Elamur Research⌕
☰
home / papers / 2025 / Self Forcing: Bridging the Train-Test Gap in Autoregressive Video Diffusion

Self Forcing: Bridging the Train-Test Gap in Autoregressive Video Diffusion

Paper · 2025 · Xun Huang, Zhengqi Li, Guande He et al.

Summary

Autoregressive video diffusion models build a video frame by frame. They are usually trained with Teacher Forcing or Diffusion Forcing, where the context frames are ground-truth or noised ground-truth. At inference, however, the model must condition on its own imperfect frames. This exposure bias causes errors to compound over time. Self Forcing fixes the mismatch by running the real autoregressive rollout during training, with KV caching, so each frame is conditioned on frames the model produced itself. The whole generated video is then scored with a video-level distribution-matching loss (DMD, SiD or GAN), not a per-frame denoising loss.

Two choices keep this sequential training affordable:

  • a few-step diffusion backbone;
  • stochastic truncation of gradients through the rollout.

A rolling KV cache supports extending videos beyond the training length. The resulting model streams video in real time on one GPU with sub-second latency. Its quality matches or beats much slower bidirectional models and CausVid-style autoregressive models. This makes interactive uses such as games, live streams and world simulation more realistic.

Key points

  • Speed: a single H100 GPU is enough to generate roughly 17 FPS, with latency under one second.
  • Loss choice: the method works with DMD, SiD and GAN objectives and beats Teacher Forcing and Diffusion Forcing baselines under each of them.
    • The baselines lose quality when moving from chunk-wise to frame-wise generation, typically drifting into oversaturation or over-sharpening.
    • Self Forcing stays stable in both setups.
  • CausVid diagnosis: CausVid (Diffusion Forcing + DMD) is identified as matching the wrong distribution. Its training samples are not what the model actually generates at test time.
  • Long videos (10 seconds):
    • Recomputing the KV cache at each window shift drops throughput to 4.6 FPS.
    • A naive rolling cache keeps speed but produces visual artifacts.
    • Training the model without access to the initial image latent removes the artifacts at 16.1 FPS.
  • Training cost:
    • Per-iteration time is similar to Teacher Forcing and Diffusion Forcing.
    • Quality on VBench is higher for the same wall-clock budget.
    • A DMD run converges in about 1.5 hours on 64 H100 GPUs.

Interesting ideas

  • I1. Parallel pre-training, sequential post-training. Pre-train cheaply with parallel objectives, then add a short post-training stage where the model consumes its own outputs. This could transfer to any autoregressive generator with continuous outputs, such as speech, motion or time series.
  • I2. Score the whole sequence, not each step. Judging the full generated sequence against real data, GAN-style, directly penalizes accumulated drift. Per-step losses cannot see that drift.
  • I3. Truncate gradients to make rollout training fit in memory. Unrolling during training is feasible if you use few inference steps and backpropagate through only a randomly chosen part of the rollout. The general lesson: costly simulation-style training becomes affordable when gradients are restricted.
  • I4. Mimic the inference-time context during training. If inference will discard the first frame from the cache, train the model without that frame as well. Matching training context to deployment conditions is a cheap way to avoid artifacts in sliding-window generation.
  • I5. Sequential rollout may not be slower in practice. Plain full attention can use fast kernels such as FlashAttention-3. Parallel training with custom causal masks pays extra overhead. When comparing training schemes, measure wall-clock quality curves rather than assuming parallel training wins.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

 Institutions

🏢MicrosoftRedmond · United States378 works · 161,630 citations🎓The University of Texas at AustinAustin · United States119 works · 57,950 citations🏢Adobe SystemsSan Jose · United States40 works · 4,251 citations

 Fields, topics and concepts

📚 Shelves 2The field of science this work belongs to (OpenAlex fields).Computer Science · 18,614 worksComputer Vision and Pattern Recognition · 3,486 works

Cite

Cite as Xun Huang, Zhengqi Li, Guande He, Mingyuan Zhou and Eli Shechtman. 2025. Self Forcing: Bridging the Train-Test Gap in Autoregressive Video Diffusion. arXiv:2506.08009.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.48550/arxiv.2506.08009 · arXiv 2506.08009

