# 🌍 Disaster Prediction Model

This repository contains a deep learning model for **disaster classification** from images, capable of identifying six disaster-related categories including fire, water damage, infrastructure damage, and more. The model is built using **ResNet50** and trained for image classification tasks.

## 🧠 Model Details

- **Architecture**: ResNet50
- **Trained On**: Kaggle Disaster Dataset
- **Image Size**: 256x256
- **Input**: RGB image
- **Output**: Disaster category
- **License**: MIT
- **Pipeline Tag**: `image-classification`
- **Main Metric**: Accuracy

## 📦 Installation & Cloning

```bash
# Install Git LFS (if not already installed)
git lfs install

# Clone the repository
git clone https://huggingface.co/Luwayy/disaster-prediction
```

## 🔍 Classes

The model predicts one of the following disaster categories:

| ID | Class Name              |
|----|-------------------------|
| 0  | Damaged_Infrastructure |
| 1  | Fire_Disaster           |
| 2  | Human_Damage            |
| 3  | Land_Disaster           |
| 4  | Non_Damage              |
| 5  | Water_Disaster          |

## 📸 Example Usage (Python)

```python
import keras
import numpy as np
from PIL import Image
import requests
from io import BytesIO

# Load the model
model = keras.layers.TFSMLayer(
    "disaster-prediction/kaggle/working/disaster_model",
    call_endpoint="serving_default"
)

# Load and preprocess image
url = 'https://www.spml.co.in/Images/blog/wdt&c-152776632.jpg'
response = requests.get(url)
img = Image.open(BytesIO(response.content)).convert("RGB").resize((256, 256))
img_array = np.array(img) / 255.0
img_array = np.expand_dims(img_array, axis=0).astype(np.float32)

# Predict
output = model(img_array)
preds = list(output.values())[0].numpy()
pred_index = np.argmax(preds)

# Class labels
labels = [
    "Damaged_Infrastructure",
    "Fire_Disaster",
    "Human_Damage",
    "Land_Disaster",
    "Non_Damage",
    "Water_Disaster"
]

print("Predicted class:", labels[pred_index])
```

## ⚙️ Preprocessing

- **Resize**: (256, 256)
- **Scale**: Normalize pixel values by dividing by 255