app.py

import gradio as gr

with gr.Blocks(title="STAR Online Inference", theme=gr.themes.Soft()) as demo:
    gr.Markdown("# STAR: Speech-to-Audio Generation via Representation Learning")
    
    gr.Markdown("""
<div style="text-align: left; padding: 10px;">
## 🗣️ Input

A brief input speech utterance for the overall audio scene.  

> Example：A cat meowing and young female speaking

#### 🎙️ Input Speech Example

#### 🎧️ Output Audio Example
</div>
---
</div>
""")