Update README.md

README.md

@@ -14,7 +14,7 @@ As generative AI (like SDXL, DALL·E, and Midjourney) becomes more accessible, t
 
 ## 🚀 Model Overview  
 
-**SDXL-Deepfake-Detector** is a fine-tuned vision transformer that classifies human faces as **Real (0)** or **AI-Generated (1)**. Trained on the [140k Real and Fake Faces](https://www.kaggle.com/datasets/xhlulu/140k-real-and-fake-faces) dataset, it achieves **91% accuracy** on held-out test data and generalizes well across diverse synthesis methods.
+**SDXL-Deepfake-Detector** is a fine-tuned vision transformer that classifies human faces as **AI-Generated (0)** or **Real (1)**. Trained on the [140k Real and Fake Faces](https://www.kaggle.com/datasets/xhlulu/140k-real-and-fake-faces) dataset.
 
 ### ✅ Key Highlights
 - **Architecture**: Fine-tuned Vision Transformer (ViT) via Hugging Face `transformers`
@@ -32,6 +32,7 @@ pip install transformers torch pillow
 ```
 ### Python Script
 ```python
+#predict.py
 import argparse
 from transformers import AutoModelForImageClassification, AutoFeatureExtractor
 from PIL import Image
@@ -80,4 +81,59 @@ def main():
 
 if __name__ == "__main__":
     main()
-```
+```
+### How to use
+```bash
+python predict.py --image path/to/image
+```
+
+## 📊 Performance & Limitations
+
+> **Note**: Final test accuracy will be reported after full evaluation. Preliminary results show strong generalization on SDXL- and diffusion-based face forgeries.
+
+### Known Limitations
+- Trained primarily on **frontal, well-lit, aligned face crops** — may underperform on:
+  - Low-resolution or blurry images
+  - Heavily occluded or non-frontal faces
+  - GAN-generated faces (e.g., StyleGAN2/3)
+- Label mapping:  
+  - `0` → `"artificial"` (AI-generated / Deepfake)  
+  - `1` → `"real"` (authentic human face)
+
+> ⚠️ This tool is **not a forensic proof**, but a probabilistic detector. Use responsibly.
+
+---
+
+## 🌱 Philosophy & Ethics
+
+This model is open-source because:
+- **Transparency** is essential in the fight against synthetic media.
+- **Accessibility** ensures researchers, journalists, and civil society can audit and use detection tools without gatekeeping.
+- **Privacy matters**: The model runs **entirely offline** — your images never leave your device.
+
+As a developer from a vulnerable community, I believe AI safety tools must be **inclusive, ethical, and human-centered** — not just technically accurate.
+
+---
+
+## 🙌 Acknowledgements
+
+- **Dataset**: [140k Real and Fake Faces](https://www.kaggle.com/datasets/xhlulu/140k-real-and-fake-faces) by xhlulu  
+- **Framework**: [Hugging Face Transformers](https://huggingface.co/docs/transformers)  
+- **Model & Code**: [GitHub Repository](https://github.com/SadraCoding/SDXL-Deepfake-Detector) | [Hugging Face Hub](https://huggingface.co/SADRACODING/SDXL-Deepfake-Detector)
+
+---
+
+## 📬 How to Contribute
+
+I welcome:
+- Bug reports and feature requests (via GitHub Issues)
+- Expanding support to **video deepfakes** or **GAN-generated faces**
+- **Bias/fairness audits** across gender, skin tone, and age
+- Multilingual documentation
+
+> 💡 **Tip for researchers**: Fine-tune this model on your domain-specific data using Hugging Face `Trainer`.
+
+---
+
+> *Built with curiosity, ethics, and a 12GB GPU — because impactful AI doesn’t require a data center, just purpose.*  
+> — Sadra Milani Moghaddam