Mirror the generator's title, theme and advanced-options layout

app.py

@@ -52,6 +52,7 @@ CANVASES = {
     "544x960 · 9:16 fast": (960, 544),
     "640x1152 · 9:16": (1152, 640),
     "768x1344 · 9:16 full": (1344, 768),
+    "544x544 · 1:1 fast": (544, 544),
     "768x768 · 1:1 full": (768, 768),
     "1024x768 · 4:3 full": (768, 1024),
     "768x1024 · 3:4 full": (1024, 768),
@@ -311,15 +312,15 @@ def _generate(prompt_embeds, text_token_tags, references, height, width, num_fra
 
 def generate(
     prompt,
-    video_path,
-    first_image_path,
-    second_image_path,
-    audio_path,
-    canvas,
-    match,
-    duration,
-    steps,
-    seed,
+    video_path=None,
+    first_image_path=None,
+    second_image_path=None,
+    audio_path=None,
+    canvas=DEFAULT_CANVAS,
+    match=True,
+    duration=5,
+    steps=28,
+    seed=42,
     progress=gr.Progress(track_tqdm=True),
 ):
     if LOAD_ERROR:
@@ -372,11 +373,22 @@ def generate(
 
 load_models()
 
-INTRO = """# MiniMax-H3 · references
+INTRO = """# MiniMax-H3
 
-Bring a subject, a style, a camera move or a voice, and generate video with a synchronized soundtrack in one pass.
-Fill as many tabs as you like — references are read **in tab order**, video then images then audio, and that order is
-part of the request.
+<div align="center">
+  <a href="https://huggingface.co/MiniMaxAI/MiniMax-H3"><strong>[ model ]</strong></a> &nbsp;
+  <a href="PAPER_URL_PLACEHOLDER"><strong>[ paper ]</strong></a> &nbsp;
+  <a href="https://www.minimax.io"><strong>[ project ]</strong></a>
+</div>
+
+**MiniMax-H3** is a 33B parameter state of the art video generation model that produces video and a
+fully synchronized soundtrack (ambience, foley, speech). Bring your own subject, camera move or voice as a
+reference — they are read **in tab order**, video then images then audio.
+"""
+
+CSS = """
+#col-container { max-width: 1180px; margin: 0 auto; }
+.dark .gradio-container { color: var(--body-text-color); }
 """
 
 with gr.Blocks(title="MiniMax-H3 references") as demo:
@@ -400,14 +412,16 @@ with gr.Blocks(title="MiniMax-H3 references") as demo:
                         second_image = gr.Image(label="Subject, style or scene", type="filepath")
                 with gr.Tab("Audio"):
                     audio = gr.Audio(label="A voice or a piece of music", type="filepath")
-            canvas = gr.Dropdown(label="Canvas", choices=list(CANVASES), value=DEFAULT_CANVAS)
-            match = gr.Checkbox(label="Match the reference soundtrack", value=True, visible=False)
-            duration = gr.Slider(
-                label="Duration (s)", minimum=MIN_DURATION, maximum=MAX_UI_DURATION, step=1, value=5
-            )
-            steps = gr.Slider(label="Steps", minimum=10, maximum=40, step=1, value=28)
-            seed = gr.Number(label="Seed", value=42, precision=0)
             run = gr.Button("Generate", variant="primary")
+            with gr.Accordion("Advanced options", open=False):
+                canvas = gr.Dropdown(label="Canvas", choices=list(CANVASES), value=DEFAULT_CANVAS)
+                match = gr.Checkbox(label="Match the reference soundtrack", value=True, visible=False)
+                duration = gr.Slider(
+                    label="Duration (s)", minimum=MIN_DURATION, maximum=MAX_UI_DURATION, step=1, value=5
+                )
+                steps = gr.Slider(label="Steps", minimum=10, maximum=40, step=1, value=28)
+                seed = gr.Number(label="Seed", value=42, precision=0)
+
         with gr.Column():
             result = gr.Video(label="Video + soundtrack")
 
@@ -426,4 +440,4 @@ with gr.Blocks(title="MiniMax-H3 references") as demo:
 
 
 if __name__ == "__main__":
-    demo.queue(max_size=4).launch(show_error=True)
+    demo.launch(show_error=True, theme=gr.themes.Citrus(), css=CSS)