@misc{huang2025self,
  title = {{Self Forcing: Bridging the Train-Test Gap in Autoregressive Video Diffusion}},
  author = {Xun Huang, Zhengqi Li, Guande He, Mingyuan Zhou and Eli Shechtman},
  year = {2025},
  doi = {10.48550/arxiv.2506.08009},
  eprint = {2506.08009},
  archivePrefix = {arXiv},
  url = {https://doi.org/10.48550/arxiv.2506.08009}
}

References

113 references, 15 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Marianne Arriola, Aaron Gokaslan, Justin T Chiu, Zhihan Yang, Zhixuan Qi, Jiaqi Han, Subham Sekhar Sahoo, and Volodymyr Kuleshov. Block diffusion: Interpolating between autoregressive and diffusion language models. In ICLR, 2025. find
  2. Abdelhak Bentaleb, May Lim, Mehmet N Akcay, Ali C Begen, Sarra Hammoudi, and Roger Zimmermann. Toward one-second latency: Evolution of live media streaming. IEEE Communications Surveys & Tutorials, 2025. Quality vs. Training Time 0.840 0.835 VBench Total Score 0.830 0.825 0.820 Diffusion Forcing Tea… find
  3. Andreas Blattmann, Tim Dockhorn, Sumith Kulal, Daniel Mendelevitch, Maciej Kilian, Dominik Lorenz, Yam Levi, Zion English, Vikram Voleti, Adam Letts, et al. Stable video diffusion: Scaling latent video diffusion models to large datasets. arXiv preprint arXiv:2311.15127, 2023. — [arXiv:2311.15127](https://arxiv.org/abs/2311.15127) find
  4. Andreas Blattmann, Robin Rombach, Huan Ling, Tim Dockhorn, Seung Wook Kim, Sanja Fidler, and Karsten Kreis. Align your latents: High-resolution video synthesis with latent diffusion models. In CVPR, 2023. find
  5. Tim Brooks, Janne Hellsten, Miika Aittala, Ting-Chun Wang, Timo Aila, Jaakko Lehtinen, Ming-Yu Liu, Alexei Efros, and Tero Karras. Generating long videos of dynamic scenes. NeurIPS, 2022. find
  6. Tim Brooks, Bill Peebles, Connor Holmes, Will DePue, Yufei Guo, Li Jing, David Schnurr, Joe Taylor, Troy Luhman, Eric Luhman, Clarence Ng, Ricky Wang, and Aditya Ramesh. Video generation models as world simulators, 2024. find
  7. Jake Bruce, Michael D Dennis, Ashley Edwards, Jack Parker-Holder, Yuge Shi, Edward Hughes, Matthew Lai, Aditi Mavalankar, Richie Steigerwald, Chris Apps, et al. Genie: Generative interactive environments. In ICML, 2024. find
  8. Boyuan Chen, Diego Martí Monsó, Yilun Du, Max Simchowitz, Russ Tedrake, and Vincent Sitzmann. Diffusion forcing: Next-token prediction meets full-sequence diffusion. In NeurIPS, 2024. find
  9. Feng Chen, Zhen Yang, Bohan Zhuang, and Qi Wu. Streaming video diffusion: Online video editing with diffusion models. arXiv preprint arXiv:2405.19726, 2024. — [arXiv:2405.19726](https://arxiv.org/abs/2405.19726) find
  10. Guibin Chen, Dixuan Lin, Jiangping Yang, Chunze Lin, Juncheng Zhu, Mingyuan Fan, Hao Zhang, Sheng Chen, Zheng Chen, Chengchen Ma, et al. Skyreels-v2: Infinite-length film generative model. arXiv preprint arXiv:2504.13074, 2025. — [arXiv:2504.13074](https://arxiv.org/abs/2504.13074) find
  11. Julian Decart, Quinn Quevedo, Spruce McIntyre, Xinlei Campbell, Robert Chen, and Wachen. Oasis: A universe in a transformer, 2024. find
