[一]深度学习Pytorch-张量定义与张量创建
[二]深度学习Pytorch-张量的操作:拼接、切分、索引和变换
[三]深度学习Pytorch-张量数学运算
[四]深度学习Pytorch-线性回归
[五]深度学习Pytorch-计算图与动态图机制
[六]深度学习Pytorch-autograd与逻辑回归
0. 往期内容1. 自动求导函数
1.1 torch.autograd.backward(tensors)1.2 torch.autograd.grad(outputs, inputs)1.3 autograd需要注意的tips 2. 逻辑回归
2.1 逻辑回归定义2.2 逻辑回归与线性回归2.3 机器学习模型训练步骤2.4 代码实战
1. 自动求导函数 1.1 torch.autograd.backward(tensors)torch.autograd.backward(tensors, grad_tensors=None, retain_graph=None, create_graph=False)
(1)功能:自动求取梯度;
(2)参数:
tensors: 用于求导的张量,如loss;
retain_graph: 是否保存计算图,如果想要保存计算图,需要设置retain_graph为True;
create_graph: 是否创建导数计算图,一般用于高阶求导;
grad_tensors: 多梯度权重;
(3)代码示例:
# ====================================== retain_graph ==============================================
# flag = True
flag = False
if flag:
w = torch.tensor([1.], requires_grad=True)
x = torch.tensor([2.], requires_grad=True)
a = torch.add(w, x)
b = torch.add(w, 1)
y = torch.mul(a, b)
y.backward(retain_graph=True)
# print(w.grad)
#想要进行第二次反向传播需要将第一次反向传播设置为retain_graph=True
y.backward()
# ====================================== grad_tensors ==============================================
# flag = True
flag = False
if flag:
w = torch.tensor([1.], requires_grad=True)
x = torch.tensor([2.], requires_grad=True)
a = torch.add(w, x) # retain_grad()
b = torch.add(w, 1)
y0 = torch.mul(a, b) # y0 = (x+w) * (w+1) dy0/dw= 5
y1 = torch.add(a, b) # y1 = (x+w) + (w+1) dy1/dw = 2
loss = torch.cat([y0, y1], dim=0) # [y0, y1]
grad_tensors = torch.tensor([1., 1.])
loss.backward(gradient=grad_tensors) # gradient 传入 torch.autograd.backward()中的grad_tensors
print(w.grad)
#输出为7,5*1+2*1=7
1.2 torch.autograd.grad(outputs, inputs)
torch.autograd.grad(outputs, inputs, grad_outputs=None, retain_graph=None, create_graph=False)
(1)功能:求取梯度;
(2)参数:
outputs: 用于求导的张量,如loss;
inputs: 需要梯度的张量;
(3)代码示例:
# ====================================== autograd.gard ==============================================
# flag = True
flag = False
if flag:
x = torch.tensor([3.], requires_grad=True)
y = torch.pow(x, 2) # y = x**2
#create_graph一定要设置为True,它是用来创建导数的计算图
grad_1 = torch.autograd.grad(y, x, create_graph=True) # grad_1 = dy/dx = 2x = 2 * 3 = 6
print(grad_1)
#输出为6,2*3=6
#grad_1是元组,所以需要用grad_1[0]来吧梯度取出来
grad_2 = torch.autograd.grad(grad_1[0], x) # grad_2 = d(dy/dx)/dx = d(2x)/dx = 2
print(grad_2)
#输出2
1.3 autograd需要注意的tips
(1)梯度不自动清零:梯度在每一次反向传播时会自动叠加,不会清零,需要使用grad.zero_()手动清零。_下划线表示in place原地操作。
(2)代码示例:
# ====================================== tips: 1 ==============================================
# flag = True
flag = False
if flag:
w = torch.tensor([1.], requires_grad=True)
x = torch.tensor([2.], requires_grad=True)
for i in range(4):
a = torch.add(w, x)
b = torch.add(w, 1)
y = torch.mul(a, b)
y.backward()
print(w.grad)
#梯度不会自动清零,每一次反向传播会自动叠加
#需要使用grad.zero_()手动清零。
w.grad.zero_()
# ====================================== tips: 2 ==============================================
# flag = True
flag = False
if flag:
w = torch.tensor([1.], requires_grad=True)
x = torch.tensor([2.], requires_grad=True)
a = torch.add(w, x)
b = torch.add(w, 1)
y = torch.mul(a, b)
print(a.requires_grad, b.requires_grad, y.requires_grad)
