♻️ TrashSort AI

A lightweight AI-based simulation environment for intelligent trash classification, built using a clean OpenEnv-style structure.

---

🚀 Live Demo

👉 https://rutiikal-trashsort-ai.hf.space

---

🧠 About the Project

TrashSort AI is a simple yet structured environment where an AI model learns to classify waste into four categories:

• Plastic
• Paper
• Metal
• Organic

The system evaluates the model across different difficulty levels and returns performance scores.

---

⚙️ How It Works

1. The environment generates a random trash type
2. The model predicts the category
3. Reward is assigned based on correctness
4. Performance is evaluated across multiple tasks:
   • Easy
   • Medium
   • Hard

---

📂 Project Structure

environment/
env.py
models.py

tasks/
easy.py
medium.py
hard.py

inference.py
openenv.yaml
Dockerfile
requirements.txt

---

▶️ Run Locally

pip install -r requirements.txt
python inference.py

---

🐳 Run with Docker

docker build -t trashsort-ai .
docker run trashsort-ai

---

📊 Sample Output

Easy Score: 0.6  
Medium Score: 0.64  
Hard Score: 0.5  
Final Score: 0.58

---

✨ Features

• Clean and modular environment design
• Multi-level evaluation (Easy, Medium, Hard)
• Simple AI prediction model
• Reproducible setup
• Docker support
• Live demo available

---

⭐ Support

If you like this project, please consider giving it a ⭐ on GitHub!

---