# Train Short, Inference Long: Training-free Horizon Extension for Autoregressive Video Generation

arXiv: https://arxiv.org/abs/2602.14027 · HTML: https://mlsanity.com/paper/2602.14027/

## Key facts

- Cited by 11 papers in the map (data through 14 Oct 2026).
- Placed in Video generation, one of 14 directions in Diffusion models.
- Builds on 30 of its 34 references that are in the map.
- First posted on arXiv 15 Feb 2026.

## Placed in

- Diffusion models / [Video generation](https://mlsanity.com/diffusion/video-generation/)

## Builds on (most cited first, 30 of 30)

- [LLaMA: Open and Efficient Foundation Language Models](https://mlsanity.com/paper/2302.13971/) (2023)
- [Qwen3 Technical Report](https://mlsanity.com/paper/2505.09388/) (2025)
- [RoFormer: Enhanced Transformer with Rotary Position Embedding](https://mlsanity.com/paper/2104.09864/) (2021)
- [Wan: Open and Advanced Large-Scale Video Generative Models](https://mlsanity.com/paper/2503.20314/) (2025)
- [Neural Tangent Kernel: Convergence and Generalization in Neural Networks](https://mlsanity.com/paper/1806.07572/) (2018)
- [Stable Video Diffusion: Scaling Latent Video Diffusion Models to Large Datasets](https://mlsanity.com/paper/2311.15127/) (2023)
- [Efficient Streaming Language Models with Attention Sinks](https://mlsanity.com/paper/2309.17453/) (2023)
- [VBench: Comprehensive Benchmark Suite for Video Generative Models](https://mlsanity.com/paper/2311.17982/) (2023)
- [AnimateDiff: Animate Your Personalized Text-to-Image Diffusion Models without Specific Tuning](https://mlsanity.com/paper/2307.04725/) (2023)
- [Train Short, Test Long: Attention with Linear Biases Enables Input Length Extrapolation](https://mlsanity.com/paper/2108.12409/) (2021)
- [One-step Diffusion with Distribution Matching Distillation](https://mlsanity.com/paper/2311.18828/) (2023)
- [Extending Context Window of Large Language Models via Positional Interpolation](https://mlsanity.com/paper/2306.15595/) (2023)
- [Movie Gen: A Cast of Media Foundation Models](https://mlsanity.com/paper/2410.13720/) (2024)
- [Self Forcing: Bridging the Train-Test Gap in Autoregressive Video Diffusion](https://mlsanity.com/paper/2506.08009/) (2025)
- [YaRN: Efficient Context Window Extension of Large Language Models](https://mlsanity.com/paper/2309.00071/) (2023)
- [Phenaki: Variable Length Video Generation From Open Domain Textual Description](https://mlsanity.com/paper/2210.02399/) (2022)
- [From Slow Bidirectional to Fast Autoregressive Video Diffusion Models](https://mlsanity.com/paper/2412.07772/) (2024)
- [LongRoPE: Extending LLM Context Window Beyond 2 Million Tokens](https://mlsanity.com/paper/2402.13753/) (2024)
- [Gated Attention for Large Language Models: Non-linearity, Sparsity, and Attention-Sink-Free](https://mlsanity.com/paper/2505.06708/) (2025)
- [LongLive: Real-time Interactive Long Video Generation](https://mlsanity.com/paper/2509.22622/) (2025)
- [Rolling Forcing: Autoregressive Long Video Diffusion in Real Time](https://mlsanity.com/paper/2509.25161/) (2025)
- [FreeNoise: Tuning-Free Longer Video Diffusion via Noise Rescheduling](https://mlsanity.com/paper/2310.15169/) (2023)
- [History-Guided Video Diffusion](https://mlsanity.com/paper/2502.06764/) (2025)
- [SANA-Video: Efficient Video Generation with Block Linear Diffusion Transformer](https://mlsanity.com/paper/2509.24695/) (2025)
- [Infinity-RoPE: Action-Controllable Infinite Video Generation Emerges From Autoregressive Self-Rollout](https://mlsanity.com/paper/2511.20649/) (2025)
- [AR-Diffusion: Asynchronous Video Generation with Auto-Regressive Diffusion](https://mlsanity.com/paper/2503.07418/) (2025)
- [LongVie: Multimodal-Guided Controllable Ultra-Long Video Generation](https://mlsanity.com/paper/2508.03694/) (2025)
- [FreeLong++: Training-Free Long Video Generation via Multi-band SpectralFusion](https://mlsanity.com/paper/2507.00162/) (2025)
- [Understanding the RoPE Extensions of Long-Context LLMs: An Attention Perspective](https://mlsanity.com/paper/2406.13282/) (2024)
- [StableWorld: Towards Stable and Consistent Long Interactive Video Generation](https://mlsanity.com/paper/2601.15281/) (2026)