#输出 True True True,因为a、b依赖于叶子节点w、x,y依赖于a、b也就是依赖于叶子节点w、x
# ====================================== tips: 3 ==============================================
# flag = True
flag = False
if flag:
a = torch.ones((1, ))
print(id(a), a)
#这里会开辟新的内存地址,两者的内存地址不一样
# a = a + torch.ones((1, ))
# print(id(a), a)
#两者的内存地址一样,+=操作是in-place操作
a += torch.ones((1, ))
print(id(a), a)
flag = True
# flag = False
if flag:
w = torch.tensor([1.], requires_grad=True)
x = torch.tensor([2.], requires_grad=True)
a = torch.add(w, x)
b = torch.add(w, 1)
y = torch.mul(a, b)
#加了这一行代码会报错,in-place操作是在原始内存中改变其数据
w.add_(1)
y.backward()
2. 逻辑回归
2.1 逻辑回归定义
2.2 逻辑回归与线性回归
对数几率回归=逻辑回归
2.3 机器学习模型训练步骤 2.4 代码实战import torch
import torch.nn as nn
import matplotlib.pyplot as plt
import numpy as np
torch.manual_seed(10)
# ============================ step 1/5 生成数据 ============================
sample_nums = 100
mean_value = 1.7
bias = 1
n_data = torch.ones(sample_nums, 2)
x0 = torch.normal(mean_value * n_data, 1) + bias # 类别0 数据 shape=(100, 2)
y0 = torch.zeros(sample_nums) # 类别0 标签 shape=(100, 1)
x1 = torch.normal(-mean_value * n_data, 1) + bias # 类别1 数据 shape=(100, 2)
y1 = torch.ones(sample_nums) # 类别1 标签 shape=(100, 1)
train_x = torch.cat((x0, x1), 0)
train_y = torch.cat((y0, y1), 0)
# ============================ step 2/5 选择模型 ============================
#g构建逻辑回归类
class LR(nn.Module):
def __init__(self):
super(LR, self).__init__()
self.features = nn.Linear(2, 1)
self.sigmoid = nn.Sigmoid()
def forward(self, x):
x = self.features(x)
x = self.sigmoid(x)
return x
lr_net = LR() # 实例化逻辑回归模型
# ============================ step 3/5 选择损失函数 ============================
#二分类交叉熵函数作为损失函数
loss_fn = nn.BCELoss()
# ============================ step 4/5 选择优化器 ============================
lr = 0.01 # 学习率
#随机梯度下降法
optimizer = torch.optim.SGD(lr_net.parameters(), lr=lr, momentum=0.9)
#momentum是引入的动量,一般设置为0.9。
# 当momentum越大时,转换为势能的能量就越大,就越有可能摆脱局部凹区域,从而进入全局凹区域。momentum主要是用于权值优化。
# 可以理解为,如果上一次的 momentum(v) 与当前的momentum负梯度方向是相同的,那这次下降的幅度就会加大,从而可以加快模型收敛。
# ============================ step 5/5 模型训练 ============================
for iteration in range(1000):
# 前向传播
y_pred = lr_net(train_x)
# 计算 loss
loss = loss_fn(y_pred.squeeze(), train_y)
# 反向传播
loss.backward()
# 更新参数
optimizer.step()
# 清空梯度
optimizer.zero_grad()
# 绘图
if iteration % 20 == 0:
mask = y_pred.ge(0.5).float().squeeze() # 以0.5为阈值进行分类
correct = (mask == train_y).sum() # 计算正确预测的样本个数
acc = correct.item() / train_y.size(0) # 计算分类准确率
plt.scatter(x0.data.numpy()[:, 0], x0.data.numpy()[:, 1], c='r', label='class 0')
plt.scatter(x1.data.numpy()[:, 0], x1.data.numpy()[:, 1], c='b', label='class 1')
w0, w1 = lr_net.features.weight[0]
w0, w1 = float(w0.item()), float(w1.item())
plot_b = float(lr_net.features.bias[0].item())
plot_x = np.arange(-6, 6, 0.1)
plot_y = (-w0 * plot_x - plot_b) / w1
plt.xlim(-5, 7)
plt.ylim(-7, 7)
plt.plot(plot_x, plot_y)
plt.text(-5, 5, 'Loss=%.4f' % loss.data.numpy(), fontdict={'size': 20, 'color': 'red'})
plt.title("Iteration: {}nw0:{:.2f} w1:{:.2f} b: {:.2f} accuracy:{:.2%}".format(iteration, w0, w1, plot_b, acc))
plt.legend()
plt.show()
plt.pause(0.5)
#模型训练停止条件
if acc > 0.99:
break



