---
license: mit
sdk: gradio
colorFrom: blue
colorTo: green
short_description: Recycling Waste Classifier (12 Classes)
sdk_version: 5.34.2
---
# Recycling Waste Classifier (12 Classes)

This is a Gradio-powered web app that classifies images of waste into 12 categories:  
**battery, biological, brown-glass, cardboard, clothes, green-glass, metal, paper, plastic, shoes, trash, white-glass**.

Upload a photo of a waste item and the model will predict the most likely material class, helping you recycle more efficiently.

## How to use
- Click "Upload" and select an image of a waste item.
- The model will display the top 3 class predictions and probabilities.

## Model Info
- **Architecture:** ResNet34
- **Framework:** fastai (PyTorch backend)
- **Dataset:** [Kaggle - Garbage Classification](https://www.kaggle.com/datasets/mostafaabla/garbage-classification)
- **Accuracy:** ~97% on validation set

## Example Categories
- battery, biological, brown-glass, cardboard, clothes, green-glass, metal, paper, plastic, shoes, trash, white-glass

## Credits
- Dataset: Kaggle Garbage Classification
- Model: fastai, PyTorch, Gradio