀个dropout层
        nn.Dropout(dropoutX2),
        nn.Linear(256, 10))
def init_weights(m):
    if type(m) == nn.Linear:
        nn.init.normal_(m.weight, std=0.01)
        
net.apply(init_weights)
trainer = torch.optim.SGD(net.parameters(), lr=lr)
d2l.train_ch3(net, train_iter, test_iter, loss, num_epochs, trainer)

dropout1, dropout2 = 0.2, 0.5
dropoutX1, dropoutX2 = 0.4, 0.8

net = nn.Sequential(nn.Flatten(),
        nn.Linear(784, 256),
        nn.ReLU(),
        # 在第⼀个全连接层之后添加⼀个dropout层
        nn.Dropout(dropout2),
        nn.Linear(256, 256),
        nn.ReLU(),
        # 在第⼆个全连接层之后添加⼀个dropout层
        nn.Dropout(dropout1),
        nn.Linear(256, 10))
def init_weights(m):
    if type(m) == nn.Linear:
        nn.init.normal_(m.weight, std=0.01)
        
net.apply(init_weights)
trainer = torch.optim.SGD(net.parameters(), lr=lr)
d2l.train_ch3(net, train_iter, test_iter, loss, num_epochs, trainer)

增加训练轮数，并将使⽤暂退法和不使⽤暂退法时获得的结果进⾏⽐较。

num_epochsX, lr, batch_size = 20, 0.5, 256
dropoutX3, dropoutX4 = 0, 0

net = nn.Sequential(nn.Flatten(),
        nn.Linear(784, 256),
        nn.ReLU(),
        # 在第⼀个全连接层之后添加⼀个dropout层
        nn.Dropout(dropoutX3),
        nn.Linear(256, 256),
        nn.ReLU(),
        # 在第⼆个全连接层之后添加⼀个dropout层
        nn.Dropout(dropoutX4),
        nn.Linear(256, 10))
def init_weights(m):
    if type(m) == nn.Linear:
        nn.init.normal_(m.weight, std=0.01)
        
net.apply(init_weights)
trainer = torch.optim.SGD(net.parameters(), lr=lr)
d2l.train_ch3(net, train_iter, test_iter, loss, num_epochsX, trainer)

