A block-pruning plus distillation method for video diffusion models, VDMini, achieves 1.25x to 2.5x speedups with modest quality loss across three video generation models.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Individual Content and Motion Dynamics Preserved Pruning for Video Diffusion Models
A block-pruning plus distillation method for video diffusion models, VDMini, achieves 1.25x to 2.5x speedups with modest quality loss across three video generation models.