目录
GPU加速
把所有的设备cpu、cuda0、cuda1等,挣到一个设备上。
.cuda()方法会搬到GPU上
.to()方法返回一个新的refoungs,是否一致取决于哪种类型数据。对一个模块使用,则模块整体更新。如果对一个tensor使用,结果不一样。data2=data.cuda(),此时data2和data不一样,会产生两个tensor,一个在cpu上一个在gpu上。
import torch
import torch.nn as nn
import torch.nn.functional as F
import torch.optim as optim
from torchvision import datasets, transforms
batch_size=200
learning_rate=0.01
epochs=10
train_loader = torch.utils.data.DataLoader(
datasets.MNIST('mnist_data/', train=True, download=True,
transform=transforms.Compose([
transforms.ToTensor(),
transforms.Normalize((0.1307,), (0.3081,))
])),
batch_size=batch_size, shuffle=True)
test_loader = torch.utils.data.DataLoader(
datasets.MNIST('mnist_data/', train=False, transform=transforms.Compose([
transforms.ToTensor(),
transforms.Normalize((0.1307,), (0.3081,))
])),
batch_size=batch_size, shuffle=True)
class MLP(nn.Module):
def __init__(self):
super(MLP, self).__init__()
self.model = nn.Sequential(
nn.Linear(784, 200),
nn.LeakyReLU(inplace=True),
nn.Linear(200, 200),
nn.LeakyReLU(inplace=True),
nn.Linear(200, 10),
nn.LeakyReLU(inplace=True),
)
def forward(self, x):
x = self.model(x)
return x
device = torch.device('cuda:0')
net = MLP().to(device)
optimizer = optim.SGD(net.parameters(), lr=learning_rate)
criteon = nn.CrossEntropyLoss().to(device)
for epoch in range(epochs):
for batch_idx, (data, target) in enumerate(train_loader):
data = data.view(-1, 28*28)
data, target = data.to(device), target.cuda()
logits = net(data)
loss = criteon(logits, target)
optimizer.zero_grad()
loss.backward()
# print(w1.grad.norm(), w2.grad.norm())
optimizer.step()
if batch_idx % 100 == 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()))
test_loss = 0
correct = 0
for data, target in test_loader:
data = data.view(-1, 28 * 28)
data, target = data.to(device), target.cuda()
logits = net(data)
test_loss += criteon(logits, target).item()
pred = logits.data.max(1)[1]
correct += pred.eq(target.data).sum()
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)))
net=MLP().to(device),CrossEntropyLoss(),data,target=data.to(device),target.cuda(),使用.to()方法把网路搬到GPU上。
minist测试实战
不断train时,它可能会记住样本,此时效果会很好,但是一换其他样本就效果不佳。
对于每一张图片,对其可能是几的概率做softmax,并且做argmax(返回最大值所在索引),返回值即表明他是几。
logits=torch.rand(4,10)
pred=F.softmax(logits,dim=1)
print('pred shape:',pred.shape)
pred_lable=pred.argmax(dim=1)
print('pred_lable:',pred_lable)
print('logits.argmax:',logits.argmax(dim=1))
label=torch.tensor([9,3,2,4])
correct=torch.eq(pred_lable,label)
print('correct',correct)
accuracy=correct.sum().float().item()/4
print(accuracy)
我们得出[5,3,2,1],假设真实是[9,3,2,4],我们需要计算accuracy=2/4=50%。
计算正确的数量,使用eq()函数,如果元素相等返回1。
correct类型是unit8,我们需要使它变成float类型。
when to test
几个batch一个test
每个epoch一个test
vidsom可视化
visdom效率比tensorboard更高一点,因为后者会把数据写到文件中,导致监听文件占据大量空间。
visdom可以直接使用tensor。
使用visdom
1.install
首先进入anaconda prompt,在我们特定环境中输入pip install visdom
2.run server damon
开启监听进程,输入python -m visdom.server
visdom实际上是一个web服务器,开启后,程序向该服务器发送数据,服务器就会把数据渲染到网页上。运行前需要先开启该web服务器。
遇到问题
- pip unstall visdom
- 从官方网页上下载最新的代码,进入下载好的文件内,pip install -e .
