Enable long video generation from short-video diffusion models without any additional training or fine-tuning.
March 27, 2026
Original Paper
Free-Lunch Long Video Generation via Layer-Adaptive O.O.D Correction
arXiv · 2603.25209
The Takeaway
Practitioners can now generate temporally consistent long videos by applying layer-adaptive out-of-distribution corrections to existing models. This bypasses the massive compute and data requirements typically needed to extend video context lengths.
From the abstract
Generating long videos using pre-trained video diffusion models, which are typically trained on short clips, presents a significant challenge. Directly applying these models for long-video inference often leads to a notable degradation in visual quality. This paper identifies that this issue primarily stems from two out-of-distribution (O.O.D) problems: frame-level relative position O.O.D and context-length O.O.D. To address these challenges, we propose FreeLOC, a novel training-free, layer-adap