Upload app.py

app.py

@@ -152,14 +152,23 @@ with gr.Blocks() as demo:
 
     gr.Markdown("""
     ---
+
+    ### Purpose
+    StyleSquirrel is designed for **style exploration**, not artist identification.  
+    It **may not report the image's artist**, even when that artist exists in the dataset.  
+    Use it to explore images with similar **colors, structures, textures, and visual motifs**, not as an attribution tool.
+
     ### Instructions
     - Drop an image in the box on the left.
-    - The "Style Tags" panel reports on tags that are stylistically similar to the query image.
-    - The "Nearest Neighbors" panel reports on e621 images that are stylistically similar to the query image.
+    - The **Style Tags** panel reports tags that are stylistically similar to the query image.
+    - The **Nearest Neighbors** panel shows e621 images that are stylistically similar to the query image.
+
     ### Notes
     - Links go to e621.net and may not be safe for work.
-    - I tried to isolate style from topic and was only partly successful.  So many reported tags and images might be topically rather than stylistically similar.
-    - The similarity metric is currently a bit naive, leading to irregularities like the "simple_background" tag being overreported due to its frequency.
+    - I tried to isolate style from topic and was only partly successful, so many reported tags/images might be topically rather than stylistically similar.
+    - The similarity metric is currently a bit naive, causing irregularities like **simple_background** being over-reported.
+    - This tool has a very small training set and does **not** work well as a style classifier.
+
     """)