谷歌的inception模型是怎么训练的

Posted 2023-03-27

参考技术A

Inception (GoogLeNet)是Google 2014年发布的Deep Convolutional Neural Network，其它几个流行的CNN网络还有QuocNet、AlexNet、BN-Inception-v2、VGG、ResNet等等。

Inception V3模型源码定义：tensorflow/contrib/slim/python/slim/nets/inception_v3.py

训练大的网络模型很耗资源，幸亏TensorFlow支持分布式：

把计算任务Distribution到服务器集群

把计算任务Distribution到多个GPU

TensorBoard可视化Inception V3模型

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

   

import tensorflow as tf

import os

import tarfile

import requests

inception_pretrain_model_url = 'http://download.tensorflow.org/models/image/imagenet/inception-2015-12-05.tgz'

# 下载inception pretrain模型

inception_pretrain_model_dir = "inception_pretrain"

if not os.path.exists(inception_pretrain_model_dir):

os.makedirs(inception_pretrain_model_dir)

filename = inception_pretrain_model_url.split('/')[-1]

filepath = os.path.join(inception_pretrain_model_dir, filename)

if not os.path.exists(filepath):

print("开始下载: ", filename)

r = requests.get(inception_pretrain_model_url, stream=True)

with open(filepath, 'wb') as f:

for chunk in r.iter_content(chunk_size=1024):

if chunk:

f.write(chunk)

print("下载完成, 开始解压: ", filename)

tarfile.open(filepath, 'r:gz').extractall(inception_pretrain_model_dir)

# TensorBoard log目录

log_dir = 'inception_log'

if not os.path.exists(log_dir):

os.makedirs(log_dir)

# 加载inception graph

inception_graph_def_file = os.path.join(inception_pretrain_model_dir, 'classify_image_graph_def.pb')

with tf.Session() as sess:

with tf.gfile.FastGFile(inception_graph_def_file, 'rb') as f:

graph_def = tf.GraphDef()

graph_def.ParseFromString(f.read())

tf.import_graph_def(graph_def, name='')

writer = tf.train.SummaryWriter(log_dir, sess.graph)

writer.close()

   

使用TensorBoard查看Graph：

1

   

$ tensorboard --logdir inception_log

   

浏览器访问：http://127.0.0.1:6006

如要转载，请保持本文完整，并注明作者@斗大的熊猫和本文原始地址： http://blog.topspeedsnail.com/archives/10919