torch 学习, 搭建网络- 手写数字识别(3)

说明

继续更新pytorch的学习

笔记

在这里插入图片描述

官方数据及处理方式

# 定义数据处理方式
train_transform = transforms.Compose([
    transforms.ToTensor()
]) 
# 构建数据dataloader

train_data = MNIST('./data', train=True, transform=train_transform, download=True)
valid_data = MNIST('./data', train=False, transform=train_transform, download=True)

train_iter = DataLoader(train_data, batch_size=128, shuffle=True)
valid_iter = DataLoader(valid_data, batch_size=128, shuffle=False)

模型的搭建

# 构建网络模型
class MNIST_Net(nn.Module):
    
    def __init__(self):
        super().__init__()
        self.f1 = nn.Linear(784, 128)
        self.f2 = nn.Linear(128, 256)
        self.f3 = nn.Linear(256, 10)
        
    def forward(self, x):
        x = x.view(-1, 784)
        x = F.relu(self.f1(x))
        x = F.relu(self.f2(x))
        x = self.f3(x)
        return x

模式识别

利用model.train()和model.eval()
** with torch.no_grad()** 不需要反向传播

完整代码

import torch.nn as nn
from torchvision.datasets import MNIST
import torch.nn.functional as F
import numpy as np
from torch.utils.data import DataLoader
from torchvision import transforms
import torch.optim as optim
import torch

# 定义数据处理方式
train_transform = transforms.Compose([
    transforms.ToTensor()
]) 
# 构建数据dataloader

train_data = MNIST('./data', train=True, transform=train_transform, download=True)
valid_data = MNIST('./data', train=False, transform=train_transform, download=True)

train_iter = DataLoader(train_data, batch_size=128, shuffle=True)
valid_iter = DataLoader(valid_data, batch_size=128, shuffle=False)

# 构建网络模型
class MNIST_Net(nn.Module):
    
    def __init__(self):
        super().__init__()
        self.f1 = nn.Linear(784, 128)
        self.f2 = nn.Linear(128, 256)
        self.f3 = nn.Linear(256, 10)
        
    def forward(self, x):
        x = x.view(-1, 784)
        x = F.relu(self.f1(x))
        x = F.relu(self.f2(x))
        x = self.f3(x)
        return x

def get_model():
    model = MNIST_Net()
    return model, optim.SGD(model.parameters(), lr=0.001)

def fit(epochs, model, loss_func, opt, train_iter, valid_iter):
    for epoch in range(epochs):
        # 
        model.train()
        for xb, yb in train_iter:
            loss_batch(model, loss_func, xb, yb, opt)
        model.eval()
        with torch.no_grad():
            losses, nums = zip(
                *[loss_batch(model, loss_func, xb, yb) for xb, yb in valid_iter]
            )
            val_loss = np.sum(np.multiply(losses, nums)) / np.sum(nums)
            print('当前step:' + str(epoch), "验证集损失:" + str(val_loss))
def loss_batch(model, loss_func, xb, yb, opt = None):
    loss = loss_func(model(xb), yb)
    if opt is not None:
        loss.backward()
        opt.step()
        opt.zero_grad()
    return loss.item(), len(xb)

model, opt = get_model()
fit(30, model, F.cross_entropy, opt, train_iter, valid_iter)

评论
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
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

打赏作者

东哥aigc

你的鼓励将是我创作的最大动力

¥1 ¥2 ¥4 ¥6 ¥10 ¥20
扫码支付:¥1
获取中
扫码支付

您的余额不足,请更换扫码支付或充值

打赏作者

实付
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