v1 fix

README.md

@@ -14,18 +14,18 @@ Based on validation results from the final training epochs:
 
 | Metric | Value |
 |--------|-------|
-| **Overall Validation Accuracy** | **94.2%** |
-| **Overall Validation Loss** | **0.0007** |
-| **Training Accuracy** | **92.7%** |
+| **Overall Validation Accuracy** | **97.1%** |
+| **Overall Validation Loss** | **0.0006** |
+| **Training Accuracy** | **95.6%** |
 
 ### Per-Format Accuracy
 
 | Format | Validation Acc | Training Acc | Quality Range |
 |--------|----------------|--------------|---------------|
-| **JPEG** | **99.6%** | 98.9% | 0-100 |
-| **WebP** | **97.2%** | 95.7% | 0-100 |
-| **AVIF** | **95.4%** | 93.3% | 0-63 |
-| **JXL** | **84.7%** | 83.1% | 0-25 |
+| **JPEG** | **99.4%** | 98.9% | 0-100 |
+| **WebP** | **97.0%** | 95.5% | 0-100 |
+| **AVIF** | **97.1%** | 95.3% | 0-100 |
+| **JXL** | **94.8%** | 92.6% | 0-100 |
 
 *Accuracy measured as predictions within ±5% range of actual quality values*
 
@@ -33,15 +33,15 @@ Based on validation results from the final training epochs:
 
 - **Multi-format Support**: Detects artifacts in JPEG, WebP, AVIF, and JXL formats
 - **Lightweight Architecture**: Only **~2M parameters** (~8MB model size)
-- **High Precision**: 99.6% on JPEG
+- **High Precision**: 99.4% on JPEG
 - **Fast Inference**: Optimized for real-time processing even on very weak HW using either CPU or GPU
 - **Format-agnostic**: Single model handles all compression types
 
 ## Limitations
 
-1. **JXL Performance**: Lower accuracy (84.7%) due to JXL's smaller quality range (0-25) and newer compression algorithm
-2. **Resolution Dependency**: Model trained on 512x512 crops; very small images may yield reduced accuracy
-3. **Quality Range**: Model assumes standard quality ranges; non-standard encoders may produce different results
+1. **Resolution Dependency**: Model trained on 512x512 crops; very small images may yield reduced accuracy
+2. **Quality Range**: Model assumes standard quality ranges; non-standard encoders may produce different results
+3. **Image Resizing**: Trying to detect compression artifacs in images which were compressed and than converted to lossless format and resized may fail or be much less accurate as these artifacts are washed by the resizing algorithms.
 
 ## Environmental Impact