Open abhishri-medewar opened 2 years ago
Our model can deal with 4K videos, although it may contain larger motion. I think there are two ways to solve this problem: (1) You can use a GPU with large memory, however, this is usually at a large cost. (2) You can divide your 4K frames into multiple 2K, 1080p or 720p frames and then run IFRNet several times on these divisions. Finally, you put the output frames together into original 4K resolution output frame.
If yes, how ? I am getting cuda out of memory error.