Update README - Run 20251012_050214

README.md

@@ -21,7 +21,7 @@ model-index:
       type: imagenet-1k
     metrics:
     - type: accuracy
-      value: 81.35
+      value: 81.55
 ---
 
 # David: Multi-Scale Crystal Classifier
@@ -51,14 +51,15 @@ as class prototypes with role-weighted similarity computation (Rose Loss).
 ## Performance
 
 ### Best Results
-- **Validation Accuracy**: 81.35%
-- **Best Epoch**: 2
-- **Final Train Accuracy**: 84.13%
+- **Validation Accuracy**: 81.55%
+- **Best Epoch**: 3
+- **Final Train Accuracy**: 85.34%
 
 ### Per-Scale Performance
-- **Scale 256**: 81.35%
-- **Scale 512**: 81.25%
-- **Scale 768**: 80.87%
+- **Scale 256**: 81.55%
+- **Scale 512**: 81.49%
+- **Scale 768**: 81.39%
+- **Scale 1024**: 80.92%
 
 
 ## Usage
@@ -181,4 +182,4 @@ Special thanks to Claude (Anthropic) for debugging assistance.
 
 ---
 
-*Generated on 2025-10-12 05:09:19*
+*Generated on 2025-10-12 05:11:38*