## Built on by (most cited first, 11 of 11)

- [LongLive-2.0: An NVFP4 Parallel Infrastructure for Long Video Generation](https://mlsanity.com/paper/2605.18739/) (2026)
- [ABot-World-0: Infinite Interactive World Rollout on a Single Desktop GPU](https://mlsanity.com/paper/2607.19191/) (2026)
- [Addressable Memory for Video World Models](https://mlsanity.com/paper/2608.07408/) (2026)
- [Attention Sink in Transformers: A Survey on Utilization, Interpretation, and Mitigation](https://mlsanity.com/paper/2604.10098/) (2026)
- [Future Forcing: Future-aware Training-free KV Cache Policy for Autoregressive Video Generation](https://mlsanity.com/paper/2605.30083/) (2026)
- [TetherCache: Stabilizing Autoregressive Long-Form Video Generation with Gated Recall and Trusted Alignment](https://mlsanity.com/paper/2606.13035/) (2026)
- [RAVEN: Real-time Autoregressive Video Extrapolation with Consistency-model GRPO](https://mlsanity.com/paper/2605.15190/) (2026)
- [FreqForcing: Autoregressive Long Video Generation via Spectral Self-Anchoring](https://mlsanity.com/paper/2607.27110/) (2026)
- [Towards Memory-Efficient Autoregressive Video Generation via Instance-Specific Parametric Absorption](https://mlsanity.com/paper/2607.00712/) (2026)
- [SAGA: Stable Acceleration Guidance for Autoregressive Video Generation](https://mlsanity.com/paper/2607.08020/) (2026)
- [Recency Forcing: Bridging the Long-Horizon Gap in Autoregressive Video Generation](https://mlsanity.com/paper/2609.19729/) (2026)

## Abstract (arXiv)

Autoregressive video diffusion models have emerged as a scalable paradigm for long video generation. However, they often suffer from severe extrapolation failure, where rapid error accumulation leads to significant temporal degradation when extending beyond training horizons. We identify that this failure primarily stems from the spectral bias of 3D positional embeddings and the lack of dynamic priors in noise sampling. To address these issues, we propose FLEX (Frequency-aware Length EXtension), a training-free inference-time framework that bridges the gap between short-term training and long-term inference. FLEX introduces Frequency-aware RoPE Modulation to adaptively interpolate under-trained low-frequency components while extrapolating high-frequency ones to preserve multi-scale temporal discriminability. This is integrated with Antiphase Noise Sampling (ANS) to inject high-frequency dynamic priors and Inference-only Attention Sink to anchor global structure. Extensive evaluations on VBench demonstrate that FLEX significantly outperforms state-of-the-art models at 6x extrapolation (30s duration) and matches the performance of long-video fine-tuned baselines at 12x scale (60s duration). As a plug-and-play augmentation, FLEX seamlessly integrates into existing inference pipelines for horizon extension. It effectively pushes the generation limits of models such as LongLive, supporting consistent and dynamic video synthesis at a 4-minute scale. Project page is available at https://ga-lee.github.io/FLEX_demo.