Tensorflow_Lite_Demo

An example Android application using TensorFLow Lite is available on Tensorflow github, Creating a project directory in tensorflow/tensorflow/contrib/lite/ , which is builted on Android studio 3.0.I have download the model of tflite format and complie the libtensorflowlite_jni.so and libtensorflowlite.jar

In the demo app, inference is done using the TensorFlow Lite Java API. The demo app classifies frames in real-time, displaying the top most probable classifications. It also displays the inference time taken to detect the object.

There are two ways to get the demo app to your device:

1,Use Android Studio to build the application. Here it's mine work ! 2,Download the source code for TensorFlow Lite and the demo and build it using bazel.I have just give building commands !

To build the demo app, run bazel:

TfLiteCameraDemo ---- The demo app classifier

#./tensorflow direction run the bellow command

bazel build --cxxopt=--std=c++11 //tensorflow/contrib/lite/java/demo/app/src/main:TfLiteCameraDemo

for more infomation about TfLiteCameraDemo

tflite_demo ---- The demo app on classifier, detector, speech

#./tensorflow direction run the bellow command

sudo bazel build -c opt --config=android_arm{,64} --cxxopt='--std=c++11' "//tensorflow/contrib/lite/examples/android:tflite_demo"

for more infomation about tflite_demo

bazel-bin/tensorflow/contrib/lite/java/libtensorflowlite.jar ---- the java api for tensorflow lite

sudo bazel build -c opt --config=android_arm{,64} --cxxopt='--std=c++11' "//tensorflow/contrib/lite/java:tensorflowlite"

bazel-bin/tensorflow/examples/android/libtensorflow_demo.so ---- the native c++ jni interface for libtensorflowlite.jar

sudo bazel build -c opt --config=android_arm{,64} --cxxopt='--std=c++11' "//tensorflow/examples/android:libtensorflow_demo.so"

attention

1, Edit your WORKSPACE to add SDK and NDK targets.

android_sdk_repository( name = "androidsdk", api_level = 25, # 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/XXXX/Software/SDK", )

android_ndk_repository( name="androidndk", path="/home/XXXX/Software/NDK/android-ndk-r13b", # 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=25)

2, Build this demo app with Bazel. The demo needs C++11. We configure the fat_apk_cpu flag to package support for 4 hardware variants. You may replace it with --config=android_arm64 on a 64-bit device and --config=android_arm for 32-bit device: For examples:

bazel build -c opt --cxxopt='--std=c++11' --fat_apk_cpu=x86,x86_64,arm64-v8a,armeabi-v7a
//tensorflow/contrib/lite/examples/android:tflite_demo

bazel build -c opt --config=android_arm{,64} --cxxopt='--std=c++11' "//tensorflow/contrib/lite/examples/android:tflite_demo"

3, Build the Tensorflow mobile demo using Bazel,which has a fuller set of supported functionality, while TensorFlow Lite supports only a limited set of operators, so not all models will work on it by default.

#bazel-bin/tensorflow/examples/android/tensorflow_demo.apk bazel build -c opt //tensorflow/examples/android:tensorflow_demo

#bazel-bin/tensorflow/contrib/android/libandroid_tensorflow_inference_java.jar bazel build //tensorflow/contrib/android:android_tensorflow_inference_java

#bazel-bin/tensorflow/contrib/android/libtensorflow_inference.so sudo bazel build -c opt //tensorflow/contrib/android:libtensorflow_inference.so --crosstool_top=//external:android/crosstool --host_crosstool_top=@bazel_tools//tools/cpp:toolchain --cpu=armeabi-v7a --cxxopt='--std=c++11' --verbose_failures for more information about [Tensorflow Mobile] (https://www.tensorflow.org/mobile/android_build)