Update README.md

README.md

@@ -7,13 +7,27 @@ dataset_info:
     dtype: string
   splits:
   - name: train
-    num_bytes: 87942822.0
+    num_bytes: 87942822
     num_examples: 760
   download_size: 87292244
-  dataset_size: 87942822.0
+  dataset_size: 87942822
 configs:
 - config_name: default
   data_files:
   - split: train
     path: data/train-*
+license: mit
+tags:
+- art
+size_categories:
+- n<1K
 ---
+
+
+# Dataset Summary
+Contains characters from One Piece animation.
+Used to finetune Stable Diffusion LoRA weights
+
+# Dataset Creation
+All data were scraped from the web (Google Image search).\
+In order to filter irrelevant images, CLIP score was calculated and the bottom 20% of the dataset was discarded.