此处纯粹作为个人学习使用,原文连接:https://www.jianshu.com/p/dc24e54aec81

这篇文章是借鉴很多博文的,作为一个关于slim库的总结

导入slim模块

import tensorflow.contrib.slim as slim

定义slim的变量

#Model Variables
weights = slim.model_variable('weights', shape = [10, 10, 3, 3],
initializer = tf.truncated_normal_initializer(stddev=0.1)
regularizer = slim.l2_regularizer(0.05),
device='/CPU:0')
model_variables = slim.get_model_variables() #获取变量吗? # Regular variables
my_var = slim.variable('my_var", shape=[20, 1],
initializer = tf.zeros_initializer())
regular_variables_and_model_variables = slim.get_variables()

# 这里的model_variable是作为模型参数保存的,variable是局部变量,不会保存。

Slim中实现一个层

input = ...
net = slim.conv2d(input, 128, [3,3], scope='conv1_1') # 代码重用
net = slim.repeat(net, 3, slim.conv2d, 256, [3,3], scope='conv3')
net = slim.max_pool2d(net, [2, 2], scope='pool2') # 处理不同参数情况
x = slim.fully_connected(x, 32, scope='fc/fc_1')
x = slim.fully_connected(x, 64, scope ='fc/fc_2')
x = slim.fuly_connected(x, 128, scope = 'fc/fc_3')
# or
slim.stack(x, slim.fully_connected, [32, 64, 128], scope='fc') # 普通方法
x = slim.conv2d(x, 32, [3, 3], scope='core/core_1')
x = slim.conv2d(x, 32, [1, 1], scope='core/core_2')
x = slim.conv2d(x, 64, [3, 3], scope='core/core_3')
x = slim.conv2d(x, 64, [1, 1], scope='core/core_4') # 简便方法:
slim.stack(x, slim.conv2d, [(32, [3,3]), (32, [1,1]), (64, [3,3]), (64, [1,1]), scopre='core')

定义相同参数的简化

with slim.arg_scope([slim.conv2d],  padding='SAME',
weights_initializer=tf.truncated_normal_initializer(stddev=0.01),
weights_regularizer=slim.l2_regularizer(0.0005)):
net = slim.conv2d(inputs, 64, [11, 11], scope='conv1')
net = slim.conv2d(net, [11,11], padding=' VALID', scope='conv2')
net = slim.conv2d(net, 256, [11, 11], scope='conv3') # arg_scope的嵌套
with slim.arg_scope([slim.conv2d, slim.fully_connected],
activation_fn=tf.nn.rely,
weights_initializer=tf.truncated_normal_initialier(stddev=0.01),
weights_regularizer=slim.l2_regularizer(0.0005)):
with slim.arg_scope([slim.conv2d], stride=1, padding='SAME'):
net = slim.conv2d(inputs, 64, [11, 11], 4, padding='VALID', scope='conv1')
net = slim.conv2d(net, 256, [5, 5],
weights_initializer=tf.truncated_normal_initializer(stddev=0.03),
scope='conv2')
net = slim.fully_connected(net, 1000, activation_fn=None, scope='fc')

训练模型

loss = slim.losses.softmax_cross_entropy(predictions, labels)
# 自定义loss模型
# define the loss functions and get the total loss.
classification_loss = slim.losses.softmax_cross_entropy(scene_predictions, scene_labels)
sum_of_squares_loss = slim.losses.sum_of_squares(depth_predictions, depth_labels)
pose_loss = MyCustomLossFunction(pose_predictions, pose_labels)
slim.losses.add_loss(pose_loss) # Letting TF-Slim know about the additional loss. # The following two ways to compute the total loss are equivalent:
regularization_loss = tf.add_n(slim.losses.get_regularization_losses())
total_loss1 = classification_loss + sum_of_squares_loss + poses_loss + regularization_loss

