如何将tensorflow程序移植到手机APP

首先，我们先介绍tensorflow上提供的demo。
我们这里用Bazel。

一.下载tensorflow。

git clone --recurse-submodules https://github.com/tensorflow/tensorflow.git

二.安装Bazel。安装教程：

https://docs.bazel.build/versions/master/install.html

三.安装NDK。只需下载解压即可。注意，这里最好下载12版本的。其他版本亲测，会出错。

https://developer.android.com/ndk/downloads/older_releases.html#ndk-12b-downloads

四.安装SDK。SDK可以单独安装，但是我们这里直接安装Android stdio ，然后会提示我们安装SDK。下载方法：

https://developer.android.com/studio/index.html

五.修改workspace文件

# Uncomment and update the paths in these entries to build the Android demo.
android_sdk_repository(
    name = "androidsdk",
    api_level = 26,
    # Ensure that you have the build_tools_version below installed in the
    # SDK manager as it updates periodically.
    build_tools_version = "26.0.1",
    # Replace with path to Android SDK on your system
    path = "/home/no1/Android/Sdk/",
)
#
# Android NDK r12b is recommended (higher may cause issues with Bazel)
android_ndk_repository(
    name="androidndk",
    path="/home/no1/android-ndk-r12b/",
    # This needs to be 14 or higher to compile TensorFlow.
    # Please specify API level to >= 21 to build for 64-bit
    # archtectures or the Android NDK will automatically select biggest
    # API level that it supports without notice.
    # Note that the NDK version is not the API level.
    api_level=26)

六.编译：

bazel build -c opt //tensorflow/examples/android:tensorflow_demo

如果出错则换用这个指令：

bazel build --noincremental_dexing //tensorflow/examples/android:tensorflow_demo

完成之后：出现如下情况表示编译完成：

然后在目录：tensorflow/bazel-bin/tensorflow/examples/android 下找到我们想要的apk，传到手机上，安装之后，可以看到有四个小功能的APP