(prototype) Tracing-based Selective Build Mobile Interpreter in Android and iOS

Author: Chen Lai <https://github.com/cccclai>, Dhruv Matani <https://github.com/dhruvbird>

Warning

Tracing-based selective build a prototype feature to minimize library size. Since the traced result relies on the model input and traced environment, if the tracer runs in a different environment than mobile interpreter, the operator list might be different from the actual used operator list and missing operators error might raise.

Introduction

This tutorial introduces a new way to custom build mobile interpreter to further optimize mobile interpreter size. It restricts the set of operators included in the compiled binary to only the set of operators actually needed by target models. It is a technique to reduce the binary size of PyTorch for mobile deployments. Tracing Based Selective Build runs a model with specific representative inputs, and records which operators were called. The build then includes just those operators.

Following are the processes to use tracing-based selective approach to build a custom mobile interpreter.

1. Prepare model with bundled input

import numpy as np
import torch
import torch.jit
import torch.utils
import torch.utils.bundled_inputs
from PIL import Image
from torchvision import transforms

# Step 1. Get the model
model = torch.hub.load('pytorch/vision:v0.7.0', 'deeplabv3_resnet50', pretrained=True)
model.eval()

scripted_module = torch.jit.script(model)
# Export full jit version model (not compatible lite interpreter), leave it here for comparison
scripted_module.save("deeplabv3_scripted.pt")
# Export lite interpreter version model (compatible with lite interpreter)
# path = "<base directory where models are stored>"

scripted_module._save_for_lite_interpreter(f"${path}/deeplabv3_scripted.ptl")

model_file = f"${path}/deeplabv3_scripted.ptl"

# Step 2. Prepare inputs for the model
input_image_1 = Image.open(f"${path}/dog.jpg")
preprocess = transforms.Compose([
    transforms.ToTensor(),
    transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225]),
])

input_tensor_1 = preprocess(input_image_1)
input_batch_1 = input_tensor_1.unsqueeze(0) # create a mini-batch as expected by the model

scripted_module = torch.jit.load(model_file)
scripted_module.forward(input_batch_1) # optional, to validate the model can run with the input_batch_1

input_image_2 = Image.open(f"${path}/deeplab.jpg")
input_tensor_2 = preprocess(input_image_2)
input_batch_2 = input_tensor_2.unsqueeze(0) # create a mini-batch as expected by the model

scripted_module = torch.jit.load(model_file)
scripted_module.forward(input_batch_2) # optional, to validate the model can run with the input_batch_2

# Step 3. Bundle the model with the prepared input from step2. Can bundle as many input as possible.
bundled_model_input = [
    (torch.utils.bundled_inputs.bundle_large_tensor(input_batch_1), ),
    (torch.utils.bundled_inputs.bundle_large_tensor(input_batch_2), )]
bundled_model = torch.utils.bundled_inputs.bundle_inputs(scripted_module, bundled_model_input)
bundled_model._save_for_lite_interpreter(f"${path}/deeplabv3_scripted_with_bundled_input.ptl")

2. Build tracer

MACOSX_DEPLOYMENT_TARGET=10.9 CC=clang CXX=clang++ MAX_JOBS=16 TRACING_BASED=1 python setup.py develop

3. Run tracer with the model with bundled input

./build/bin/model_tracer --model_input_path ${path}/deeplabv3_scripted_with_bundled_input.ptl --build_yaml_path ${path}/deeplabv3_scripted.yaml

Android

Get the Image Segmentation demo app in Android: https://github.com/pytorch/android-demo-app/tree/master/ImageSegmentation

1. Tracing-based build libtorch lite for android: Build libtorch for android for all 4 android abis (armeabi-v7a, arm64-v8a, x86, x86_64) by running

SELECTED_OP_LIST=${path}/deeplabv3_scripted.yaml TRACING_BASED=1  ./scripts/build_pytorch_android.sh

if it will be tested on Pixel 4 emulator with x86, use cmd BUILD_LITE_INTERPRETER=1 ./scripts/build_pytorch_android.sh x86 to specify abi to save build time.

