Pytorch实现线性回归中的梯度推导和验证，线性回归用的torch.nn实现

代码：主要进行前向和梯度计算，没有进行反向传播

import numpy as np

x_values = [i for i in range(11)]
x_train = np.array(x_values, dtype=np.float32)
x_train = x_train.reshape(-1,1)
print(x_train.shape)

y_values = [2*i+1 for i in x_values]
y_train = np.array(y_values, dtype=np.float32)
y_train = y_train.reshape(-1, 1)
print(y_train.shape)


import torch
import torch.nn as nn

class LinearRegressionModel(nn.Module):
    def __init__(self, input_dim, output_dim):
        super(LinearRegressionModel, self).__init__()
        self.linear = nn.Linear(input_dim, output_dim)

    def forward(self, x):
        out = self.linear(x)
        return out

input_dim = 1
output_dim = 1

model = LinearRegressionModel(input_dim, output_dim)

epochs = 1
learning_rate = 0.01
optimizer = torch.optim.SGD(model.parameters(), lr=learning_rate)
criterion = nn.MSELoss()


# torch.manual_seed(7) # cpu
# torch.cuda.manual_seed(7) #gpu
# np.random.seed(7) #numpy
# random.seed(7) #random and transforms
# torch.backends.cudnn.deterministic=True # cudnn

for epoch in range(epochs):
    epoch += 1

    inputs = torch.from_numpy(x_train)
    labels = torch.from_numpy(y_train)

    optimizer.zero_grad()
    model.linear.weight.data = torch.tensor([2], dtype=torch.float32).reshape([1,1])
    # model.linear.bias.data = torch.tensor([1], dtype=torch.float32).reshape([1,1])
    # torch.tensor().reshape([1,1])

    outputs = model(inputs)

    loss = criterion(outputs, labels)

    print(loss)
    print(torch.sum(pow(outputs-labels, 2))/11)
    print(outputs - labels)

    loss.backward()

    print(model.linear.weight.grad)
    print(model.linear.bias.grad)
    print(2*torch.mean(outputs - labels))

    print(model.linear.weight)
    print(model.linear.bias)

    # optimizer.step()

    if epoch % 50 == 0:
        print('epoch {}, loss{}'.format(epoch, loss.item()))

MSELoss 就是 

对bias进行求导就是

 

同理

 

代码2：进行一次反向传播

import numpy as np

x_values = [i for i in range(11)]
x_train = np.array(x_values, dtype=np.float32)
x_train = x_train.reshape(-1,1)
print(x_train.shape)

y_values = [2*i+1 for i in x_values]
y_train = np.array(y_values, dtype=np.float32)
y_train = y_train.reshape(-1, 1)
print(y_train.shape)

import torch
import torch.nn as nn

class LinearRegressionModel(nn.Module):
    def __init__(self, input_dim, output_dim):
        super(LinearRegressionModel, self).__init__()
        self.linear = nn.Linear(input_dim, output_dim)

    def forward(self, x):
        out = self.linear(x)
        return out

input_dim = 1
output_dim = 1

model = LinearRegressionModel(input_dim, output_dim)

epochs = 1
learning_rate = 0.01
optimizer = torch.optim.SGD(model.parameters(), lr=learning_rate)
criterion = nn.MSELoss()


torch.manual_seed(7) # cpu
torch.cuda.manual_seed(7) #gpu
np.random.seed(7) #numpy
# random.seed(7) #random and transforms
torch.backends.cudnn.deterministic=True # cudnn

for epoch in range(epochs):
    epoch += 1

    inputs = torch.from_numpy(x_train)
    labels = torch.from_numpy(y_train)

    optimizer.zero_grad()
    model.linear.weight.data = torch.tensor([2], dtype=torch.float32).reshape([1,1])
    # model.linear.bias.data = torch.tensor([1], dtype=torch.float32).reshape([1,1])
    # torch.tensor().reshape([1,1])

    outputs = model(inputs)

    loss = criterion(outputs, labels)

    print(loss)
    print(torch.sum(pow(outputs-labels, 2))/11)
    print(outputs - labels)

    loss.backward()

    print(model.linear.weight)
    print(model.linear.bias)

    print(model.linear.weight.grad)
    print(model.linear.bias.grad)
    # print(2*torch.mean(outputs - labels))

    optimizer.step()

    print('After back propagation..............')

    print(model.linear.weight)
    print(model.linear.bias)

    print(model.linear.weight.grad)
    print(model.linear.bias.grad)


    if epoch % 50 == 0:
        print('epoch {}, loss{}'.format(epoch, loss.item()))