feat: Make app more informative

app.py

@@ -157,13 +157,36 @@ def generate_caption(audio_file):
     return outputs
 
 
-demo = gr.Interface(
-    fn=caption_audio,
-    inputs=gr.Audio(type="filepath", label="Upload an audio file"),
-    outputs=gr.Textbox(label="Generated Caption"),
-    title="SonicVerse",
-    description="Upload an audio file to generate a caption using SonicVerse"
-)
+with gr.Blocks(title="SonicVerse") as demo:
+    gr.Markdown("""
+        # 🎼 SonicVerse: Music Captioning Demo
+
+        Welcome to **SonicVerse**, a multi-task music captioning model that provides natural language descriptions of input clips.
+
+        🎵 Captions include music features such as:
+        - Genre
+        - Mood
+        - Instrumentation
+        - Vocals
+        - Key
+
+        📘 [Read the Paper](https://arxiv.org/abs/2506.15154)
+
+         🖥️ [Replicate locally](https://github.com/amaai-lab/SonicVerse)
+
+
+        ⚠️ **Note:** You can upload audio of any length, but due to compute limits on Hugging Face Spaces,
+        it is recommended to keep clips under 30 seconds unless you have a Pro account or run this locally.
+    """)
+
+    with gr.Row():
+        audio_input = gr.Audio(type="filepath", label="Upload your music clip")
+        caption_output = gr.Textbox(label="Generated Caption", lines=8)
+
+    submit_btn = gr.Button("Generate Caption")
+    submit_btn.click(fn=caption_audio, inputs=audio_input, outputs=caption_output)
+
+
 
 if __name__ == "__main__":
     demo.launch()