Update README.md

README.md

@@ -1,3 +1,30 @@
----
-license: mit
----
+# 🎭 SDXL-Deepfake-Detector
+
+**A high-performance deep learning model fine-tuned for the binary classification of real vs. fake faces, specifically targeting synthesized images, including those potentially generated by powerful models like SDXL.**
+
+This project was developed by **[Sadra Milani Moghadam](https://sadramilani.ir)**.
+
+---
+
+## 🚀 Model Description
+
+The **SDXL-Deepfake-Detector** is a specialized image classification model designed to distinguish between authentic human faces and synthetically generated (deepfake) faces. It leverages the power of transfer learning to provide robust, high-accuracy detection.
+
+### Key Features & Performance
+
+| Feature | Value / Metric | Notes |
+| :--- | :--- | :--- |
+| **Task** | Binary Image Classification | Real Face vs. Deepfake Face |
+| **Dataset Used** | [140K Real and Fake Faces (Kaggle)](https://www.kaggle.com/datasets/xhlulu/140k-real-and-fake-faces) | Comprehensive dataset of 140,000 high-quality images. |
+| **Test Accuracy** | **0.91** (91%) | Demonstrates strong performance in distinguishing image authenticity. |
+| **Hardware Used** | NVIDIA RTX 3060 (12GB VRAM) | Optimized training time due to suitable GPU hardware. |
+
+---
+
+## 💻 Usage with Hugging Face Transformers
+
+You can easily load and run this model for inference using the Hugging Face `transformers` library in Python.
+
+---
+license: mit
+---