Update data/dot_single_video/precess_dataset_with_dot_single_video_return_position.py (#1)

- Update data/dot_single_video/precess_dataset_with_dot_single_video_return_position.py (17a08051d333c816ce1b52b48dfca0dcbb5b2903)


Co-authored-by: zzw <[email protected]>

data/dot_single_video/precess_dataset_with_dot_single_video_return_position.py

@@ -220,6 +220,7 @@ if __name__ == '__main__':
     tensor_frames = spatial_transform(tensor_frames).unsqueeze(0) / 255.
     
     # 2、get video tensor frames..
+    assert inference_config['return_flow'] == False, "for this mode, you need set this false"         # ! note, important..
     with torch.no_grad():
         pred = dot_model({"video": tensor_frames.to(device)}, **inference_config)    # input [b, f, c, h, w], torch.float32, 0~1 input