# slim读取保存模型的方法

# Create some variables.
v1 = slim.variable(name='v1', ...)
v2 = slim.variable(name=''nested/v2', ...)
... # Get list of variables to restore (which contains only 'v2')
variables_to_restore = slim.get_variables_by_name("v2") # Create the saver which will be used to restore the varialbes.
restorer = tf.train.Saver(variables_to_restore) with tf.Session() as sess:
# Restore variables from disk.
restores.restore(sess, "/tmp/model.ckpt")
print("Model restored.") # 为模型添加变量前缀
# 假设我们定义的网络变量是conv1/weights, 而从VGG记载的变量名为#vgg16/conv1/weights, 正常load肯定会报错
def name_in_checkpoint(var):
return 'vgg16/' + var.op.name variables_to_restore = slim.get_model_variables()
variables_to_restore = {name_in_checkpoint(var):var for var in variables_to_restore}
restorer = tf.train.Saver(variables_to_restore) with tf.Session() as sess:
# Restore variables from disk.
restorer.restore(sess, "/tmp/model.ckpt")

训练模型

在该例中, slim.learning.train根据train_op计算损失、应用梯度step. logdir指定checkpoints和event文件的存储路径。我们可以限制梯度step到任何数值。这里我们采用1000步。最后, save_summaries_secs=300表示每5分钟计算一次summaries, save_interval_secs=600表示每10分钟保存一次模型的checkpoint

g = tf.Graph()

# Create the model and specify the losses...
... total_loss = slim.losses.get_total_loss()
optimizer = tf.train.GradientDescentOptimizer(learning_rate) # create_train_op ensures that each time we ask for the loss, the update_ops
# are run and the gradients being computed are applied too.
train_op = slim.learning.create_train_op(total_loss, optimizer)
logdir = ... # Where checkpoints are stored. slim.learning.train(
train_op,
logdir,
number_of_steps=1000,
save_summaries_secs=300,
save_interval_secs=600)

Fine-Tuning a Model on a different task

假设我们有一个已经预训练好的VGG16的模型。这个模型是在拥有1000分类的ImageNet数据集上进行训练的。但是,现在我们想把它应用只具有20个分类的Pascal VOC数据集上。为了能这样做,我们可以通过利用除最后一些全连接层的其它预训练模型来初始化新模型的达到目的:

# Load the Pascal VOC data
image, label = MyPascalVocDataLoader(...)
images, labels = tf.train.batch([image, label], batch_size = 32) # Create the model
predictions = vgg.vgg_16(images)
train_op = slim.learning.create_train_op(...) # Specify where the Model, trained on ImageNet, was saved.
model_path = '/path/to/pre_trained_on_imagenet.checkpoint'
metric_ops.py
# Specify where the new model will live:
log_dir = from_checkpoint_'/path/to/my_pascal_model_dir/' # Restore only the convolutional layers:
variables_to_restore = slim.get_variables_to_restore(exclude=['fc6', 'fc7', 'fc8'])
init_fn = assign_from_checkpoint_fn(model_path, variables_to_restore) # Start training.
slim.learning.train(train_op, log_dir, init_fn=init_fn)

evaluation loop

import tensorflow as tf

slim = tf.contrib.slim

# Load the data
images, labels = load_data(...) # Define the network
predictions = MyModel(images) # Choose the metrics to compute:
names_to_values, names_to_updates = slim.metrics.aggregate_metric_map({
'accuracy': slim.metrics.accuracy(predictions, labels),
'precision': slim.metrics.precision(predictions, labels),
'recall': slim.metrics.recall(mean_relative_errors, 0.3),
}) # Create the summary ops such that they also print out to std output:
summary_ops = []
for metric_name, metric_value in names_to_values.iteritems():
op = tf.summary.scalar(metric_name, metric_value)
op = tf.Print(op, [metric_value], metric_name)
summary_ops.append(op) num_examples = 10000
batch_size = 32
num_batches = math.ceil(num_examples / float(batch_size)) # Setup the global step.
slim.get_or_create_global_step() output_dir = ... # Where the summaries are stored.
eval_interval_secs = ... # How often to run the evaluation.
slim.evaluation.evaluation_loop(
'local',
checkpoint_dir,
log_dir,
num_evals=num_batches,
eval_op=names_to_updates.values(),
summary_op=tf.summary.merge(summary_ops),
eval_interval_secs=eval_interval_secs)

tensorflow slim代码使用的更多相关文章

  1. 使用笔记:TF辅助工具--tensorflow slim(TF-Slim)

    如果抛开Keras,TensorLayer,tfLearn,tensroflow 能否写出简介的代码? 可以!slim这个模块是在16年新推出的,其主要目的是来做所谓的“代码瘦身” 一.简介 slim ...