  12. Chaorui Deng, Deyao Zhu, Kunchang Li, Shi Guang, and Haoqi Fan. Causal diffusion transformers for generative modeling. arXiv preprint arXiv:2412.12095, 2024. — [arXiv:2412.12095](https://arxiv.org/abs/2412.12095) find
  13. Haoge Deng, Ting Pan, Haiwen Diao, Zhengxiong Luo, Yufeng Cui, Huchuan Lu, Shiguang Shan, Yonggang Qi, and Xinlong Wang. Autoregressive video generation without vector quantization. In ICLR, 2025. find
  14. Emily L Denton et al. Unsupervised learning of disentangled representations from video. In NeurIPS, 2017. find
  15. Juechu Dong, Boyuan Feng, Driss Guessous, Yanbo Liang, and Horace He. Flex attention: A programming model for generating optimized attention kernels. ArXiv, abs/2412.05496, 2024.
  16. Kaifeng Gao, Jiaxin Shi, Hanwang Zhang, Chunping Wang, Jun Xiao, and Long Chen. Ca2-vdm: Efficient autoregressive video diffusion model with causal generation and cache sharing. arXiv preprint arXiv:2411.16375, 2024. — [arXiv:2411.16375](https://arxiv.org/abs/2411.16375) find
  17. Songwei Ge, Thomas Hayes, Harry Yang, Xi Yin, Guan Pang, David Jacobs, Jia-Bin Huang, and Devi Parikh. Long video generation with time-agnostic vqgan and time-sensitive transformer. In ECCV, 2022. find
  18. Ian J Goodfellow, Jean Pouget-Abadie, Mehdi Mirza, Bing Xu, David Warde-Farley, Sherjil Ozair, Aaron Courville, and Yoshua Bengio. Generative adversarial nets. In NeurIPS, 2014.
  19. Albert Gu and Tri Dao. Mamba: Linear-time sequence modeling with selective state spaces. In COLM, 2024. find
  20. Yuchao Gu, Weijia Mao, and Mike Zheng Shou. Long-context autoregressive video modeling with next-frame prediction. arXiv preprint arXiv:2503.19325, 2025. — [arXiv:2503.19325](https://arxiv.org/abs/2503.19325) find
  21. Daya Guo, Dejian Yang, Haowei Zhang, Junxiao Song, Ruoyu Zhang, Runxin Xu, Qihao Zhu, Shirong Ma, Peiyi Wang, Xiao Bi, et al. Deepseek-r1: Incentivizing reasoning capability in llms via reinforcement learning. arXiv preprint arXiv:2501.12948, 2025.
  22. Yuwei Guo, Ceyuan Yang, Ziyan Yang, Zhibei Ma, Zhijie Lin, Zhenheng Yang, Dahua Lin, and Lu Jiang. Long context tuning for video generation. arXiv preprint arXiv:2503.10589, 2025. — [arXiv:2503.10589](https://arxiv.org/abs/2503.10589) find
  23. Agrim Gupta, Lijun Yu, Kihyuk Sohn, Xiuye Gu, Meera Hahn, Fei-Fei Li, Irfan Essa, Lu Jiang, and José Lezama. Photorealistic video generation with diffusion models. In ECCV, 2024. find
  24. Yoav HaCohen, Nisan Chiprut, Benny Brazowski, Daniel Shalem, Dudu Moshe, Eitan Richardson, Eran Levin, Guy Shiran, Nir Zabari, Ori Gordon, et al. Ltx-video: Realtime video latent diffusion. arXiv preprint arXiv:2501.00103, 2024. find
  25. Jonathan Ho, William Chan, Chitwan Saharia, Jay Whang, Ruiqi Gao, Alexey A. Gritsenko, Diederik P. Kingma, Ben Poole, Mohammad Norouzi, David J. Fleet, and Tim Salimans. Imagen video: High definition video generation with diffusion models. ArXiv, abs/2210.02303, 2022. — [arXiv:2210.02303](https://arxiv.org/abs/2210.02303) find
  26. Jonathan Ho, Tim Salimans, Alexey Gritsenko, William Chan, Mohammad Norouzi, and David J Fleet. Video diffusion models. In NeurIPS, 2022. find
