Update README

app.py

@@ -237,11 +237,15 @@ with gr.Blocks(css=css, theme=gr.themes.Origin()) as demo:
         """
 <b>Official demo</b> for our working paper <b>AdaFace: A Versatile Face Encoder for Zero-Shot Diffusion Model Personalization</b>.<br>
 
+❗️**What's New**❗️
+- Support switching between two model styles: **Realistic** and **Anime**.
+- If you just changed the model style, the first image/video generation will take extra 20~30 seconds for loading new model weight.
+
 ❗️**Tips**❗️
 - You can upload one or more subject images for generating ID-specific video.
-- If the face dominates the video frames, try increasing the Weight of ID prompt in the negative prompt, at the cost of slight drop of ID authenticity.
-- If the face loses focu, try increasing the guidance scale. At the same time, increase the Weight of ID prompt in the negative prompt proportionally.
-- If the motion is weird, e.g., running, try increasing the number of sampling steps.
+- If the face dominates the video frames, try decreasing the guidance scale.
+- If the face loses focus, try increasing the guidance scale.
+- If the motion is weird, e.g., the prompt is "... running", try increasing the number of sampling steps.
 - Usage explanations and demos: [Readme](https://huggingface.co/spaces/adaface-neurips/adaface-animate/blob/main/README2.md).
 - AdaFace Text-to-Image: <a href="https://huggingface.co/spaces/adaface-neurips/adaface" style="display: inline-flex; align-items: center;">
   AdaFace