Update README - Run 20251012_050214

README.md

@@ -21,7 +21,7 @@ model-index:
       type: imagenet-1k
     metrics:
     - type: accuracy
-      value: 79.48
+      value: 80.82
 ---
 
 # David: Multi-Scale Crystal Classifier
@@ -51,12 +51,13 @@ as class prototypes with role-weighted similarity computation (Rose Loss).
 ## Performance
 
 ### Best Results
-- **Validation Accuracy**: 79.48%
-- **Best Epoch**: 0
-- **Final Train Accuracy**: 67.24%
+- **Validation Accuracy**: 80.82%
+- **Best Epoch**: 1
+- **Final Train Accuracy**: 82.27%
 
 ### Per-Scale Performance
-- **Scale 256**: 79.48%
+- **Scale 256**: 80.82%
+- **Scale 512**: 80.74%
 
 
 ## Usage
@@ -179,4 +180,4 @@ Special thanks to Claude (Anthropic) for debugging assistance.
 
 ---
 
-*Generated on 2025-10-12 05:04:37*
+*Generated on 2025-10-12 05:07:00*