<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Tensorflow on 111qqz's blog</title><link>https://111qqz.com/en/tags/tensorflow/</link><description>Recent content in Tensorflow on 111qqz's blog</description><generator>Hugo -- gohugo.io</generator><language>en</language><copyright>© 2015-2026 111qqz</copyright><lastBuildDate>Thu, 24 Aug 2017 09:12:58 +0000</lastBuildDate><atom:link href="https://111qqz.com/en/tags/tensorflow/index.xml" rel="self" type="application/rss+xml"/><item><title>tensorflow input pipline 学习笔记</title><link>https://111qqz.com/en/post/acm-icpc/2017/2017-08-24-tensorflow-input-pipline-notes/</link><pubDate>Thu, 24 Aug 2017 09:12:58 +0000</pubDate><guid>https://111qqz.com/en/post/acm-icpc/2017/2017-08-24-tensorflow-input-pipline-notes/</guid><description>&lt;p&gt;参考资料：&lt;/p&gt;
&lt;p&gt;&lt;a href="https://www.tensorflow.org/api_guides/python/reading_data" target="_blank" rel="noreferrer"&gt;tf_doc_Reading data&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;&lt;a href="https://ischlag.github.io/2016/06/19/tensorflow-input-pipeline-example/" target="_blank" rel="noreferrer"&gt;TENSORFLOW INPUT PIPELINE EXAMPLE&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;&lt;a href="http://shartoo.github.io/tensorflow-inputpipeline/" target="_blank" rel="noreferrer"&gt;tensorflow：理解tensorflow中的输入管道&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;第二个参考资料是第一个的翻译版本，翻译的水平一般，建议看原文，不是很长。&lt;/p&gt;</description></item><item><title>tensorflow 合并模型</title><link>https://111qqz.com/en/post/acm-icpc/2017/2017-08-21-tensorflow-model-merging/</link><pubDate>Mon, 21 Aug 2017 06:56:22 +0000</pubDate><guid>https://111qqz.com/en/post/acm-icpc/2017/2017-08-21-tensorflow-model-merging/</guid><description>&lt;p&gt;在这里存个备份，还有些问题没有解决。&lt;/p&gt;
&lt;p&gt;raise ValueError(&amp;ldquo;GraphDef cannot be larger than 2GB.&amp;rdquo;)&lt;/p&gt;
&lt;p&gt;记录一些思路好了。现在是没有生成.meta文件，爆掉应该是因为所有的变量都加载到了默认图里。&lt;/p&gt;</description></item><item><title>tensorflow checkpoint 学习笔记</title><link>https://111qqz.com/en/post/acm-icpc/2017/2017-08-21-tensorflow-checkpoint-notes/</link><pubDate>Mon, 21 Aug 2017 02:03:45 +0000</pubDate><guid>https://111qqz.com/en/post/acm-icpc/2017/2017-08-21-tensorflow-checkpoint-notes/</guid><description>&lt;p&gt;参考资料：&lt;/p&gt;
&lt;p&gt;&lt;a href="https://stackoverflow.com/questions/36195454/what-is-the-tensorflow-checkpoint-meta-file" target="_blank" rel="noreferrer"&gt;What is the TensorFlow checkpoint meta file?&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;&lt;a href="https://stackoverflow.com/questions/35733917/tensorflow-restoring-variables-from-from-multiple-checkpoints" target="_blank" rel="noreferrer"&gt;TensorFlow: Restoring variables from from multiple checkpoints&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;合并模型的时候发现.meta一直在累加，而其他数据文件没有改变。因此来探究一下checkpoint的几个文件的含义。&lt;/p&gt;</description></item><item><title>tensorflow variable 学习笔记</title><link>https://111qqz.com/en/post/acm-icpc/2017/2017-08-20-tensorflow-variable-notes/</link><pubDate>Sun, 20 Aug 2017 09:36:00 +0000</pubDate><guid>https://111qqz.com/en/post/acm-icpc/2017/2017-08-20-tensorflow-variable-notes/</guid><description>&lt;p&gt;参考资料：&lt;/p&gt;
&lt;p&gt;&lt;a href="https://www.tensorflow.org/programmers_guide/variables" target="_blank" rel="noreferrer"&gt;programmers_guide/variables&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;&lt;a href="https://www.tensorflow.org/api_docs/python/tf/Variable" target="_blank" rel="noreferrer"&gt;tf/Variable&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;之前感觉对tensorflow 的variable的理解不是很深刻&amp;hellip;跑个模型啥的倒不会有什么问题，但是涉及分布式，模型并行之类的，感觉有些地方还是要理解得仔细一点比较好。&lt;/p&gt;</description></item><item><title>tensorflow Session 学习笔记</title><link>https://111qqz.com/en/post/acm-icpc/2017/2017-08-20-tensorflow-session-notes/</link><pubDate>Sun, 20 Aug 2017 08:21:57 +0000</pubDate><guid>https://111qqz.com/en/post/acm-icpc/2017/2017-08-20-tensorflow-session-notes/</guid><description>&lt;p&gt;&lt;a href="https://www.tensorflow.org/api_docs/python/tf/Session" target="_blank" rel="noreferrer"&gt;tensorflow-session官方文档&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;说下我自己的理解：&lt;/p&gt;
