Update README.md

README.md

@@ -46,7 +46,9 @@ def _transform(n_px):
 
 See the documentation for [`CLIPImageProcessor` for details](https://huggingface.co/docs/transformers/model_doc/clip#transformers.CLIPImageProcessor).
 
-Also, despite the names `style vector` and `content vector`, I have noticed by visual inspection that both are basically equally good for style embedding. I don't know why, but I guess that's life?
+Also, despite the names `style vector` and `content vector`, I have noticed by visual inspection that both are basically equally good for style embedding. I don't know why, but I guess that's life? (No, it's actually not supposed to happen. I don't know why it didn't really disentangle style and content. Maybe that's a question for a small research paper.)
+
+You can see for yourself by changing the line `style_output = output["style_output"].squeeze(0)` to `style_output = output["content_output"].squeeze(0)` in the demo. The resulting t-SNE is still clustering by style, to my eyes equally well.
 
 ## How to use it