  27. Wenyi Hong, Ming Ding, Wendi Zheng, Xinghan Liu, and Jie Tang. Cogvideo: Large-scale pretraining for text-to-video generation via transformers. In ICLR, 2023.
  28. Jinyi Hu, Shengding Hu, Yuxuan Song, Yufei Huang, Mingxuan Wang, Hao Zhou, Zhiyuan Liu, Wei-Ying Ma, and Maosong Sun. Acdit: Interpolating autoregressive conditional modeling and diffusion transformer. arXiv preprint arXiv:2412.07720, 2024. — [arXiv:2412.07720](https://arxiv.org/abs/2412.07720) find
  29. Nick Huang, Aaron Gokaslan, Volodymyr Kuleshov, and James Tompkin. The gan is dead; long live the gan! a modern gan baseline. In NeurIPS, 2024. find
  30. Zemin Huang, Zhengyang Geng, Weijian Luo, and Guo-jun Qi. Flow generator matching. arXiv preprint arXiv:2410.19310, 2024. — [arXiv:2410.19310](https://arxiv.org/abs/2410.19310) find
  31. Ziqi Huang, Yinan He, Jiashuo Yu, Fan Zhang, Chenyang Si, Yuming Jiang, Yuanhan Zhang, Tianxing Wu, Qingyang Jin, Nattapol Chanpaisit, Yaohui Wang, Xinyuan Chen, Limin Wang, Dahua Lin, Yu Qiao, and Ziwei Liu. VBench: Comprehensive benchmark suite for video generative models. In CVPR, 2024. find
  32. Simon Jenni and Paolo Favaro. On stabilizing generative adversarial training with noise. In CVPR, 2019. find
  33. Yang Jin, Zhicheng Sun, Ningyuan Li, Kun Xu, Hao Jiang, Nan Zhuang, Quzhe Huang, Yang Song, Yadong Mu, and Zhouchen Lin. Pyramidal flow matching for efficient video generative modeling. In ICLR, 2025. find
  34. Alexia Jolicoeur-Martineau. The relativistic discriminator: a key element missing from standard gan. In ICLR, 2019. find
  35. Jihwan Kim, Junoh Kang, Jinyoung Choi, and Bohyung Han. Fifo-diffusion: Generating infinite videos from text without training. In NeurIPS, 2024. find
  36. Diederik Kingma, Tim Salimans, Ben Poole, and Jonathan Ho. Variational diffusion models. In NeurIPS, 2021. find
  37. Diederik P Kingma and Max Welling. Auto-encoding variational bayes. In ICLR, 2014.
  38. Dan Kondratyuk, Lijun Yu, Xiuye Gu, Jose Lezama, Jonathan Huang, Grant Schindler, Rachel Hornung, Vighnesh Birodkar, Jimmy Yan, Ming-Chang Chiu, et al. Videopoet: A large language model for zero-shot video generation. In ICML, 2024. find
  39. Weijie Kong, Qi Tian, Zijian Zhang, Rox Min, Zuozhuo Dai, Jin Zhou, Jiangfeng Xiong, Xin Li, Bo Wu, Jianwei Zhang, et al. Hunyuanvideo: A systematic framework for large video generative models. arXiv preprint arXiv:2412.03603, 2024.
  40. Alex M Lamb, Anirudh Goyal ALIAS PARTH GOYAL, Ying Zhang, Saizheng Zhang, Aaron C Courville, and Yoshua Bengio. Professor forcing: A new algorithm for training recurrent networks. In NeurIPS, 2016. find
  41. Qing Li, Xun Tang, Junkun Peng, Yuanzheng Tan, and Yong Jiang. Latency reducing in real-time internet video transport: A survey. SSRN 4654242, 2023. find
  42. Shuang Li, Yihuai Gao, Dorsa Sadigh, and Shuran Song. Unified video action model. arXiv preprint arXiv:2503.00200, 2025. — [arXiv:2503.00200](https://arxiv.org/abs/2503.00200) find
  43. Tianhong Li, Yonglong Tian, He Li, Mingyang Deng, and Kaiming He. Autoregressive image generation without vector quantization. In NeurIPS, 2024. find
  44. Zhengqi Li, Qianqian Wang, Noah Snavely, and Angjoo Kanazawa. Infinitenature-zero: Learning perpetual view generation of natural scenes from single images. In ECCV, 2022. find