- 退回去,开启监听进程
- 复制地址,打开浏览器
功能
lines:single trace
画一个曲线
首先创建一条直线,有x,y,且赋初始值,win可以理解为id,大窗口是一个environment。把窗口名字命名为train_loss,且只有一个点(0,0)
from visdom import Visdom
viz=Visdom()
viz.line([0.],[0.],win='train_loss',opts=dict(title='train loss'))
viz.line([loss.item()],[global_step],win='train_loss',update='append')
global_step是x作表,image可以直接接受tensor,对于直线类型,接受的是numpy数据,指定为append,添加到当前直线后面,如果不指定,会全部覆盖掉。
lines:single trace
多条曲线
两条直线,我们需要两个y。
from visdom import Visdom
viz=Visdom()
viz.line([[0.0,0.0]],[0.],win='test',opts=dict(title='test loss & acc',legend=['loss','acc']))
viz.line([[test_loss,correct/len(test_loader.dataset)]],
[global_step],win='test',update='append')
完整代码
import torch
import torch.nn as nn
import torch.nn.functional as F
import torch.optim as optim
from torchvision import datasets, transforms
from visdom import Visdom
batch_size=200
learning_rate=0.01
epochs=10
train_loader = torch.utils.data.DataLoader(
datasets.MNIST('mnist_data/', train=True, download=True,
transform=transforms.Compose([
transforms.ToTensor(),
# transforms.Normalize((0.1307,), (0.3081,))
])),
batch_size=batch_size, shuffle=True)
test_loader = torch.utils.data.DataLoader(
datasets.MNIST('mnist_data/', train=False, transform=transforms.Compose([
transforms.ToTensor(),
# transforms.Normalize((0.1307,), (0.3081,))
])),
batch_size=batch_size, shuffle=True)
class MLP(nn.Module):
def __init__(self):
super(MLP, self).__init__()
self.model = nn.Sequential(
nn.Linear(784, 200),
nn.LeakyReLU(inplace=True),
nn.Linear(200, 200),
nn.LeakyReLU(inplace=True),
nn.Linear(200, 10),
nn.LeakyReLU(inplace=True),
)
def forward(self, x):
x = self.model(x)
return x
device = torch.device('cuda:0')
net = MLP().to(device)
optimizer = optim.SGD(net.parameters(), lr=learning_rate)
criteon = nn.CrossEntropyLoss().to(device)
viz=Visdom()
viz.line([0.],[0.],win='train_loss',opts=dict(title='train loss'))
viz.line([[0.0,0.0]],[0.],win='test',opts=dict(title='test loss & acc',legend=['loss','acc']))
global_step=0
for epoch in range(epochs):
for batch_idx, (data, target) in enumerate(train_loader):
data = data.view(-1, 28*28)
data, target = data.to(device), target.cuda()
logits = net(data)
loss = criteon(logits, target)
optimizer.zero_grad()
loss.backward()
# print(w1.grad.norm(), w2.grad.norm())
optimizer.step()
global_step+=1
viz.line([loss.item()],[global_step],win='train_loss',update='append')
if batch_idx % 100 == 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()))
test_loss = 0
correct = 0
for data, target in test_loader:
data = data.view(-1, 28 * 28)
data, target = data.to(device), target.cuda()
logits = net(data)
test_loss += criteon(logits, target).item()
pred = logits.argmax(dim=1)
correct += pred.eq(target).sum().float().item()
viz.line([[test_loss,correct/len(test_loader.dataset)]],
[global_step],win='test',update='append')
viz.images(data.view(-1, 1, 28, 28), win='x')
viz.text(str(pred.detach().cpu().numpy()),win='pred',
opts=dict(title='pred'))
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)))