其他
【他山之石】PyTorch 手把手搭建神经网络 (MNIST)
“他山之石,可以攻玉”,站在巨人的肩膀才能看得更高,走得更远。在科研的道路上,更需借助东风才能更快前行。为此,我们特别搜集整理了一些实用的代码链接,数据集,软件,编程技巧等,开辟“他山之石”专栏,助你乘风破浪,一路奋勇向前,敬请关注。
地址:https://www.zhihu.com/people/sun-xu-sen-30
01
(<PIL.Image.Image image mode=L size=28x28 at 0x175EF44F160>, 5)
02
input:torch.Size([64, 1, 28, 28]) after conv1: torch.Size([64, 32, 26, 26]) after conv2: torch.Size([64, 64, 24, 24]) after max_pool2d: torch.Size([64, 64, 12, 12]) after flatten: torch.Size([64, 9216]) after fc1: torch.Size([64, 128]) after fc2: torch.Size([64, 10]) after log_softmax: torch.Size([64, 10])
from __future__ import print_function
import argparse
import torch
import torch.nn as nn
import torch.nn.functional as F
import torch.optim as optim
from torchvision import datasets, transforms
from torch.optim.lr_scheduler import StepLR
class Net(nn.Module):
def __init__(self):
super(Net, self).__init__()
self.conv1 = nn.Conv2d(1, 32, 3, 1)
self.conv2 = nn.Conv2d(32, 64, 3, 1)
self.dropout1 = nn.Dropout(0.25)
self.dropout2 = nn.Dropout(0.5)
self.fc1 = nn.Linear(9216, 128)
self.fc2 = nn.Linear(128, 10)
def forward(self, x):
x = self.conv1(x)
# print("after conv1: {}".format(x.shape))
x = F.relu(x)
x = self.conv2(x)
# print("after conv2: {}".format(x.shape))
x = F.relu(x)
x = F.max_pool2d(x, 2)
# print("after max_pool2d: {}".format(x.shape))
x = self.dropout1(x)
x = torch.flatten(x, 1)
# print("after flatten: {}".format(x.shape))
x = self.fc1(x)
# print("after fc1: {}".format(x.shape))
x = F.relu(x)
x = self.dropout2(x)
x = self.fc2(x)
# print("after fc2: {}".format(x.shape))
output = F.log_softmax(x, dim=1)
# print("after log_softmax: {}".format(output.shape))
return output
transform = transforms.Compose([
transforms.ToTensor(),
transforms.Normalize((0.1307,), (0.3081,))
])
taet1 = datasets.MNIST('../data', train=True, download=True,
transform=transform)
train_loader = torch.utils.data.DataLoader(dataset1, **train_kwargs)
# 优化器
optimizer = optim.Adadelta(model.parameters(), lr=args.lr)
# 调整学习率
scheduler = StepLR(optimizer, step_size=1, gamma=args.gamma)
03
def train(args, model, device, train_loader, optimizer, epoch):
# 设置为train模式
model.train()
# 分批次读入数据
for batch_idx, (data, target) in enumerate(train_loader):
data, target = data.to(device), target.to(device)
# 优化器梯度置零
optimizer.zero_grad()
# 数据通过网络
output = model(data)
# 计算损失
loss = F.nll_loss(output, target)
# 反向传播
loss.backward()
# 权重更新
optimizer.step()
if batch_idx % args.log_interval == 0:
print('Train Epoch: {} [{}/{} ({:.0f}%)]\tLoss: {:.6f}'.format(
epoch, batch_idx * len(data), len(train_loader.dataset),
100. * batch_idx / len(train_loader), loss.item()))
if args.dry_run:
break
04
def test(model, device, test_loader):
# 设置为test模式
model.eval()
test_loss = 0
correct = 0
# 不进行计算图的构建,即没有grad_fn属性
with torch.no_grad():
for data, target in test_loader:
data, target = data.to(device), target.to(device)
output = model(data)
test_loss += F.nll_loss(output, target, reduction='sum').item() # sum up batch loss
pred = output.argmax(dim=1, keepdim=True) # get the index of the max log-probability
correct += pred.eq(target.view_as(pred)).sum().item()
