Mask Forcing: Improving Autoregressive Video Diffusion Distillation via Dual-Noise Masking Rollout
Open paper
Benchmark
Model
Rank
Results
text-to-video-generation-on-vbench
Causal Forcing + Mask Forcing (chunk-wise)
#30
VBench Total Score: 82.76
text-to-video-generation-on-vbench
LongLive + Mask Forcing (chunk-wise)
#39
VBench Quality Score: 83.71
Rank counts only results with a code link.