Preserve Your Own Correlation: A Noise Prior for Video Diffusion Models

Benchmark Model Rank Results
text-to-video-generation-on-ucf-101PYoCo (Zero-shot, 64x64)–FVD16: 355.19
video-generation-on-ucf-101PYoCo (Zero-shot, 64x64, text-conditional)–FVD16: 355.19Inception Score: 47.76
video-generation-on-ucf-101PYoCo (Zero-shot, 64x64, unconditional)–FVD16: 310Inception Score: 60.01