Update app.py

app.py

@@ -83,35 +83,68 @@ def generate_audio(text, length,
     return params['autoencoder']['sr'], pred
 
 
-# Gradio Interface
-def gradio_interface():
-    # Input components
-    text_input = gr.Textbox(label="Text Prompt", value="the sound of dog barking")
-    length_input = gr.Slider(minimum=1, maximum=10, step=1, value=10, label="Audio Length (in seconds)")
-
-    # Advanced settings
-    guidance_scale_input = gr.Slider(minimum=1.0, maximum=10, step=0.1, value=5, label="Guidance Scale")
-    guidance_rescale_input = gr.Slider(minimum=0.0, maximum=1, step=0.05, value=0.75, label="Guidance Rescale")
-    ddim_steps_input = gr.Slider(minimum=25, maximum=200, step=5, value=100, label="DDIM Steps")
-    eta_input = gr.Slider(minimum=0.0, maximum=1.0, step=0.1, value=1, label="Eta")
-    random_seed_input = gr.Slider(minimum=0, maximum=MAX_SEED, step=1, value=0,)
-
-    randomize_seed = gr.Checkbox(label="Randomize seed", value=False)
-
-    # Output component
-    output_audio = gr.Audio(label="Converted Audio", type="numpy")
-
-    # Interface
-    gr.Interface(
-        fn=generate_audio,
-        inputs=[text_input, length_input, guidance_scale_input, guidance_rescale_input, ddim_steps_input, eta_input,
-                random_seed_input, randomize_seed],
-        outputs=output_audio,
-        title="EzAudio Text-to-Audio Generator",
-        description="Generate audio from text using a diffusion model. Adjust advanced settings for more control.",
-        allow_flagging="never"
-    ).launch()
-
-
-if __name__ == "__main__":
-    gradio_interface()
+# Examples (if needed for the demo)
+examples = [
+    "the sound of rain falling softly",
+    "a dog barking in the distance",
+    "light guitar music is playing",
+]
+
+# CSS styling (optional)
+css = """
+#col-container {
+    margin: 0 auto;
+    max-width: 1280px;
+}
+"""
+
+# Gradio Blocks layout
+with gr.Blocks(css=css, theme=gr.themes.Soft()) as demo:
+    with gr.Column(elem_id="col-container"):
+        gr.Markdown("""
+        # EzAudio Text-to-Audio Generator
+        Generate audio from text using a diffusion transformer. Adjust advanced settings for more control.
+        """)
+
+        # Basic Input: Text prompt and Audio Length
+        with gr.Row():
+            text_input = gr.Textbox(
+                label="Text Prompt",
+                show_label=False,
+                max_lines=2,
+                placeholder="Enter your prompt",
+                container=False,
+                value="a dog barking in the distance"
+            )
+            length_input = gr.Slider(minimum=1, maximum=10, step=1, value=10, label="Audio Length (in seconds)")
+
+        # Output Component
+        result = gr.Audio(label="Result", type="numpy")
+
+        # Advanced settings in an Accordion
+        with gr.Accordion("Advanced Settings", open=False):
+            guidance_scale = gr.Slider(minimum=1.0, maximum=10, step=0.1, value=5.0, label="Guidance Scale")
+            guidance_rescale = gr.Slider(minimum=0.0, maximum=1, step=0.05, value=0.75, label="Guidance Rescale")
+            ddim_steps = gr.Slider(minimum=25, maximum=200, step=5, value=100, label="DDIM Steps")
+            eta = gr.Slider(minimum=0.0, maximum=1.0, step=0.1, value=1.0, label="Eta")
+            seed = gr.Slider(minimum=0, maximum=MAX_SEED, step=1, value=0, label="Seed")
+            randomize_seed = gr.Checkbox(label="Randomize Seed", value=False)
+
+        # Examples block
+        gr.Examples(
+            examples=examples,
+            inputs=[text_input]
+        )
+
+        # Run button
+        run_button = gr.Button("Generate")
+
+        # Define the trigger and input-output linking
+        run_button.click(
+            fn=generate_audio,
+            inputs=[text_input, length_input, guidance_scale, guidance_rescale, ddim_steps, eta, seed, randomize_seed],
+            outputs=[result]
+        )
+
+# Launch the Gradio demo
+demo.launch()