Android Studio搭建Tensorflow Lite项目和加载tflite模型文件

小咪咪 2022-12-21 00:52 196阅读 0赞

# 搭建Tensorflow Lite项目 #

新建一个安卓工程后，将项目目录的展示方式设置成“Project”模式，会看到两个同名的gradle文件，对这两个文件进行修改。  
![在这里插入图片描述][watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3UwMTEyMDg5ODQ_size_16_color_FFFFFF_t_70_pic_center]

## 修改项目的build.gradle文件 ##

buildscript {
        repositories {
        //修改1
            google()
            jcenter()
            mavenLocal()
            
        }
        dependencies {
            classpath 'com.android.tools.build:gradle:3.2.1'
    
            // NOTE: Do not place your application dependencies here; they belong
            // in the individual module build.gradle files
        }
    }
    
    allprojects {
        repositories {
         //修改2
            google()
            jcenter()
            mavenLocal()
            
        }
    }
    
    task clean(type: Delete) {
        delete rootProject.buildDir
    }

## 修改app的build.gradle文件 ##

apply plugin: 'com.android.application'
    
    android {
        compileSdkVersion 30
        buildToolsVersion "30.0.1"
        defaultConfig {
            applicationId "com.example.word_ocr"
            minSdkVersion 21
            targetSdkVersion 30
            versionCode 1
            versionName "1.0"
            testInstrumentationRunner "androidx.test.runner.AndroidJUnitRunner"
        }
        buildTypes {
            release {
                minifyEnabled false
                proguardFiles getDefaultProguardFile('proguard-android-optimize.txt'), 'proguard-rules.pro'
            }
        }
        // 修改1  设置不编译模型文件
        aaptOptions {
            noCompress "tflite"
        }
    }
    
    dependencies {
        implementation fileTree(dir: 'libs', include: ['*.jar'])
        implementation 'androidx.appcompat:appcompat:1.2.0'
        implementation 'androidx.constraintlayout:constraintlayout:2.0.4'
        implementation 'com.google.android.material:material:1.2.1'
        testImplementation 'junit:junit:4.12'
        androidTestImplementation 'androidx.test.ext:junit:1.1.2'
        androidTestImplementation 'androidx.test.espresso:espresso-core:3.3.0'
        //修改2  添加TensorFlow Lite依赖库
        implementation('org.tensorflow:tensorflow-lite:0.0.0-nightly') { changing = true }
        implementation('org.tensorflow:tensorflow-lite-gpu:0.0.0-nightly') { changing = true }
        implementation('org.tensorflow:tensorflow-lite-support:0.0.0-nightly') { changing = true }
    
    }

以上操作时，注意保持网路通畅，不要使用代理。

## 添加模型文件 ##

在src/main目录下新建一个assets目录，将转换好的model.tflite文件放到这个目录下。

# 加载tflite模型文件 #

可以参考这篇文章：[Tensorflow Lite初探（Android）][Tensorflow Lite_Android]  
模型的输入输出对应：[DataType error: cannot resolve DataType of \[\[\[D][DataType error_ cannot resolve DataType of _D]，  
[Android——Tensorflow-Lite简单使用][Android_Tensorflow-Lite]  
[Android加载tensorflow模型文件（.tflite）][Android_tensorflow_.tflite]  
[手写数字识别–Android Studio 加载tensorflow模型][Android Studio _tensorflow]

参考：  
[使用tensorflow lite遇到的一些坑][tensorflow lite]  
[AS build.gradle——tensorflow-lite官方项目两个build.gradle详解以及无法下载依赖的解决方法][AS build.gradle_tensorflow-lite_build.gradle]  
[Android加载tensorflow模型文件（.tflite）][Android_tensorflow_.tflite]

[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3UwMTEyMDg5ODQ_size_16_color_FFFFFF_t_70_pic_center]: /images/20221120/d6468bf772074240ae235d0632b321ea.png
[Tensorflow Lite_Android]: https://blog.csdn.net/bhneo/article/details/78835224
[DataType error_ cannot resolve DataType of _D]: https://blog.csdn.net/u011208984/article/details/109741274
[Android_Tensorflow-Lite]: https://blog.csdn.net/qq_27634797/article/details/90513246?utm_medium=distribute.pc_relevant.none-task-blog-BlogCommendFromMachineLearnPai2-2.compare&depth_1-utm_source=distribute.pc_relevant.none-task-blog-BlogCommendFromMachineLearnPai2-2.compare
[Android_tensorflow_.tflite]: https://blog.csdn.net/liuhongshuo2012/article/details/106712250/
[Android Studio _tensorflow]: https://blog.csdn.net/qq_38956849/article/details/85041381
[tensorflow lite]: https://subaochen.github.io/android/2019/08/12/tensorflow-lite-note/
[AS build.gradle_tensorflow-lite_build.gradle]: https://blog.csdn.net/umbrellalalalala/article/details/88345208