pytorch 预测手写体数字_PyTorch学习笔记7 - 使用PyTorch完成手写体数字识别

本篇笔记是PyTorch学习的最后一篇，使用PyTorch搭建神经网络，完成手写体数字识别问题。

1. 导入数据并可视化

import torch

import torch.nn as nn

import torch.nn.functional as F

import torch.optim as optim

from torchvision import datasets, transforms

#定义batch size

batch_size = 64

#下载MNIST数据集

train_dataset = datasets.MNIST(root='./data/',

train=True,

transform=transforms.ToTensor(),

download=True)

test_dataset = datasets.MNIST(root='./data/',

train=False,

transform=transforms.ToTensor())

#将下载的MNIST数据导入到dataloader中

train_loader = torch.utils.data.DataLoader(dataset=train_dataset,

batch_size=batch_size,

shuffle=True)

test_loader = torch.utils.data.DataLoader(dataset=test_dataset,

batch_size=batch_size,

s