liuyinglao / DeepLabV3Example

Demo using DeeplabV3 with Pytorch Live

DeepLabV3 Example

The repository contains code for a PyTorch Live image segmentation prototype. The prototype uses the DeepLabV3 model for the semantic segmentation task and runs on-device. It runs on Android only for now.

NOTE: This example uses an unreleased version of PyTorch Live including an API that is currently under development and can change for the final release.

How was this project bootstrapped?

The project was bootstrapped with the following command:

npx torchlive-cli@nightly init DeepLabV3Example --template react-native-template-pytorch-live@nightly

Unused packages were removed and react-native upgraded to version 0.64.3.

Screenshots

Android	iOS

Run project in emulator or on a device

Prerequisites

Install React Native development depencencies. Follow the instructions for Setting up the development environment as provided on the React Native website.

Install project dependencies

Run yarn install to install the project dependencies.

Start Metro server

Start the Metro server, which is needed to build the app bundle (containing the transpiled TypeScript code in the <PROJECT>/src directory).

yarn start

Android

Build the apk for Android and install and run on the emulator (or on a physical device if connected via USB).

yarn android

See instructions on the React Native website for how to build the app in release variant.

iOS

Install CocoaPod dependencies

(cd ios && pod install)

Build the prototype app for iOS and run it in the simulator.

yarn ios

or use the following command to open the Xcode workspace in Xcode to build and run it.

xed ios/DeepLabV3Example.xcworkspace

See instructions on the React Native website for how to build the app in release scheme.