Create README.md

README.md

@@ -1,17 +1,25 @@
----
-dataset_info:
-  features:
-  - name: image
-    dtype: image
-  - name: text
-    dtype: string
-  splits:
-  - name: train
-    num_bytes: 222190761.8
-    num_examples: 2475
-  download_size: 232442024
-  dataset_size: 222190761.8
----
-# Dataset Card for "Dungeons-and-Diffusion"
-
-[More Information needed](https://github.com/huggingface/datasets/blob/main/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
+This dataset includes ~2500 images of fantasy RPG character art. This dataset has a distribution of races and classes, though only races are annotated right now.
+Additionally, BLIP captions were generated for all examples. 
+Thus, there are two datasets- one with the human generated race annotation formatted as 'D&D Character, {race}'
+BLIP captions are formatted as 'D&D Character, {race} {caption}' for example: 'D&D Character, drow a woman with horns and horns'
+Distribution of races:
+({'kenku': 31,
+  'drow': 162,
+  'tiefling': 285,
+  'dwarf': 116,
+  'dragonborn': 110,
+  'gnome': 72,
+  'orc': 184,
+  'aasimar': 74,
+  'kobold': 61,
+  'aarakocra': 24,
+  'tabaxi': 123,
+  'genasi': 126,
+  'human': 652,
+  'elf': 190,
+  'goblin': 80,
+  'halfling': 52,
+  'centaur': 22,
+  'firbolg': 76,
+  'goliath': 35})
+There is a high chance some images are mislabelled! Please feel free to enrich this dataset with whatever attributes you think might be useful!