Hi Abedi, I would like to clarify the some implementation in transforms.py. In the paper, it was said that there were 50 timesteps extracted from the video, so I was wondering, are these time steps consecutive, (i.e. 1.jpg, 2.jpg, etc) or are they extracted throughout the video (i.e. 1.jpg, 3.jpg, ..., 100.jpg ).
Hi Abedi, I would like to clarify the some implementation in transforms.py. In the paper, it was said that there were 50 timesteps extracted from the video, so I was wondering, are these time steps consecutive, (i.e. 1.jpg, 2.jpg, etc) or are they extracted throughout the video (i.e. 1.jpg, 3.jpg, ..., 100.jpg ).