Update README - Run 20251012_060013

README.md

@@ -21,7 +21,7 @@ model-index:
       type: imagenet-1k
     metrics:
     - type: accuracy
-      value: 82.98
+      value: 83.04
 ---
 
 # David: Multi-Scale Crystal Classifier
@@ -51,19 +51,21 @@ as class prototypes with role-weighted similarity computation (Rose Loss).
 ## Performance
 
 ### Best Results
-- **Validation Accuracy**: 82.98%
-- **Best Epoch**: 7
-- **Final Train Accuracy**: 89.95%
+- **Validation Accuracy**: 83.04%
+- **Best Epoch**: 9
+- **Final Train Accuracy**: 91.00%
 
 ### Per-Scale Performance
-- **Scale 256**: 82.98%
-- **Scale 512**: 83.15%
-- **Scale 768**: 83.09%
-- **Scale 1024**: 83.08%
-- **Scale 1280**: 82.99%
-- **Scale 1536**: 82.93%
-- **Scale 1792**: 82.88%
-- **Scale 2048**: 82.67%
+- **Scale 256**: 83.04%
+- **Scale 512**: 83.12%
+- **Scale 768**: 83.20%
+- **Scale 1024**: 83.21%
+- **Scale 1280**: 83.25%
+- **Scale 1536**: 83.13%
+- **Scale 1792**: 83.15%
+- **Scale 2048**: 83.14%
+- **Scale 2304**: 82.96%
+- **Scale 2560**: 82.71%
 
 
 ## Usage
@@ -186,4 +188,4 @@ Special thanks to Claude (Anthropic) for debugging assistance.
 
 ---
 
-*Generated on 2025-10-12 06:35:55*
+*Generated on 2025-10-12 06:40:29*