---
license: mit
tags:
- cybersecurity
- threat-detection
- malware
- phishing
- anomaly-detection
library_name: sklearn
pipeline_tag: tabular-classification
---

# CyberForge AI Security Models

Production-ready machine learning models for real-time cybersecurity threat detection.

## Models Included

| Model | Task | Accuracy | F1 Score | Inference |
|-------|------|----------|----------|-----------|
| phishing_detection | Detect phishing URLs | 98.9% | 0.989 | 0.02ms |
| malware_detection | Identify malware | 99.8% | 0.998 | 0.001ms |
| anomaly_detection | Network anomalies | 99.9% | 0.999 | 0.007ms |
| web_attack_detection | Web attacks | 100% | 1.000 | 0.03ms |

## Quick Start

```python
from cyberforge_inference import CyberForgePredictor

# Initialize
predictor = CyberForgePredictor()

# Predict
result = predictor.predict("phishing_detection", features)
print(f"Threat: {result['prediction']}, Confidence: {result['confidence']}")
```

## API Usage

```python
import requests

response = requests.post(
    "https://your-api-endpoint/predict",
    json={
        "model": "phishing_detection",
        "features": {...}
    }
)
```

## Features

- **Real-time inference** < 1ms per prediction
- **Multiple threat types**: Phishing, Malware, Anomalies, Web Attacks
- **Production-ready**: Optimized for high-throughput
- **Backend integration**: Compatible with Node.js/Python backends

## Training Data

Models trained on 50,000+ samples from:
- [CyberForge Datasets](https://huggingface.co/datasets/Che237/cyberforge-datasets)

## License

MIT License - Free for commercial and personal use.