&lt;p&gt;session中文一般叫会话，可以理解成op执行时候需要的一层虚拟化的封装。&lt;/p&gt;</description></item><item><title>Distributed Tensorflow : Cannot assign a device for operation save</title><link>https://111qqz.com/en/post/acm-icpc/2017/2017-08-14-distributed-tensorflow-cannot-assign-a-device-for-operation-save/</link><pubDate>Mon, 14 Aug 2017 01:55:15 +0000</pubDate><guid>https://111qqz.com/en/post/acm-icpc/2017/2017-08-14-distributed-tensorflow-cannot-assign-a-device-for-operation-save/</guid><description>&lt;p&gt;是在使用分布式tensorflow遇到的一个错误&lt;/p&gt;
&lt;p&gt;报错如下：&lt;/p&gt;
&lt;p&gt;InvalidArgumentError (see above for traceback): Cannot assign a device for operation &amp;lsquo;save/Rest│| 2 GeForce GTX 1080 On | 0000:08:00.0 Off | N/A |
oreV2_888&amp;rsquo;: Operation was explicitly assigned to /job:worker/task:0/device:CPU:0 but available │| 24% 39C P8 12W / 180W | 0MiB / 8114MiB | 0% Default |
devices are [ /job:localhost/replica:0/task:0/cpu:0, /job:localhost/replica:0/task:0/gpu:0 ]. Make sure the device specification refers to a valid device.&lt;/p&gt;</description></item><item><title>分布式 tensorflow 学习笔记(非最终版)</title><link>https://111qqz.com/en/post/acm-icpc/2017/2017-08-07-tensorflowe-notes/</link><pubDate>Mon, 07 Aug 2017 12:54:23 +0000</pubDate><guid>https://111qqz.com/en/post/acm-icpc/2017/2017-08-07-tensorflowe-notes/</guid><description>&lt;p&gt;感觉资料不是很多，先收集资料好了。&lt;/p&gt;
&lt;p&gt;&lt;a href="https://www.tensorflow.org/deploy/distributed" target="_blank" rel="noreferrer"&gt;tf-distributed官网文档&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;&lt;a href="https://stackoverflow.com/questions/41600321/distributed-tensorflow-the-difference-between-in-graph-replication-and-between" target="_blank" rel="noreferrer"&gt;SO-between-graph和in-graph的区别&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;&lt;a href="https://github.com/tensorflow/models/blob/master/inception/README.md" target="_blank" rel="noreferrer"&gt;inception.README.md&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;&lt;a href="https://www.tensorflow.org/versions/r0.12/api_docs/python/train/other_functions_and_classes" target="_blank" rel="noreferrer"&gt;SyncReplicasOptimizer&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;&lt;a href="https://stackoverflow.com/questions/41480149/how-does-ps-work-in-distribute-tensorflow" target="_blank" rel="noreferrer"&gt;SO_How does ps work in distribute Tensorflow?&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;update:在多个nodes（机）上跑。。。tf默认是异步更新的。。。同步的话。。大概需要syncreplicasoptimizer?&lt;/p&gt;</description></item><item><title>tensorflow Supervisor 学习笔记</title><link>https://111qqz.com/en/post/acm-icpc/2017/2017-08-04-tensorflow-supervisor-notes/</link><pubDate>Fri, 04 Aug 2017 09:22:44 +0000</pubDate><guid>https://111qqz.com/en/post/acm-icpc/2017/2017-08-04-tensorflow-supervisor-notes/</guid><description>&lt;p&gt;update:supervisor的缺点是遇到问题只会抛异常，所以现在有一个better的管理工具,&lt;a href="https://www.tensorflow.org/api_docs/python/tf/train/MonitoredSession" target="_blank" rel="noreferrer"&gt;MonitoredSession&lt;/a&gt;&lt;/p&gt;</description></item><item><title>TensorFlow Architecture 学习笔记（二）Adding a New Op</title><link>https://111qqz.com/en/post/acm-icpc/2017/2017-08-02-tensorflow-architecture-notes-2/</link><pubDate>Wed, 02 Aug 2017 03:07:37 +0000</pubDate><guid>https://111qqz.com/en/post/acm-icpc/2017/2017-08-02-tensorflow-architecture-notes-2/</guid><description>&lt;h1 class="relative group"&gt;Adding a New Op
 &lt;div id="adding-a-new-op" class="anchor"&gt;&lt;/div&gt;
 
