Training-free wrapper accelerates causal video diffusion to 50 FPS without quality loss
The authors present UnStep, a training-free wrapper that accelerates already-distilled causal video diffusion models by running them with fewer denoising steps and a limited temporal KV cache. Two inference-only mechanisms—clean-cache refinement (renoise and refine) and truncated SVD on attention projections—recover the quality lost by these reductions, while an optimized runtime stack further boosts throughput. UnStep achieves 50 frames per second on a single H100 GPU and 77 FPS on a GB200 GPU without quality degradation, all without retraining.
