头歌-Autograd

第1关：求导原理

import torch

Tensor = torch.Tensor


def task(x: Tensor) -> Tensor:
    

    y = 2*x + 6

    z = y * y * 3


    out = torch.mean(3*z)
    
    ########## Begin ##########
    x.requires_grad = True
    out.backward()
    ########## End ##########

    result = x.grad

    return result

第2关：静态图动态图设计

import torch


def task():
    w = torch.tensor([1.], requires_grad=True)

    x = torch.tensor([2.], requires_grad=True)

    ########## Begin ##########
    a = w+x 
    b = w+0
    y = a+b
    ########## End ##########

    y.backward()

    print(y)
    print(w.is_leaf, x.is_leaf, a.is_leaf, b.is_leaf, y.is_leaf)

第3关：正向反向传播

import torch
from torch import nn, optim
from torch.autograd import Variable
import numpy as np

x_train = np.array([[3.3], [4.4], [5.5], [6.71], [6.93], [4.168],
                    [9.779], [6.182], [7.59], [2.167], [7.042],
                    [10.791], [5.313], [7.997], [3.1]], dtype=np.float32)

y_train = np.array([[1.7], [2.76], [2.09], [3.19], [1.694], [1.573],
                    [3.366], [2.596], [2.53], [1.221], [2.827],
                    [3.465], [1.65], [2.904], [1.3]], dtype=np.float32)


x_train = torch.from_numpy(x_train)

y_train = torch.from_numpy(y_train)


# Linear Regression Model
class LinearRegression(nn.Module):
    def __init__(self):
        super(LinearRegression, self).__init__()
        self.linear = nn.Linear(1, 1)  # input and output is 1 dimension

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


model = LinearRegression()
# 定义loss和优化函数
criterion = nn.MSELoss()
optimizer = optim.SGD(model.parameters(), lr=1e-4)

def train():
    # 开始训练
    num_epochs = 2000
    for epoch in range(num_epochs):
        inputs = Variable(x_train)
        target = Variable(y_train)

        # 正向传播
        ########## Begin ##########
        out = model(inputs)
        ########## End ##########
        
        loss = criterion(out, target)
        # 反向传播
        optimizer.zero_grad()

        ########## Begin ##########
        loss.backward()
        ########## End ##########

        optimizer.step()

        if (epoch+1) % 20 == 0:
            print('Epoch[{}/{}], loss: {:.6f}'
                .format(epoch+1, num_epochs, loss.data))
    
    return loss.data

第4关：损失函数

import torch


def task(y1, y2):
    '''
    任务要求：使用pytorch中的均方误差（MSE）求y1与y2之间的均方误差
    '''
    ########## Begin ##########
    loss = torch.nn.MSELoss()
    ########## End ##########
    output = loss(y1, y2)
    return output