 &lt;span
 class="absolute top-0 w-6 transition-opacity opacity-0 -start-6 not-prose group-hover:opacity-100 select-none"&gt;
 &lt;a class="text-primary-300 dark:text-neutral-700 !no-underline" href="#adding-a-new-op" aria-label="Anchor"&gt;#&lt;/a&gt;
 &lt;/span&gt;
 
&lt;/h1&gt;
&lt;pre&gt;&lt;code&gt; * [目录](https://www.tensorflow.org/extend/adding_an_op#top_of_page)
 * [定义运算的接口](https://www.tensorflow.org/extend/adding_an_op#define_the_ops_interface)
 * [实现运算的核心部分(kernels)](https://www.tensorflow.org/extend/adding_an_op#implement_the_kernel_for_the_op)

 * [多线程cpu kernels](https://www.tensorflow.org/extend/adding_an_op#multi-threaded_cpu_kernels)
 * [GPU kernels](https://www.tensorflow.org/extend/adding_an_op#gpu_kernels)


 * [构建运算库](https://www.tensorflow.org/extend/adding_an_op#build_the_op_library)

 * [用系统编译器编译你的运算（TensorFlow binary installation）](https://www.tensorflow.org/extend/adding_an_op#compile_the_op_using_your_system_compiler_tensorflow_binary_installation)
 * [使用bazel编译你的运算(TensorFlow source installation)](https://www.tensorflow.org/extend/adding_an_op#compile_the_op_using_bazel_tensorflow_source_installation)


 * [在 Python 中使用你的运算](https://www.tensorflow.org/extend/adding_an_op#use_the_op_in_python)
 * [验证你添加的运算可以工作](https://www.tensorflow.org/extend/adding_an_op#verify_that_the_op_works)
 * [在你的运算中添加高级特性](https://www.tensorflow.org/extend/adding_an_op#building_advanced_features_into_your_op)

 * [条件检查和验证](https://www.tensorflow.org/extend/adding_an_op#conditional_checks_and_validation)
 * [Op registration](https://www.tensorflow.org/extend/adding_an_op#op_registration)
 * [GPU Support](https://www.tensorflow.org/extend/adding_an_op#gpu_support)
 * [用python 实现梯度](https://www.tensorflow.org/extend/adding_an_op#implement_the_gradient_in_python)
 * [Shape functions in C++](https://www.tensorflow.org/extend/adding_an_op#shape_functions_in_c)
&lt;/code&gt;&lt;/pre&gt;
&lt;p&gt;对于要添加原生tensorflow中没有定义的运算的需求，首先建议在python层面，能不能将需要的op用其他原生的op拼凑起来。&lt;/p&gt;</description></item><item><title>TensorFlow Architecture 学习笔记（一）</title><link>https://111qqz.com/en/post/acm-icpc/2017/2017-08-01-tensorflow-architecture-notes-1/</link><pubDate>Tue, 01 Aug 2017 03:01:12 +0000</pubDate><guid>https://111qqz.com/en/post/acm-icpc/2017/2017-08-01-tensorflow-architecture-notes-1/</guid><description>&lt;p&gt;&lt;strong&gt;这篇文章不会涉及tensorflow的具体使用，而是专注于介绍tensorflow的架构，目的是让开发者能够对tensorflow现有框架进行自定义的扩展。&lt;/strong&gt;&lt;/p&gt;</description></item><item><title>how to copy &amp; modify nets model on tensorflow slim</title><link>https://111qqz.com/en/post/%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0/%E8%AE%A1%E7%AE%97%E6%9C%BA%E8%A7%86%E8%A7%89/2017-07-19-how-to-copy-modify-nets-model-on-tensorflow-slim/</link><pubDate>Wed, 19 Jul 2017 06:21:40 +0000</pubDate><guid>https://111qqz.com/en/post/%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0/%E8%AE%A1%E7%AE%97%E6%9C%BA%E8%A7%86%E8%A7%89/2017-07-19-how-to-copy-modify-nets-model-on-tensorflow-slim/</guid><description>&lt;p&gt;想要修改tensorflow-slim 中 nets中的某个model,例如明明为kk_v2.py&lt;/p&gt;</description></item><item><title>tensorflow slim 源码分析</title><link>https://111qqz.com/en/post/%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0/%E8%AE%A1%E7%AE%97%E6%9C%BA%E8%A7%86%E8%A7%89/2017-07-16-tensorflow-slim-code-notes/</link><pubDate>Sun, 16 Jul 2017 13:10:04 +0000</pubDate><guid>https://111qqz.com/en/post/%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0/%E8%AE%A1%E7%AE%97%E6%9C%BA%E8%A7%86%E8%A7%89/2017-07-16-tensorflow-slim-code-notes/</guid><description>&lt;p&gt;py的源码看起来还是很愉快的。。。（虽然熟练成程度完全不如cpp。。。。&lt;/p&gt;</description></item></channel></rss>