  45. Zongyi Li, Shujie Hu, Shujie Liu, Long Zhou, Jeongsoo Choi, Lingwei Meng, Xun Guo, Jinyu Li, Hefei Ling, and Furu Wei. Arlon: Boosting diffusion transformers with autoregressive models for long video generation. In ICLR, 2025. find
  46. Feng Liang, Akio Kodaira, Chenfeng Xu, Masayoshi Tomizuka, Kurt Keutzer, and Diana Marculescu. Looking backward: Streaming video-to-video translation with feature banks. In ICLR, 2025. find
  47. Shanchuan Lin, Xin Xia, Yuxi Ren, Ceyuan Yang, Xuefeng Xiao, and Lu Jiang. Diffusion adversarial post-training for one-step video generation. arXiv preprint arXiv:2501.08316, 2025. — [arXiv:2501.08316](https://arxiv.org/abs/2501.08316) find
  48. Yaron Lipman, Ricky TQ Chen, Heli Ben-Hamu, Maximilian Nickel, and Matthew Le. Flow matching for generative modeling. In ICLR, 2023.
  49. Andrew Liu, Richard Tucker, Varun Jampani, Ameesh Makadia, Noah Snavely, and Angjoo Kanazawa. Infinite nature: Perpetual view generation of natural scenes from a single image. In ICCV, 2021. find
  50. Haozhe Liu, Shikun Liu, Zijian Zhou, Mengmeng Xu, Yanping Xie, Xiao Han, Juan C Pérez, Ding Liu, Kumara Kahatapitiya, Menglin Jia, et al. Mardini: Masked autoregressive diffusion for video generation at scale. arXiv preprint arXiv:2410.20280, 2024. — [arXiv:2410.20280](https://arxiv.org/abs/2410.20280) find
  51. Xingchao Liu, Chengyue Gong, and Qiang Liu. Flow straight and fast: Learning to generate and transfer data with rectified flow. In ICLR, 2023.
  52. Yaofang Liu, Yumeng Ren, Xiaodong Cun, Aitor Artola, Yang Liu, Tieyong Zeng, Raymond H Chan, and Jean-michel Morel. Redefining temporal modeling in video diffusion: The vectorized timestep approach. arXiv preprint arXiv:2410.03160, 2024. — [arXiv:2410.03160](https://arxiv.org/abs/2410.03160) find
  53. Zhijun Liu, Shuai Wang, Sho Inoue, Qibing Bai, and Haizhou Li. Autoregressive diffusion transformer for text-to-speech synthesis. arXiv preprint arXiv:2406.05551, 2024. — [arXiv:2406.05551](https://arxiv.org/abs/2406.05551) find
  54. Weijian Luo, Tianyang Hu, Shifeng Zhang, Jiacheng Sun, Zhenguo Li, and Zhihua Zhang. Diff-instruct: A universal approach for transferring knowledge from pre-trained diffusion models. In NeurIPS, 2023. find
  55. Weijian Luo, Zemin Huang, Zhengyang Geng, J Zico Kolter, and Guo-jun Qi. One-step diffusion distillation through score implicit matching. NeurIPS, 2024. find
  56. Xiaofeng Mao, Zhengkai Jiang, Fu-Yun Wang, Jiangning Zhang, Hao Chen, Mingmin Chi, Yabiao Wang, and Wenhan Luo. Osv: One step is enough for high-quality image to video generation. In CVPR, 2025. find
  57. William Merrill and Ashish Sabharwal. The parallelism tradeoff: Limitations of log-precision transformers. TACL, 2023. find
  58. Lars Mescheder, Andreas Geiger, and Sebastian Nowozin. Which training methods for gans do actually converge? In ICML, 2018. find
  59. Sicheng Mo, Thao Nguyen, Xun Huang, Siddharth Srinivasan Iyer, Yijun Li, Yuchen Liu, Abhishek Tandon, Eli Shechtman, Krishna Kumar Singh, Yong Jae Lee, et al. X-fusion: Introducing new modality to frozen large language models. arXiv preprint arXiv:2504.20996, 2025. — [arXiv:2504.20996](https://arxiv.org/abs/2504.20996) find
