Update app.py

app.py

@@ -221,6 +221,11 @@ if __name__ == "__main__":
         checkbox_label_text_color_selected="*button_primary_text_color",
     )
     css = """
+        #col-container {
+            margin: 0 auto;
+            max-width: 1024px;
+        }
+
         .custom-log * {
             font-style: italic;
             font-size: 22px !important;
@@ -259,11 +264,6 @@ if __name__ == "__main__":
         }
         """
     with gr.Blocks(css=css, title="AnySplat Demo", theme=theme) as demo:
-        gr.Markdown(
-            """
-            <h1 style='text-align: center;'>AnySplat: Feed-forward 3D Gaussian Splatting from Unconstrained Views</h1>
-            """
-        )
 
         gr.Markdown(
             """ # AnySplat – Feed-forward 3D Gaussian Splatting from Unconstrained Views
@@ -279,128 +279,129 @@ if __name__ == "__main__":
         scene_name = gr.Textbox(label="scene_name", visible=False, value="None")
         image_type = gr.Textbox(label="image_type", visible=False, value="None")
 
-        with gr.Row():
-            with gr.Column():
-                input_video = gr.Video(label="Upload Video", interactive=True)
-                input_images = gr.File(
-                    file_count="multiple",
-                    label="Upload Images",
-                    interactive=True,
-                )
-    
-                image_gallery = gr.Gallery(
-                    label="Preview",
-                    columns=4,
-                    height="300px",
-                    show_download_button=True,
-                    object_fit="contain",
-                    preview=True,
-                )
-
-
-            with gr.Column():
+        with gr.Column(elem_id="col-container"):
+            with gr.Row():
                 with gr.Column():
-                    reconstruction_output = gr.Model3D(
-                        label="3D Reconstructed Gaussian Splat",
-                        height=540,
-                        zoom_speed=0.5,
-                        pan_speed=0.5,
-                        camera_position=[20, 20, 20],
+                    input_video = gr.Video(label="Upload Video", interactive=True)
+                    input_images = gr.File(
+                        file_count="multiple",
+                        label="Upload Images",
+                        interactive=True,
+                    )
+        
+                    image_gallery = gr.Gallery(
+                        label="Preview",
+                        columns=4,
+                        height="300px",
+                        show_download_button=True,
+                        object_fit="contain",
+                        preview=True,
                     )
     
-                with gr.Row():
+    
+                with gr.Column():
+                    with gr.Column():
+                        reconstruction_output = gr.Model3D(
+                            label="3D Reconstructed Gaussian Splat",
+                            height=540,
+                            zoom_speed=0.5,
+                            pan_speed=0.5,
+                            camera_position=[20, 20, 20],
+                        )
+        
+                    with gr.Row():
+                        with gr.Row():
+                            rgb_video = gr.Video(
+                                label="RGB Video", interactive=False, autoplay=True
+                            )
+                            depth_video = gr.Video(
+                                label="Depth Video",
+                                interactive=False,
+                                autoplay=True,
+                            )
+        
                     with gr.Row():
-                        rgb_video = gr.Video(
-                            label="RGB Video", interactive=False, autoplay=True
+                        submit_btn = gr.Button(
+                            "Reconstruct", scale=1, variant="primary"
                         )
-                        depth_video = gr.Video(
-                            label="Depth Video",
-                            interactive=False,
-                            autoplay=True,
+                        clear_btn = gr.ClearButton(
+                            [
+                                input_video,
+                                input_images,
+                                reconstruction_output,
+                                target_dir_output,
+                                image_gallery,
+                                rgb_video,
+                                depth_video,
+                            ],
+                            scale=1,
                         )
-    
-                with gr.Row():
-                    submit_btn = gr.Button(
-                        "Reconstruct", scale=1, variant="primary"
-                    )
-                    clear_btn = gr.ClearButton(
-                        [
-                            input_video,
-                            input_images,
-                            reconstruction_output,
-                            target_dir_output,
-                            image_gallery,
-                            rgb_video,
-                            depth_video,
-                        ],
-                        scale=1,
-                    )
 
