【Pytroch】Ubuntu下通过graphviz实现网络可视化方法

安装graphviz

pip install graphviz
sudo apt install graphviz
dot -version

若安装成功则会显示：

Using device: dot:dot:core
The plugin configuration file:
/usr/lib/x86_64-linux-gnu/graphviz/config6a
was successfully loaded.
render : cairo dot dot_json fig gd json json0 map mp pic pov ps svg tk vml vrml xdot xdot_json
layout : circo dot fdp neato nop nop1 nop2 osage patchwork sfdp twopi
textlayout : textlayout
device : canon cmap cmapx cmapx_np dot dot_json eps fig gd gd2 gif gv imap imap_np ismap jpe jpeg jpg json json0 mp pdf pic plain plain-ext png pov ps ps2 svg svgz tk vml vmlz vrml wbmp x11 xdot xdot1.2 xdot1.4 xdot_json xlib
loadimage : (lib) eps gd gd2 gif jpe jpeg jpg png ps svg xbm

添加环境变量

export PATH=$PATH:/usr/lib/x86_64-linux-gnu/graphviz

生成网络图片

运行下面代码

import torch
import torch.nn as nn
from torch.autograd import Variable
from torchviz import make_dot

class LeNet(nn.Module):
    def __init__(self):
        super(LeNet, self).__init__()
        self.conv1 = nn.Sequential(     #input_size=(1*28*28)
            nn.Conv2d(1, 6, 5, 1, 2),
            nn.ReLU(),      #(6*28*28)
            nn.MaxPool2d(kernel_size=2, stride=2),  #output_size=(6*14*14)
        )
        self.conv2 = nn.Sequential(
            nn.Conv2d(6, 16, 5),
            nn.ReLU(),      #(16*10*10)
            nn.MaxPool2d(2, 2)  #output_size=(16*5*5)
        )
        self.fc1 = nn.Sequential(
            nn.Linear(16 * 5 * 5, 120),
            nn.ReLU()
        )
        self.fc2 = nn.Sequential(
            nn.Linear(120, 84),
            nn.ReLU()
        )
        self.fc3 = nn.Linear(84, 10)

    # 定义前向传播过程，输入为x
    def forward(self, x):
        x = self.conv1(x)
        x = self.conv2(x)
        # nn.Linear()的输入输出都是维度为一的值，所以要把多维度的tensor展平成一维
        x = x.view(x.size()[0], -1)
        x = self.fc1(x)
        x = self.fc2(x)
        x = self.fc3(x)
        return x

dummy_input = Variable(torch.rand(13, 1, 28, 28)) #假设输入13张1*28*28的图片
model = LeNet()

y = model(dummy_input)
g = make_dot(y)
g.render('LeNet_model', view=False)

会生成一个LeNet_model和LeNet_model.pdf，LeNet_model即为网络图：

参考：

pytorch 网络结构可视化方法汇总（三种实现方法详解）
ExecutableNotFound: failed to execute [‘dot’, ‘-Tsvg’], make sure the Graphviz executables are on yo