SELECTED_OP_LIST=${path}/deeplabv3_scripted.yaml TRACING_BASED=1  ./scripts/build_pytorch_android.sh x86

After the build finish, it will show the library path:

BUILD SUCCESSFUL in 55s
134 actionable tasks: 22 executed, 112 up-to-date
+ find /Users/chenlai/pytorch/android -type f -name '*aar'
+ xargs ls -lah
-rw-r--r--  1 chenlai  staff    13M Feb 11 11:48 /Users/chenlai/pytorch/android/pytorch_android/build/outputs/aar/pytorch_android-release.aar
-rw-r--r--  1 chenlai  staff    36K Feb  9 16:45 /Users/chenlai/pytorch/android/pytorch_android_torchvision/build/outputs/aar/pytorch_android_torchvision-release.aar

2. Use the PyTorch Android libraries built from source in the ImageSegmentation app: Create a folder libs in the path, the path from repository root will be ImageSegmentation/app/libs. Copy pytorch_android-release to the path ImageSegmentation/app/libs/pytorch_android-release.aar. Copy pytorch_android_torchvision (downloaded from Pytorch Android Torchvision Nightly) to the path ImageSegmentation/app/libs/pytorch_android_torchvision.aar. Update the dependencies part of ImageSegmentation/app/build.gradle to

dependencies {
    implementation 'androidx.appcompat:appcompat:1.2.0'
    implementation 'androidx.constraintlayout:constraintlayout:2.0.2'
    testImplementation 'junit:junit:4.12'
    androidTestImplementation 'androidx.test.ext:junit:1.1.2'
    androidTestImplementation 'androidx.test.espresso:espresso-core:3.3.0'


    implementation(name:'pytorch_android-release', ext:'aar')
    implementation(name:'pytorch_android_torchvision', ext:'aar')

    implementation 'com.android.support:appcompat-v7:28.0.0'
    implementation 'com.facebook.fbjni:fbjni-java-only:0.0.3'
}

Update all projects part in ImageSegmentation/build.gradle to

allprojects {
    repositories {
        google()
        jcenter()
        flatDir {
            dirs 'libs'
        }
    }
}

3. Test app: Build and run the ImageSegmentation app in Android Studio

iOS

Get ImageSegmentation demo app in iOS: https://github.com/pytorch/ios-demo-app/tree/master/ImageSegmentation

1. Build libtorch lite for iOS:

SELECTED_OP_LIST=${path}/deeplabv3_scripted.yaml TRACING_BASED=1 IOS_PLATFORM=SIMULATOR ./scripts/build_ios.sh

2. Remove Cocoapods from the project (this step is only needed if you ran pod install):

pod deintegrate

3. Link ImageSegmentation demo app with the custom built library:

Open your project in XCode, go to your project Target’s Build Phases - Link Binaries With Libraries, click the + sign and add all the library files located in build_ios/install/lib. Navigate to the project Build Settings, set the value Header Search Paths to build_ios/install/include and Library Search Paths to build_ios/install/lib. In the build settings, search for other linker flags. Add a custom linker flag below -all_load. Finally, disable bitcode for your target by selecting the Build Settings, searching for Enable Bitcode, and set the value to No.

4. Build and test the app in Xcode.

Conclusion

In this tutorial, we demonstrated a new way to custom build PyTorch’s efficient mobile interpreter - tracing-based selective build, in an Android and iOS app.

We walked through an Image Segmentation example to show how to bundle inputs to a model, generated operator list by tracing the model with bundled input, and build a custom torch library from source with the operator list from tracing result.

The custom build is still under development, and we will continue improving its size in the future. Note, however, that the APIs are subject to change in future versions.

Thanks for reading! As always, we welcome any feedback, so please create an issue here <https://github.com/pytorch/pytorch/issues>`.

Learn More

• To learn more about PyTorch Mobile, please refer to PyTorch Mobile Home Page <https://pytorch.org/mobile/home/>

• To learn more about Image Segmentation, please refer to the Image Segmentation DeepLabV3 on Android Recipe <https://pytorch.org/tutorials/beginner/deeplabv3_on_android.html>_