traffic-sign-android

Traffic sign classifier trained using fastai and deployed to run in real time (60-70 FPS) on Android using PyTorch Mobile.

The Android app is based on pytorch/android-demo-app, an official demo app for PyTorch Mobile project.

Training the model

To train the model, launch Traffic_Sign_Classifier.ipynb notebook using Google Colab, preferably using GPU runtime.

This repo can be directly imported using Google Colab GitHub import feature.

To build and launch the Android app, open Android/build.gradle in Android Studio.

Name		Name	Last commit message	Last commit date
Latest commit History 12 Commits
Android		Android
images		images
models		models
README.md		README.md
Traffic_Sign_Classifier.html		Traffic_Sign_Classifier.html
Traffic_Sign_Classifier.ipynb		Traffic_Sign_Classifier.ipynb
WRITEUP.md		WRITEUP.md
signnames.csv		signnames.csv