-        # ---------------------- Examples section ----------------------
-
-        examples = [
-            [None, "examples/video/re10k_1eca36ec55b88fe4.mp4", "re10k", "1eca36ec55b88fe4", "2", "Real", "True",],
-            [None, "examples/video/bungeenerf_colosseum.mp4", "bungeenerf", "colosseum", "8", "Synthetic", "True",],
-            [None, "examples/video/fox.mp4", "InstantNGP", "fox", "14", "Real", "True",],
-            [None, "examples/video/matrixcity_street.mp4", "matrixcity", "street", "32", "Synthetic", "True",],
-            [None, "examples/video/vrnerf_apartment.mp4", "vrnerf", "apartment", "32", "Real", "True",],
-            [None, "examples/video/vrnerf_kitchen.mp4", "vrnerf", "kitchen", "17", "Real", "True",],
-            [None, "examples/video/vrnerf_riverview.mp4", "vrnerf", "riverview", "12", "Real", "True",],
-            [None, "examples/video/vrnerf_workshop.mp4", "vrnerf", "workshop", "32", "Real", "True",],
-            [None, "examples/video/fillerbuster_ramen.mp4", "fillerbuster", "ramen", "32", "Real", "True",],
-            [None, "examples/video/meganerf_rubble.mp4", "meganerf", "rubble", "10", "Real", "True",],
-            [None, "examples/video/llff_horns.mp4", "llff", "horns", "12", "Real", "True",],
-            [None, "examples/video/llff_fortress.mp4", "llff", "fortress", "7", "Real", "True",],
-            [None, "examples/video/dtu_scan_106.mp4", "dtu", "scan_106", "20", "Real", "True",],
-            [None, "examples/video/horizongs_hillside_summer.mp4", "horizongs", "hillside_summer", "55", "Synthetic", "True",],
-            [None, "examples/video/kitti360.mp4", "kitti360", "kitti360", "64", "Real", "True",],
-        ]
-
-        def example_pipeline(
-            input_images,
-            input_video,
-            dataset_name,
-            scene_name,
-            num_images_str,
-            image_type,
-            is_example,
-        ):
-            """
-            1) Copy example images to new target_dir
-            2) Reconstruct
-            3) Return model3D + logs + new_dir + updated dropdown + gallery
-            We do NOT return is_example. It's just an input.
-            """
-            target_dir, image_paths = handle_uploads(input_video, input_images)
-            plyfile, video, depth_colored = gradio_demo(target_dir)
-            return plyfile, video, depth_colored, target_dir, image_paths
-
-        gr.Markdown("Click any row to load an example.", elem_classes=["example-log"])
-
-        gr.Examples(
-            examples=examples,
-            inputs=[
+            # ---------------------- Examples section ----------------------
+    
+            examples = [
+                [None, "examples/video/re10k_1eca36ec55b88fe4.mp4", "re10k", "1eca36ec55b88fe4", "2", "Real", "True",],
+                [None, "examples/video/bungeenerf_colosseum.mp4", "bungeenerf", "colosseum", "8", "Synthetic", "True",],
+                [None, "examples/video/fox.mp4", "InstantNGP", "fox", "14", "Real", "True",],
+                [None, "examples/video/matrixcity_street.mp4", "matrixcity", "street", "32", "Synthetic", "True",],
+                [None, "examples/video/vrnerf_apartment.mp4", "vrnerf", "apartment", "32", "Real", "True",],
+                [None, "examples/video/vrnerf_kitchen.mp4", "vrnerf", "kitchen", "17", "Real", "True",],
+                [None, "examples/video/vrnerf_riverview.mp4", "vrnerf", "riverview", "12", "Real", "True",],
+                [None, "examples/video/vrnerf_workshop.mp4", "vrnerf", "workshop", "32", "Real", "True",],
+                [None, "examples/video/fillerbuster_ramen.mp4", "fillerbuster", "ramen", "32", "Real", "True",],
+                [None, "examples/video/meganerf_rubble.mp4", "meganerf", "rubble", "10", "Real", "True",],
+                [None, "examples/video/llff_horns.mp4", "llff", "horns", "12", "Real", "True",],
+                [None, "examples/video/llff_fortress.mp4", "llff", "fortress", "7", "Real", "True",],
+                [None, "examples/video/dtu_scan_106.mp4", "dtu", "scan_106", "20", "Real", "True",],
+                [None, "examples/video/horizongs_hillside_summer.mp4", "horizongs", "hillside_summer", "55", "Synthetic", "True",],
+                [None, "examples/video/kitti360.mp4", "kitti360", "kitti360", "64", "Real", "True",],
+            ]
+    
+            def example_pipeline(
                 input_images,
                 input_video,
                 dataset_name,
                 scene_name,
-                num_images,
+                num_images_str,
                 image_type,
                 is_example,
-            ],
-            outputs=[
-                reconstruction_output,
-                rgb_video,
-                depth_video,
-                target_dir_output,
-                image_gallery,
-            ],
-            fn=example_pipeline,
-            cache_examples=False,
-            examples_per_page=50,
-        )
-
-        gr.Markdown("<p style='text-align: center; font-style: italic; color: #666;'>We thank VGGT for their excellent gradio implementation!</p>")
+            ):
+                """
+                1) Copy example images to new target_dir
+                2) Reconstruct
+                3) Return model3D + logs + new_dir + updated dropdown + gallery
+                We do NOT return is_example. It's just an input.
+                """
+                target_dir, image_paths = handle_uploads(input_video, input_images)
+                plyfile, video, depth_colored = gradio_demo(target_dir)
+                return plyfile, video, depth_colored, target_dir, image_paths
+    
+            gr.Markdown("Click any row to load an example.", elem_classes=["example-log"])
+    
+            gr.Examples(
+                examples=examples,
+                inputs=[
+                    input_images,
+                    input_video,
+                    dataset_name,
+                    scene_name,
+                    num_images,
+                    image_type,
+                    is_example,
+                ],
+                outputs=[
+                    reconstruction_output,
+                    rgb_video,
+                    depth_video,
+                    target_dir_output,
+                    image_gallery,
+                ],
+                fn=example_pipeline,
+                cache_examples=False,
+                examples_per_page=50,
+            )
+
+            gr.Markdown("<p style='text-align: center; font-style: italic; color: #666;'>We thank VGGT for their excellent gradio implementation!</p>")
 
         submit_btn.click(
             fn=clear_fields,