Update README.md

README.md

@@ -40,7 +40,7 @@ configs:
 <img src="WAON-Bench.jpg" alt="Overview of WAON-Bench" width="100%"/>
 
 WAON-Bench is a manually curated image classification dataset designed to benchmark Vision-Language models on Japanese culture.
-The dataset contains 374 classes across 8 categories: animals, buildings, events, everyday life, food, nature, scenery, and traditions. Each class includes 5 images.
+The dataset contains 374 classes across 8 categories (animals, buildings, events, everyday life, food, nature, scenery, and traditions), with 5 images per class, totaling 1,870 examples.
 
 ## How to Use