added note on avoiding NSFW

Finally had time to look into SDXL's filtering. They're sparse on documentation but it seems like people say it's filtered well enough that we really don't need the SD1.5 safety filter. On the other hand, since we're using the SDXLightning version, which has even less info available, I'm adding the note just for caution.

app.py

@@ -279,6 +279,8 @@ with gr.Blocks(css=css, head=js_head) as demo:
                  [b4, embs, ys, calibrate_prompts],
                  [b1, b2, b3, b4, img, embs, ys, calibrate_prompts])
     with gr.Row():
-        html = gr.HTML('''<div style='text-align:center; font-size:20px'>You will calibrate for several prompts and then roam.</ div>''')
+        html = gr.HTML('''<div style='text-align:center; font-size:20px'>You will calibrate for several prompts and then roam. </ div><br><br><br>
+<div style='text-align:center; font-size:14px'>Note that while the SDXL model is unlikely to produce NSFW images, it still may be possible, and users should avoid NSFW content when rating.
+</ div>''')
 
 demo.launch()  # Share your demo with just 1 extra parameter 🚀