/mobile-pest-identification

A Lightweight Deep Learning Model for Automated Pest Detection on Mobile

Primary LanguagePythonCreative Commons Zero v1.0 UniversalCC0-1.0

A Lightweight Deep Learning Model for Automated Pest Detection on Mobile.

User Guide

Step 1. To clone this repository, run:
git clone <this repo>
Step 2. To install dependencies run:
pip install -r requirements.txt
Step 3. Download pest-imagery IP102v1.1 dataset:
Note: This dataset contains more than 75,000 images belonging to 102 categories, only a subset of which is showcased in this repository.

Step 4. To run the model locally, run:
python src/pest.py

This will load the data, train the model and save the mobile optimized version in src/model/pest/. For deploying the model to a mobile app, see PyTorch mobile.

Demonstration on the Wikilimo App: