Update inference.py

inference.py

@@ -60,7 +60,6 @@ def convert_video(model,
     assert output_type in ['video', 'png_sequence'], 'Only support "video" and "png_sequence" output modes.'
     assert seq_chunk >= 1, 'Sequence chunk must be >= 1'
     assert num_workers >= 0, 'Number of workers must be >= 0'
-    assert output_video_mbps == None or output_type == 'video', 'Mbps is not available for png_sequence output.'
     
     # Initialize transform
     if input_resize is not None: