- [MINOR] [SOURCE] [UPDATE] 1. update app.py

app.py

@@ -63,10 +63,8 @@ def inference_one_video(input_video):
     """
     TESTING
     """
-    # Convert input video to tensor and adjust dimensions
-    input_video_tensor = torch.from_numpy(input_video).permute(0, 3, 1, 2)
     views, _ = spatial_temporal_view_decomposition(
-        input_video_tensor, dopt["sample_types"], temporal_samplers
+        input_video, dopt["sample_types"], temporal_samplers
     )
 
     for k, v in views.items():