Update README.md

README.md

@@ -17,36 +17,75 @@ Made for the **Forbidden Vision** ComfyUI custom nodes
 
 ---
 
-## Dataset
+## 🎯 Why These Models Exist
 
-All three models share a core mixed-domain dataset specifically curated for diffusion-generated content:
+Traditional face models fail where it matters most for AI art workflows:
 
-- images from CivitAI diffusion outputs (SDXL, SD1.5, Pony, Illustrious)
-- curated images from Danbooru (mixed anime styles)
-- real photographs from various sources
-- NSFW images from each domain without filtering
+| **Problem** | **Why It Matters** |
+|-------------|-------------------|
+| 🎨 **Domain-locked** | Existing models excel at *either* anime *or* realistic—never both |
+| 🔞 **NSFW blindness** | Most models trained only on SFW data break on adult content |
+| 🎲 **Generation artifacts** | Standard datasets don't include diffusion model quirks and failures |
+| 🔄 **No rotation handling** | Can't automatically correct face angles across mixed styles |
 
-**Total dataset size:** ~11k manually annotated images
-
-This mixed approach ensures the models work reliably across:
-- ✓ Both realistic and anime art styles
-- ✓ Difficult angles, occlusions, and expressions  
-- ✓ Low-quality generations and artifacts
-- ✓ SFW and NSFW content equally
-- ✓ Edge cases that break traditional face detectors
+**These models solve all four.**
 
 ---
 
+## 📊 Training Foundation
+
+### The Dataset Difference
+
+Built from **11,000+ manually annotated images** across the domains that actually matter for AI generation:
+
+<table>
+<tr>
+<td width="50%">
+
+**🎨 Multi-Domain Coverage**
+- SDXL, SD1.5, Pony, Illustrious outputs
+- Curated Danbooru (anime styles)
+- Real photography
+- Full NSFW inclusion (no filtering)
+
+</td>
+<td width="50%">
+
+**💎 Edge Case Priority**
+- ✓ Extreme angles & occlusions
+- ✓ Failed/broken generations
+- ✓ Low-quality artifacts
+- ✓ Unusual expressions & poses
+- ✓ Everything other models ignore
+
+</td>
+</tr>
+</table>
+
+### What This Means For You
+
+```
+Traditional models: Trained on clean celebrity faces
+         ↓
+    Fail on real workflows
+
+These models: Trained on what you actually generate
+         ↓
+    Work when you need them
+```
+
+**One model family. Every domain. Zero compromises.**
+
 ## Model Details
 
 ### Face Detection (YOLOv11-Small)
 
-**Purpose:** Primary face detection and bounding box localization
+**Purpose:** Primary face detection with high recall and very tight face boxes
 
 **Training Approach:**
-- [TRAINING DETAILS PLACEHOLDER]
+- After every training run, I ran the model on a new mixed dataset, hardmining failures and improving the dataset until an acceptable performance was reached
+- Used offline custom augmentation on the initial set to complement light yolo training script augmentations
 - Trained at 640px resolution (inference should use same resolution)
-- [AUGMENTATION DETAILS PLACEHOLDER]
 
 **Why YOLOv11-Small instead of nano?**  
 More reliable detection across mixed realistic/anime domains with acceptable speed tradeoff.