Update README.md

README.md

@@ -1,20 +1,25 @@
----
-license: apache-2.0
-dataset_info:
-  features:
-  - name: image
-    dtype: image
-  - name: text
-    dtype: string
-  splits:
-  - name: train
-    num_bytes: 98326877.0
-    num_examples: 140
-  download_size: 98322766
-  dataset_size: 98326877.0
-configs:
-- config_name: default
-  data_files:
-  - split: train
-    path: data/train-*
----
+---
+license: apache-2.0
+dataset_info:
+  features:
+  - name: image
+    dtype: image
+  - name: text
+    dtype: string
+  splits:
+  - name: train
+    num_bytes: 98326877.0
+    num_examples: 140
+  download_size: 98322766
+  dataset_size: 98326877.0
+configs:
+- config_name: default
+  data_files:
+  - split: train
+    path: data/train-*
+---
+
+#
+
+This dataset was automatically annotated using the florence 2 vision model with Fluxgym. This data which then was put some rounds of human annotation to ensure all image description pairs had appropriate triggerwords and descriptions:
+https://github.com/cocktailpeanut/fluxgym