Update README - Run 20251012_132646

README.md

@@ -21,7 +21,7 @@ model-index:
       type: imagenet-1k
     metrics:
     - type: accuracy
-      value: 69.51
+      value: 70.13
 ---
 
 # David: Multi-Scale Crystal Classifier
@@ -51,13 +51,13 @@ as class prototypes with role-weighted similarity computation (Rose Loss).
 ## Performance
 
 ### Best Results
-- **Validation Accuracy**: 69.51%
-- **Best Epoch**: 1
-- **Final Train Accuracy**: 69.52%
+- **Validation Accuracy**: 70.13%
+- **Best Epoch**: 2
+- **Final Train Accuracy**: 70.31%
 
 ### Per-Scale Performance
-- **Scale 256**: 68.85%
-- **Scale 512**: 69.34%
+- **Scale 256**: 69.58%
+- **Scale 512**: 69.88%
 
 
 ## Usage
@@ -191,4 +191,4 @@ Special thanks to Claude (Anthropic) for debugging assistance.
 
 ---
 
-*Generated on 2025-10-12 13:31:43*
+*Generated on 2025-10-12 13:33:22*