李沐动手学深度学习 https://zh-v2.d2l.ai/chapter_introduction/index.html

JamyJ
关注
  • 0
    点赞
  • 踩
  • 2
    收藏
    觉得还不错? 一键收藏
  • 1
    评论
多层感知机代码实现.py
05-04
import torch from torchvision import datasets from torchvision import transforms import torch.nn as nn import torch.optim as optim train_data = datasets.MNIST(root="D:/Python_file/Deep_Learning/...
动手学深度学习笔记第四章（多层感知器）
qq_46144849的博客
01-13 883
深度学习知识
动手学深度学习--4.多层感知机
weixin_43845922的博客
10-20 1861
动手学深度学习
李沐动手学习深度学习——4.2练习
weixin_45496725的博客
03-02 674
多层感知机的从零开始实现
《从深度学习基础到车道线模型训练与部署》学习（三）
majunyongzhangduo的博客
01-05 940
1.Python常用库之一：Numpy库 NumPy是Python语言的一个扩充程序库。支持高级大量的维度数组与矩阵运算，此外也针对数组运算提供大量的数学函数库。总而言之，包含： 一个强大的N维数组对象 ndarray 广播功能函数 整合 C/C++/Fortran 代码的工具 线性代数、傅里叶变换、随机数生成等功能 2.Python常用库之二：Pandas库 Pandas是用于数据操纵和分析，建立在Numpy之上的。Pandas为Python带来了两种新的数据结构：Pandas Series和
pytorch中backward()函数详解
m0_45275062的博客
09-03 4004
在最开始使用Pytorch时，关于反向传播函数，有时会报错，简单的说，如果你的Y值是个标量，那么直接使用: y.backward() 就可以了，但是如果你的Y是个向量或矩阵，那么就不一样: y.backward(torch.ones_like(x)) 具体原因如下： (19条消息) pytorch中backward()函数详解_Camlin_Z的博客-CSDN博客 ...
多层感知机+快问快答
qq_60567866的博客
05-30 447
import torch from torch import nn from d2l import torch as d2l batch_size = 256 train_iter, test_iter = d2l.load_data_fashion_mnist(batch_size) #28*28 隐藏层， num_inputs, num_outputs, num_hiddens = 784, 10, 256 #这里有个小问题，我们是全部设成随机数 一会.
TensorFlow实现多层感知机.py
03-30
本代码是在TensorFlow实现softmax regression模型之后的优化，增加了一层隐含层来提高拟合度，并使用dropout减轻过拟合、自适应学习速率的Adagrad，以及可以解决梯度弥散的激活函数ReLU。 本代码按照《TensorFlow...
深度学习Tensorflow逻辑回归多层感知机深度神经网络搭建基础学习
最新发布
03-30
它分别包括使用Tensorflow1.1和Keras2.2搭建简单感知机、逻辑回归、多分类问题、多层感知器(MLP)、深度神经网络等的模型的代码和案例。 最后给出了使用Tensorflow搭建深度神经网络的通用模板。 适合Tensorflow深度...
多层感知机matlab实现.rar
02-29
多层感知机matlab代码实现，分类是采用较难分类的半月形两类问题分类，可方便深度学习初学者了解深度学习反馈传播机制。
多层感知机 python pytorch
02-13
多层感知机的基本知识 深度学习主要关注多层模型。在这里，我们将以多层感知机（multilayer perceptron，MLP）为例，介绍多层神经网络的概念。
PyTorch基础（一）
onion___的博客
03-29 457
1. 张量 在同构的意义下： 第零阶张量 （r = 0） 为标量 （Scalar）， 第一阶张量 （r = 1） 为向量 （Vector）， 第二阶张量 （r = 2） 则成为矩阵 （Matrix）， 第三阶以上的统称为多维张量。 如果张量中只有一个元素，可以用tensor.item()取出其数值 Tensor的基本数据类型有五种： 32位浮点型：torch.FloatTensor。 (默认) 64位整型：torch.LongTensor。 32位整型：torch.IntTensor。 16位整型：tor
多层感知机+代码实现
weixin_45337294的博客
03-16 3406
感知机 给定输入X，权重w，偏移b，感知机输出 训练感知机 initialize w=0 and b=0 #此处的w直接取0 reprat if yi[<wi,xi>+b]≤0 then #此处表示预测错误，要使得≤0，若[<wi,xi>+b]≤0，对应如上公式，yi=-1，如此yi[<wi,xi>+b]必定不会≤0。反之同理 w<-w+yixi and b<-b+yi#对w和b进行更新 end if until all clas
神经网络与深度学习--课堂笔记（1）
shuaibidebei的博客
03-16 106
神经网络与深度学习--课堂笔记（1） 主要介绍神经元、激活函数、BP神经网络的算法推导
动手学深度学习学习记录——多层感知机
qq_43571742的博客
03-24 129
动手学深度学习学习记录
动手学深度学习 第四章 多层感知机 练习（Pytorch版）
Rita's Blog
05-09 3260
学习《动手学深度学习》过程中完成的作业
动手学深度学习Day04
南邮在读本科生的博客
01-27 288
动手学深度学习Day04
动手深度学习笔记（十六）4.2. 多层感知机的从零开始实现
juluwangriyue的博客
02-16 1985
动手深度学习笔记（十六）4.2. 多层感知机的从零开始实现4. 多层感知机4.2. 多层感知机的从零开始实现4.2.1. 初始化模型参数4.2.2. 激活函数4.2.3. 模型4.2.4. 损失函数4.2.5. 训练4.2.6. 小结4.2.7. 练习 4. 多层感知机 4.2. 多层感知机的从零开始实现 我们已经在 4.1节中描述了多层感知机（MLP）， 现在让我们尝试自己实现一个多层感知机。 为了与之前softmax回归（ 3.6节 ） 获得的结果进行比较， 我们将继续使用Fashion-MNIST图像
mlp多层感知机与深度学习
09-03
mlp多层感知机与深度学习密切相关。深度学习是一种机器学习的方法，它基于神经网络的多层结构进行建模和训练。多层感知机(MLP)是最简单的神经网络结构之一，也是深度学习中常用的模型之一。 MLP由多个计算层组成，包括输入层、若干个隐层和输出层。每个隐层都由多个隐单元组成，而隐单元的个数是根据数据集的复杂度来确定的。对于简单的数据集，可以选择较少的隐单元，而对于复杂的数据集，可以选择更多的隐单元，甚至可以添加多个隐层。 例如，当我们使用MLP进行分类任务时，可以使用两个隐层，每个隐层都包含10个隐单元。这样的设置可以通过修改MLPClassifier类的hidden_layer_sizes属性来实现。具体的代码如下所示： ``` mlp = MLPClassifier(solver='lbfgs', random_state=0, hidden_layer_sizes=[10, 10]) ``` 这个设置意味着我们有两个隐层，每个隐层都有10个隐单元。 因此，通过使用MLP多层感知机，我们可以实现深度学习中的模型构建和训练，以解决各种机器学习问题。<span class="em">1</span><span class="em">2</span><span class="em">3</span> #### 引用[.reference_title] - *1* *2* *3* [深度学习简介与MLP多层感知机](https://blog.csdn.net/qq_43355223/article/details/86593078)[target="_blank" data-report-click={"spm":"1018.2226.3001.9630","extra":{"utm_source":"vip_chatgpt_common_search_pc_result","utm_medium":"distribute.pc_search_result.none-task-cask-2~all~insert_cask~default-1-null.142^v93^chatsearchT3_1"}}] [.reference_item style="max-width: 100%"] [ .reference_list ]

“相关推荐”对你有帮助么？

  • 非常没帮助
  • 没帮助
  • 一般
  • 有帮助
  • 非常有帮助
提交
评论 1
添加红包

请填写红包祝福语或标题

个

红包个数最小为10个

元

红包金额最低5元

当前余额3.43元 前往充值 >
需支付：10.00元
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付元
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明：

1.余额是钱包充值的虚拟货币，按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载，可以购买VIP、付费专栏及课程。

余额充值