---
language: en
license: apache-2.0
tags:
- computer-vision
- image-classification
- waste-management
- mobilenetv2
- sustainability
datasets:
- waste-classification-data
---

# 🗑️ Trash Sorter AI

Classifies waste as **Organic** or **Recyclable** with 93.37% accuracy.

## Quick Start

```python
# 1. Install
!pip install tensorflow

# 2. Load model
from huggingface_hub import hf_hub_download
from tensorflow import keras

model_path = hf_hub_download(
    repo_id="tahazyan/trash-sorter-ai",
    filename="trash_sorter.keras"
)
model = keras.models.load_model(model_path)

# 3. Use
import cv2
import numpy as np

def classify_waste(image_path):
    img = cv2.imread(image_path)
    img = cv2.resize(img, (224, 224))
    img = img / 255.0
    img = np.expand_dims(img, axis=0)

    prediction = model.predict(img)[0]
    classes = ["Organic", "Recyclable"]
    return classes[np.argmax(prediction)]
```

## Model Details

**Architecture:** MobileNetV2 + Custom layers  
**Input:** 224×224 RGB images  
**Output:** 2 classes (Organic, Recyclable)  
**Accuracy:** 93.37% on validation set  
**Training data:** 22,564 images  
**Training time:** ~6 hours on Colab GPU

## Example Results

| Organic Waste | Recyclable Waste |
|--------------|------------------|
| Food scraps  | Plastic bottles |
| Yard waste   | Glass jars      |
| Paper towels | Aluminum cans   |

## About the Creator

This model was created as part of an AI waste management project to help reduce improper waste disposal.

## License

Apache 2.0 - Free for personal and commercial use.