Pytorch 網(wǎng)絡(luò)結(jié)構(gòu)可視化

點(diǎn)擊上方“程序員大白”，選擇“星標(biāo)”公眾號(hào)
重磅干貨，第一時(shí)間送達(dá)

文自：深度學(xué)習(xí)這件小事

可以通過以下的命令進(jìn)行安裝

conda install pytorch-nightly -c pytorch
conda install graphviz
conda install torchvision
conda install tensorwatch

本教程基于以下的版本：

torchvision.__version__   '0.2.1'
torch.__version__         '1.2.0.dev20190610'
sys.version               '3.6.8 |Anaconda custom (64-bit)| (default, Dec 30 2018, 01:22:34)
[GCC 7.3.0]'

import sys
import torch
import tensorwatch as tw
import torchvision.models

alexnet_model = torchvision.models.alexnet()
tw.draw_model(alexnet_model, [1, 3, 224, 224])

載入alexnet，draw_model函數(shù)需要傳入三個(gè)參數(shù)，第一個(gè)為model，第二個(gè)參數(shù)為input_shape，第三個(gè)參數(shù)為orientation，可以選擇'LR'或者'TB'，分別代表左右布局與上下布局。

在notebook中，執(zhí)行完上面的代碼會(huì)顯示如下的圖，將網(wǎng)絡(luò)的結(jié)構(gòu)及各個(gè)層的name和shape進(jìn)行了可視化。

可以通過model_stats方法統(tǒng)計(jì)各層的參數(shù)情況。

tw.model_stats(alexnet_model, [1, 3, 224, 224])

[MAdd]: Dropout is not supported!
[Flops]: Dropout is not supported!
[Memory]: Dropout is not supported!
[MAdd]: Dropout is not supported!
[Flops]: Dropout is not supported!
[Memory]: Dropout is not supported!
[MAdd]: Dropout is not supported!
[Flops]: Dropout is not supported!
[Memory]: Dropout is not supported!
[MAdd]: Dropout is not supported!
[Flops]: Dropout is not supported!
[Memory]: Dropout is not supported!
[MAdd]: Dropout is not supported!
[Flops]: Dropout is not supported!
[Memory]: Dropout is not supported!
[MAdd]: Dropout is not supported!
[Flops]: Dropout is not supported!
[Memory]: Dropout is not supported!

alexnet_model.features

Sequential(
  (0): Conv2d(3, 64, kernel_size=(11, 11), stride=(4, 4), padding=(2, 2))
  (1): ReLU(inplace=True)
  (2): MaxPool2d(kernel_size=3, stride=2, padding=0, dilation=1, ceil_mode=False)
  (3): Conv2d(64, 192, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2))
  (4): ReLU(inplace=True)
  (5): MaxPool2d(kernel_size=3, stride=2, padding=0, dilation=1, ceil_mode=False)
  (6): Conv2d(192, 384, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  (7): ReLU(inplace=True)
  (8): Conv2d(384, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  (9): ReLU(inplace=True)
  (10): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  (11): ReLU(inplace=True)
  (12): MaxPool2d(kernel_size=3, stride=2, padding=0, dilation=1, ceil_mode=False)
)

alexnet_model.classifier

Sequential(
  (0): Dropout(p=0.5)
  (1): Linear(in_features=9216, out_features=4096, bias=True)
  (2): ReLU(inplace=True)
  (3): Dropout(p=0.5)
  (4): Linear(in_features=4096, out_features=4096, bias=True)
  (5): ReLU(inplace=True)
  (6): Linear(in_features=4096, out_features=1000, bias=True)
)

參考：

https://github.com/microsoft/tensorwatch

推薦閱讀
國(guó)產(chǎn)小眾瀏覽器因屏蔽視頻廣告，被索賠100萬（后續(xù)）
年輕人“不講武德”：因看黃片上癮，把網(wǎng)站和786名女主播起訴了
中國(guó)聯(lián)通官網(wǎng)被發(fā)現(xiàn)含木馬腳本，可向用戶推廣色情APP
張一鳴：每個(gè)逆襲的年輕人，都具備的底層能力

關(guān)于程序員大白

程序員大白是一群哈工大，東北大學(xué)，西湖大學(xué)和上海交通大學(xué)的碩士博士運(yùn)營(yíng)維護(hù)的號(hào)，大家樂于分享高質(zhì)量文章，喜歡總結(jié)知識(shí)，歡迎關(guān)注[程序員大白]，大家一起學(xué)習(xí)進(jìn)步！