changing the text again

app.py

@@ -104,10 +104,11 @@ def predict(prompt, weights, seed, scale=7.5, steps=50):
 
 
 MESSAGE = '''
-Our method helps you achieve two amazing things:
+Our method helps you achieve three amazing things:
 
 1. Edit your generated images iteratively without damaging any important concepts.
 2. Generate any view of objects that the original Stable Diffusion implementation couldn't produce. For example, you can generate a "peacock, back view" by using "peacock, front view" as the negative prompt. Compare our method to [Stable Diffusion](https://huggingface.co/spaces/stabilityai/stable-diffusion).
+3. Alleviate the multihead problem in text-to-3D. Check out our work on this at [perp-neg.github.io](https://perp-neg.github.io/).
 
 To use our demo, simply enter your main prompt first, followed by a set of positive and negative prompts separated by "|". When only one prompt is provided and the weight of that prompt is 1, it is identical to using Stable Diffusion. We provided those as examples for the sake of comparison of our algorithm to Stable Diffusion. Put the weight of main prompt as 1.
 
@@ -118,7 +119,7 @@ MESSAGE_END = '''
 
 Unlike the original implementation, our method ensures that everything provided as the main prompt remains intact even when there is an overlap between the positive and negative prompts.
 
-We've also integrated the idea of robust view generation in text-to-3D to avoid the multihead problem. Check out our work on this at [perp-neg.github.io](https://perp-neg.github.io/).
+We've also integrated the idea of robust view generation in text-to-3D to avoid the multihead problem. For more details, please check out our work on this at [perp-neg.github.io](https://perp-neg.github.io/).
 
 '''