1 paper · 1 filter
Tianyu Zhang, Shuai Huo, Dengchao Jin +7
Video diffusion models can reconstruct visually compelling content from severely compressed representations at ultra-low bitrates, but they impose a rigid deployment trade-off. Bid…