test_loss /= len(test_loader.dataset)
print('\nTest set: Average loss: {:.4f}, Accuracy: {}/{} ({:.0f}%)\n'.format(
test_loss, correct, len(test_loader.dataset),
100. * correct / len(test_loader.dataset)))
05
def main():
# Training settings
# 声明一个parser
parser = argparse.ArgumentParser(description='PyTorch MNIST Example')
# 添加参数
parser.add_argument('--batch-size', type=int, default=64, metavar='N',
help='input batch size for training (default: 64)')
parser.add_argument('--test-batch-size', type=int, default=1000, metavar='N',
help='input batch size for testing (default: 1000)')
parser.add_argument('--epochs', type=int, default=14, metavar='N',
help='number of epochs to train (default: 14)')
parser.add_argument('--lr', type=float, default=1.0, metavar='LR',
help='learning rate (default: 1.0)')
parser.add_argument('--gamma', type=float, default=0.7, metavar='M',
help='Learning rate step gamma (default: 0.7)')
parser.add_argument('--no-cuda', action='store_true', default=False,
help='disables CUDA training')
parser.add_argument('--dry-run', action='store_true', default=False,
help='quickly check a single pass')
parser.add_argument('--seed', type=int, default=1, metavar='S',
help='random seed (default: 1)')
parser.add_argument('--log-interval', type=int, default=10, metavar='N',
help='how many batches to wait before logging training status')
parser.add_argument('--save-model', action='store_true', default=False,
help='For Saving the current Model')
# 读取命令行参数
args = parser.parse_args()
use_cuda = not args.no_cuda and torch.cuda.is_available()
torch.manual_seed(args.seed)
device = torch.device("cuda" if use_cuda else "cpu")
train_kwargs = {'batch_size': args.batch_size}
test_kwargs = {'batch_size': args.test_batch_size}
if use_cuda:
cuda_kwargs = {'num_workers': 1,
'pin_memory': True,
'shuffle': True}
train_kwargs.update(cuda_kwargs)
test_kwargs.update(cuda_kwargs)
transform = transforms.Compose([
transforms.ToTensor(),
transforms.Normalize((0.1307,), (0.3081,))
])
dataset1 = datasets.MNIST('../data', train=True, download=True,
transform=transform)
dataset2 = datasets.MNIST('../data', train=False,
transform=transform)
train_loader = torch.utils.data.DataLoader(dataset1, **train_kwargs)
test_loader = torch.utils.data.DataLoader(dataset2, **test_kwargs)
model = Net().to(device)
optimizer = optim.Adadelta(model.parameters(), lr=args.lr)
scheduler = StepLR(optimizer, step_size=1, gamma=args.gamma)
for epoch in range(1, args.epochs + 1):
train(args, model, device, train_loader, optimizer, epoch)
test(model, device, test_loader)
scheduler.step()
if args.save_model:
torch.save(model.state_dict(), "mnist_cnn.pt")
06
07
本文目的在于学术交流,并不代表本公众号赞同其观点或对其内容真实性负责,版权归原作者所有,如有侵权请告知删除。
“他山之石”历史文章
autograd源码剖析
怎样才能让你的模型更加高效运行?
来自日本程序员的纯C++深度学习库tiny-dnn
MMTracking: OpenMMLab 一体化视频目标感知平台
深度学习和机器视觉top组都在研究什么
pytorch常见的坑汇总
pytorch 中张量基本操作
pytorch计算模型FLOPs和Params
保姆级教程:个人深度学习工作站配置指南
整理 Deep Learning 调参 tricks
Tensorflow模型保存方式大汇总
利用Tensorflow构建CNN图像多分类模型及图像参数、数据维度变化情况实例分析
pytorch中optimizer对loss的影响
使用PyTorch 1.6 for Android
更多他山之石专栏文章,
请点击文章底部“阅读原文”查看
分享、点赞、在看,给个三连击呗!