深度学习与PyTorch笔记21

Visdom可视化

TensorboardX

使用pip install tensorboardX命令来安装。
tensorboardX在pytorch中怎么使用：
它有一个SummaryWriter，新建一个SummaryWriter实例，把要监听的数据起一个名字，(‘data/scalar’（名字）,dummy_s1[0]（数据）,n_iter（x坐标，代表的是哪一个epoch）)。
tensorboard本质上抽取的是一个numpy的数据，如果要和tensor做一个match的话，必须先把tensor转换到cpu上面（.cpu()），然后再转化成numpy数据（data.numpy()）,才能够赋值给tensorboard。

from tensorboardX import SummaryWriter
writer=SummaryWriter()
writer.add_scalar('data/scalar',dummy_s1[0],n_iter)
writer.add_scalar('data/scalar_group',{'xsinx':n_iter*np.sin(n_iter),
                                       'xcosx':n_iter*np.cos(n_iter),
                                       'arctanx':np.arctan(n_iter),},n_iter)
writer.add_image('Image',x,n_iter)
writer.add_text('Text','text logged at step:'+str(n_iter),n_iter)

for name,param in resnet18.named_parampython -m visdom.servereters():
    writer.add_histogram(name,param.clone().cpu().data.numpy(),n_iter)
writer.close()

Visdom from Facebook

使用pip install visdom命令来安装。安装完成后开启监听的进程，确保程序运行前开启visdom，使用python -m visdom.server来开启。
windows开启时容易出现错误：

解决这个错误的方法就是重新安装visdom。首先pip uninstall visdom卸载之前的visdom，然后从官方网页上下载源代码，解压，然后进入到那个目录下面去cd 目录名，再进入cd visdom-master，然后pip install -e.。安装完成之后退回cd ../..，再python -m visdom.server。出现监听命令然后复制地址打开浏览器就行了。


地址：http://localhost:8097

lines:single trace

展示一个最常见的功能，画一个曲线。
首先创建一个Visdom实例（viz=Visdom()），再创建一条直线，然后把当前的数据添加到直线上面去，创建一条直线的指令（viz.line([0.](y初始值),[0.](x初始值),win='train_loss'（唯一的标志符，可以理解为ID，创建一个小窗口）,opts=dict(title='train loss'（命名）))）和添加数据指令（viz.line([loss.item()](对于非image数据，传入的还是numpy),[global_step](x坐标),win='train_loss',update='append'(指定update操作，添加到当前直线后面，不指定的话会全部覆盖掉。))）一定要配合起来使用。

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')

我把loss.item()改为1，把global_step改为1，出现此图

lines:multi-traces

多条曲线
创建[y1,y2]，legend是一个标识符，代表y1，y2的label。

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,coorrect/len(test_loader.dataset)]],
         [global_step],win='test',update='append')

自己添加一些数据，viz.line([[1.0,0.5]],[10],win='test',update='append')

visual X

在visdom中传入照片可以直接输入tensor。

from visdom import Visdom
viz=Visdom()
viz.images(data.view(-1,1,28,28),win='x')
viz.text(str(pred.detach().cpu().numpy()),win='pred',
         opts=dict(title='pred'))

总结

下面复制完整代码及结果图

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('../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('../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).float().sum().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)))