  2. 解决TensorFlow最新代码编译错误问题

    老是有个习惯,看到开源代码更新了,总是想更新到最新版,如果置之不理的话,就感觉自己懒惰了或有的不负责任了,这个也可能是一种形式的强迫症吧: 前几天晚上git pull TensorFlow,完事后也没 ...

  3. tensorflow没有代码提示的问题

    在tensorflow包下的__init__.py文件中定义了一个contrib变量表示tensorflow.contrib包下的内容,但是tensorflow.contrib这个包是懒加载的,也就是 ...

  4. google tensorflow bert代码分析

    参考网上博客阅读了bert的代码,记个笔记.代码是 bert_modeling.py 参考的博客地址: https://blog.csdn.net/weixin_39470744/article/de ...

  5. tensorflow训练代码

    from tensorflow.examples.tutorials.mnist import input_data import tensorflow as tf mnist = input_dat ...

  6. tensorflow TensorArray 代码例子

    import tensorflow as tf import numpy as np B=3 D=4 T=5 tf.reset_default_graph() xs=tf.placeholder(sh ...

  7. Tensorflow模型代码调试问题

    背景: 不知道大家有没有这样的烦恼:在使用Tensorflow搭建好模型调试的过程中,经常会碰到一些问题,当时花了不少时间把这个问题解决了,一段时间后,又出现了同样的问题,却怎么也不记得之前是怎么解决 ...

  8. TensorFlow Slim 的常用操作

    https://blog.csdn.net/mzpmzk/article/details/81706379

  9. tensorflow slim训练以及到安卓部署教程

    https://blog.csdn.net/chenyuping333/article/details/81537551 https://blog.csdn.net/u012328159/articl ...

随机推荐

  1. 常用注解解析(因为不太明白@component和@configuration写了)

    1.@controller 控制器(注入服务) 用于标注控制层,相当于struts中的action层 2.@service 服务(注入dao) 用于标注服务层,主要用来进行业务的逻辑处理 3.@rep ...

  2. center----Iframe 用法的详细讲解

    把iframe解释成“浏览器中的浏览器“很是恰当 <iframe frameborder=0 width=170 height=100 marginheight=0 marginwidth=0 ...

  3. 开发一个这样的 APP 要多长时间?

    作者:蒋国刚 www.cnblogs.com/guogangj/p/4676836.html 这是一个“如有雷同,纯属巧合”的故事,外加一些废话,大家请勿对号入座.开始了…… 我有些尴尬地拿着水杯,正 ...

  4. json解决ajax跨域的原理

    jsonp只能解决GET类型的ajax请求跨域问题 jsonp请求不是ajax请求,而是一般的get请求 基本原理 浏览器端: 动态生成<script>来请求后台接口(src就是接口的ur ...

  5. javascript中的toString()

    基本介绍 javascript中的toString方法是我们在写前端时经常要用的一个函数,也就是将我们的变量转换成字符串的方法. javascript中各种类型的toString方法 javascri ...

  6. 《Netty Redis Zookeeper 高并发实战》 勘误

    <Netty Redis Zookeeper 高并发实战> 勘误与申明 疯狂创客圈 Java 高并发[ 亿级流量聊天室实战]实战系列 [博客园总入口 ] 勘误一 文字问题: Page1 J ...

  7. 易优CMS:channel的基础用法

    [基础用法] 名称:channel 功能:易优常用标记,可以循环嵌套标签.通常用于网站导航以获取站点栏目信息,方便网站会员分类浏览整站信息 语法: {eyou:channel type='top' r ...

  8. 弹指间,网页灰飞烟灭——Google灭霸彩蛋实现

    不知道大家有没有看这段时间最火的一部电影<复仇者联盟4:终局之战>,作为漫威迷的我还没看,为什么呢?因为太贵了,刚上映的那周,一张IMAX厅的票价已经达到了299的天价,作为搬砖民工是舍不 ...

  9. Grafana+Prometheus 监控 MySQL

    转自:Grafana+Prometheus 监控 MySQL 架构图 环境 IP 环境 需装软件 192.168.0.237 mysql-5.7.20 node_exporter-0.15.2.lin ...

  10. Format a Business Object Caption 设置业务对象标题的格式

    In this lesson, you will learn how to format the caption of a detail form that displays a business o ...