TensorFlow 簡介
2015 年 11 月 9 日陆错，Google Research 發(fā)布了文章：TensorFlow - Google’s latest machine learning system, open sourced for everyone，正式宣布其新一代機器學(xué)習(xí)系統(tǒng)開源腊瑟。
至于 Google 為什么要開源 TensorFlow喂分，官方的說法是：
If TensorFlow is so great, why open source it rather than keep it proprietary? The answer is simpler than you might think: We believe that machine learning is a key ingredient to the innovative products and technologies of the future. Research in this area is global and growing fast, but lacks standard tools. By sharing what we believe to be one of the best machine learning toolboxes in the world, we hope to create an open standard for exchanging research ideas and putting machine learning in products. Google engineers really do use TensorFlow in user-facing products and services, and our research group intends to share TensorFlow implementations along side many of our research publications.
Here's Why Google Is Open-Sourcing Some Of Its Most Important Technology 文章中援引了 TensorFlow 開發(fā)者的說法：
The decision to open-source was the brainchild of Jeff Dean, who felt that the company’s innovation efforts were being hampered by the slow pace of normal science. Google researchers would write a paper, which would then be discussed at a conference some months later. Months after that somebody else would write another paper building on their work.
Dean saw that open-sourcing TensorFlow could significantly accelerate the process. Rather than having to wait for the next paper or conference, Google’s researchers could actively collaborate with the scientific community in real-time. Smart people outside of Google could also improve the source code and, by sharing machine learning techniques more broadly, it would help populate the field with more technical talent.
“Having this system open sourced we’re able to collaborate with many other researchers at universities and startups, which gives us new ideas about how we can advance our technology. Since we made the decision to open-source, the code runs faster, it can do more things and it’s more flexible and convenient,” says Rajat Monga, who leads the TensorFlow team.
毫無意外地锦庸，TensorFlow 在 Github 上的 Repo 在很短的時間內(nèi)就收獲了大量的 Star 和 Fork，學(xué)術(shù)界和工業(yè)界都對其表示了巨大的興趣蒲祈，并投身于 TensorFlow 的社區(qū)和 Google 一起完善和改進 TensorFlow甘萧。
然而，當(dāng)時在 Github 做基準(zhǔn)測試梆掸、目前就職于 Facebook AI 部門的程序員 Soumith 發(fā)布了文章 Benchmark TensorFlow（中文解讀）扬卷，對 TensorFlow 和其他主流深度學(xué)習(xí)框架的性能進行了比較，結(jié)果差強人意酸钦。當(dāng)然怪得，Google 團隊表示會繼續(xù)優(yōu)化，并在后面的版本中支持分布式卑硫。
2016 年 4 月 13 日徒恋，Google 通過文章 Announcing TensorFlow 0.8 – now with distributed computing support! 正式發(fā)布支持分布式的 TensorFlow 0.8 版本，結(jié)合之前對 CPU 和 GPU 的支持欢伏，TensorFlow 終于可以被用于實際的大數(shù)據(jù)生產(chǎn)環(huán)境中了入挣。
2016 年 4 月 29 日，開發(fā)出目前最強圍棋 AI 的 Google 旗下 DeepMind 宣布：DeepMind moves to TensorFlow颜懊，這在業(yè)界被認(rèn)為 TensorFlow 終于可以被當(dāng)作 TensorFlow 在工業(yè)界發(fā)展的里程碑事件财岔，極大提升了 TensorFlow 使用者的研究熱情。
The Good, Bad & Ugly of TensorFlow（中文翻譯）對目前 TensorFlow 的優(yōu)缺點做了詳細(xì)的分析河爹。
Windows 平臺
Tensorflow 官方文檔 中匠璧，對 Android Demo 的編譯介紹時用到了 bazel，該工具對 Windows 平臺的支持處于實驗階段咸这，就不推薦了夷恍，Github 上有一個 使用 NDK 在 Anroid Studio 中進行編譯的示例工程，大家可以 clone 下來使用媳维。
Ubuntu 14.04
這里假定 Ubuntu 14.04 系統(tǒng)上還沒有 Android 開發(fā)環(huán)境酿雪。
安裝 Java 1.8
$ sudo apt-get install software-properties-common
$ sudo add-apt-repository ppa:webupd8team/java
