Draft-and-Target Sampling for Video Generation Policy
This addresses efficiency issues in video generation for robotics, offering a practical improvement for real-time applications, though it is incremental as it builds on existing diffusion methods.
The paper tackles the high computational cost and long inference time of video generation models used as robot policies by proposing Draft-and-Target Sampling, a training-free diffusion inference paradigm that achieves up to 2.1x speedup with minimal compromise to success rate.
Video generation models have been used as a robot policy to predict the future states of executing a task conditioned on task description and observation. Previous works ignore their high computational cost and long inference time. To address this challenge, we propose Draft-and-Target Sampling, a novel diffusion inference paradigm for video generation policy that is training-free and can improve inference efficiency. We introduce a self-play denoising approach by utilizing two complementary denoising trajectories in a single model, draft sampling takes large steps to generate a global trajectory in a fast manner and target sampling takes small steps to verify it. To further speedup generation, we introduce token chunking and progressive acceptance strategy to reduce redundant computation. Experiments on three benchmarks show that our method can achieve up to 2.1x speedup and improve the efficiency of current state-of-the-art methods with minimal compromise to the success rate. Our code is available.