Update README.md

README.md

@@ -36,4 +36,14 @@ configs:
     path: data/validation-*
   - split: test
     path: data/test-*
+license: cc-by-4.0
+language:
+- ja
 ---
+# MangaOCR Dataset
+
+## Dataset Details
+
+This is the MangaOCR dataset. We constructed this dataset by consolidating annotations from the [Manga109](http://www.manga109.org/) dataset and the [manga onomatopoeia dataset](https://github.com/ku21fan/COO-Comic-Onomatopoeia).
+
+It contains approximately **209K** narrative text instances, spanning a wide variety of visual styles and layouts.