$ sudo apt-get update
$ sudo apt-get install oracle-java8-installer
配置 Java 環(huán)境變量，將下面的內(nèi)容添加到 /etc/environment:
JAVA_HOME="/usr/lib/jvm/java-8-oracle"
安裝 bazel
$ echo "deb [arch=amd64] http://storage.googleapis.com/bazel-apt stable jdk1.8" | sudo tee /etc/apt/sources.list.d/bazel.list
$ curl https://bazel.io/bazel-release.pub.gpg | sudo apt-key add -
$ sudo apt-get update && sudo apt-get install bazel
$ sudo apt-get upgrade bazel
詳細(xì)的說明可以參考 bazel 的官方文檔侄刽。
下載 tensorflow
$ cd ~/
$ git clone https://github.com/tensorflow/tensorflow.git
之后的步驟基本來自 TensorFlow on Android 的翻譯：
下載解壓 Android SDK
$ wget https://dl.google.com/android/android-sdk_r24.4.1-linux.tgz
$ tar xvzf android-sdk_r24.4.1-linux.tgz -C ~/tensorflow
更新 SDK：
$ cd ~/tensorflow/android-sdk-linux
# 如果希望在熟悉的 SDK Manager 中進行操作指黎，可以去掉下面命令中的 --no-ui
$ tools/android update sdk --no-ui
下載解壓 NDK
$ wget https://dl.google.com/android/repository/android-ndk-r12b-linux-x86_64.zip
$ unzip android-ndk-r12b-linux-x86_64.zip -d ~/tensorflow
下載 tensorflow 的 model
$ cd ~/tensorflow
$ wget https://storage.googleapis.com/download.tensorflow.org/models/inception5h.zip -O /tmp/inception5h.zip
$ unzip /tmp/inception5h.zip -d tensorflow/examples/android/assets/
修改 WORKSPACE
$ gedit ~/tensorflow/WORKSPACE
反注釋 android_sdk_repository 和 android_ndk_repository 部分，用下面的內(nèi)容替換：
android_sdk_repository(
name = "androidsdk",
api_level = 24,
build_tools_version = "24.0.3",
# Replace with path to Android SDK on your system
path = "/home/ross/Downloads/android-sdk-linux",
)
android_ndk_repository(
name="androidndk",
path="/home/ross/Downloads/android-ndk-r12b",
api_level=24)
編譯 tensorflow 的 Android Demo App：
$ cd ~/tensorflow
$ bazel build //tensorflow/examples/android:tensorflow_demo
如果一切順利就會在最后看到下面的提示：
bazel-bin/tensorflow/examples/android/tensorflow_demo_deploy.jar
bazel-bin/tensorflow/examples/android/tensorflow_demo_unsigned.apk
bazel-bin/tensorflow/examples/android/tensorflow_demo.apk
INFO: Elapsed time: 109.114s, Critical Path: 37.45s
Android Demo 分析
整個 Demo 的目錄結(jié)構(gòu)和使用 Jni 的 Android 工程是相同的州丹，在 ~/tensorflow/tensorflow/examples/android/jni 目錄下醋安，放著 native 的代碼：
├── imageutils_jni.cc
├── __init__.py
├── rgb2yuv.cc
├── rgb2yuv.h
├── yuv2rgb.cc
└── yuv2rgb.h
Java interface 相關(guān)的 Java 類在 https://github.com/tensorflow/tensorflow/tree/master/tensorflow/contrib/android 目錄里面杂彭，可以考慮將其直接集成到自己的項目中。
Demo 所需的 native 實現(xiàn)在 https://github.com/tensorflow/tensorflow/tree/master/tensorflow/examples/android/jni 目錄里面吓揪。
https://github.com/tensorflow/tensorflow/blob/master/tensorflow/examples/android/src/org/tensorflow/demo/TensorFlowImageListener.java 里面定義了用到的 tensorflow model亲怠，protobuf 格式，識別結(jié)果的 labels 等：
private static final Logger LOGGER = new Logger();
private static final boolean SAVE_PREVIEW_BITMAP = false;
private static final String MODEL_FILE = "file:///android_asset/tensorflow_inception_graph.pb";
private static final String LABEL_FILE =
"file:///android_asset/imagenet_comp_graph_label_strings.txt";
private static final int NUM_CLASSES = 1001;
private static final int INPUT_SIZE = 224;
private static final int IMAGE_MEAN = 117;
如果想使用自己的模型柠辞，使用 tensorflow 解決其他的問題团秽，通過修改上面提到的代碼和模塊來完成。TensorFlow on Android 文章就提到了具體的步驟叭首。
最后习勤，Tensorflow 也支持移植到 iOS 應(yīng)用中，可以參考 TalkingData SDK Team 的技術(shù)博客文章 iOS 開發(fā)迎來機器學(xué)習(xí)的春天--- TensorFlow放棒。
