Plant Classification and Q&A System

This application allows users to:

1. Upload images of plants for classification
2. Browse a plant library for information
3. Ask questions about specific plants

Features

• Plant Classification: Upload images to identify plant species
• Plant Library: Browse a comprehensive database of plant information
• Question Answering: Ask specific questions about plants and get AI-generated responses

Deployment on Hugging Face Spaces

This application is configured for deployment on Hugging Face Spaces using Docker.

Repository Structure

├── main.py                # FastAPI backend
├── classifier.py          # Plant classification model
├── embedder.py            # Image embedding model
├── llm_huggingface.py     # Plant QA system
├── retrieval_system.py    # Image retrieval system
├── Dockerfile             # Docker configuration
├── requirements.txt       # Python dependencies
└── frontend/             # React frontend

Model Files

This application references the following model files, which are automatically downloaded during startup:

• dino_best.pth: Classification model
• embedding_best.pth: Image embedding model
• plant_index_2.pkl: Retrieval system index
• merge_metadata.json: Plant metadata

These files are sourced from the Hugging Face repository: hqta1110/plant-classification-models

Local Development

1. Install Python dependencies:

   pip install -r requirements.txt
   

2. Set up the frontend:

   cd frontend
   npm install
   npm run build
   

3. Run the FastAPI server:

   uvicorn main:app --reload
   

Usage

1. Classification: Upload a plant image to identify its species
2. Library Mode: Browse the catalog of plants and access detailed information
3. Q&A: Ask questions about specific plants and receive detailed answers

License

[Specify your license information here]