fix order bug

app.py

@@ -127,7 +127,7 @@ def main():
             ["image-text", "A dog|A car|A bird", "assets/dog_image.jpg", None, None],
             ["image-text", "A dog|A car|A bird", "assets/car_image.jpg", None, None],
             ["audio-text", "A dog|A car|A bird",  None, "assets/bird_audio.wav", None],
-            ["video-text", "A dog|A car|A bird", None, "assets/dog_video.mp4", None],
+            ["video-text", "A dog|A car|A bird", None, None, "assets/dog_video.mp4"],
         ],
         description="""<p>This is a simple demo of ImageBind for zero-shot cross-modal understanding (now including image classification, audio classification, and video classification). Please refer to the original <a href='https://arxiv.org/abs/2305.05665' target='_blank'>paper</a> and <a href='https://github.com/facebookresearch/ImageBind' target='_blank'>repo</a> for more details.<br>
                     To test your own cases, you can upload an image, an audio or a video, and provide the candidate texts separated by "|".<br>