  60. Mang Ning, Mingxiao Li, Jianlin Su, Albert Ali Salah, and Itir Onal Ertugrul. Elucidating the exposure bias in diffusion models. In ICLR, 2024. find
  61. Jack Parker-Holder, Philip Ball, Jake Bruce, Vibhavari Dasagi, Kristian Holsheimer, Christos Kaplanis, Alexandre Moufarek, Guy Scully, Jeremy Shar, Jimmy Shi, Stephen Spencer, Jessica Yung, Michael Dennis, Sultan Kenjeyev, Shangbang Long, Vlad Mnih, Harris Chan, Maxime Gazeau, Bonnie Li, Fabio Pardo… find
  62. William S Peebles and Saining Xie. Scalable diffusion models with transformers. In ICCV, 2023.
  63. Ryan Po, Yotam Nitzan, Richard Zhang, Berlin Chen, Tri Dao, Eli Shechtman, Gordon Wetzstein, and Xun Huang. Long-context state-space video world models. arXiv preprint arXiv:2505.20171, 2025. — [arXiv:2505.20171](https://arxiv.org/abs/2505.20171) find
  64. Adam Polyak, Amit Zohar, Andrew Brown, Andros Tjandra, Animesh Sinha, Ann Lee, Apoorv Vyas, Bowen Shi, Chih-Yao Ma, Ching-Yao Chuang, et al. Movie gen: A cast of media foundation models. arXiv preprint arXiv:2410.13720, 2024. — [arXiv:2410.13720](https://arxiv.org/abs/2410.13720) find
  65. Marc’Aurelio Ranzato, Sumit Chopra, Michael Auli, and Wojciech Zaremba. Sequence level training with recurrent neural networks. In ICLR, 2016.
  66. Shuhuai Ren, Shuming Ma, Xu Sun, and Furu Wei. Next block prediction: Video generation via semi-auto-regressive modeling. arXiv preprint arXiv:2502.07737, 2025. — [arXiv:2502.07737](https://arxiv.org/abs/2502.07737) find
  67. David Ruhe, Jonathan Heek, Tim Salimans, and Emiel Hoogeboom. Rolling diffusion models. In ICML, 2024. find
  68. Masaki Saito, Eiichi Matsumoto, and Shunta Saito. Temporal generative adversarial nets with singular value clipping. In ICCV, 2017. find
  69. Sand-AI. Magi-1: Autoregressive video generation at scale, 2025. find
  70. Axel Sauer, Frederic Boesel, Tim Dockhorn, Andreas Blattmann, Patrick Esser, and Robin Rombach. Fast high-resolution image synthesis with latent adversarial diffusion distillation. In SIGGRAPH Asia 2024 Conference Papers, pages 1–11, 2024. find
  71. Florian Schmidt. Generalization in generation: A closer look at exposure bias. EMNLP-IJCNLP 2019, page 157, 2019. find
  72. Jay Shah, Ganesh Bikshandi, Ying Zhang, Vijay Thakkar, Pradeep Ramani, and Tri Dao. Flashattention-3: Fast and accurate attention with asynchrony and low-precision. In NeurIPS, 2024.
  73. Kiwhan Song, Boyuan Chen, Max Simchowitz, Yilun Du, Russ Tedrake, and Vincent Sitzmann. History-guided video diffusion. arXiv preprint arXiv:2502.06764, 2025. — [arXiv:2502.06764](https://arxiv.org/abs/2502.06764) find
  74. Yang Song, Prafulla Dhariwal, Mark Chen, and Ilya Sutskever. Consistency models. In ICML, 2023. find
  75. Yang Song, Conor Durkan, Iain Murray, and Stefano Ermon. Maximum likelihood training of score-based diffusion models. In NeurIPS, 2021. find
  76. Mingzhen Sun, Weining Wang, Gen Li, Jiawei Liu, Jiahui Sun, Wanquan Feng, Shanshan Lao, SiYu Zhou, Qian He, and Jing Liu. Ar-diffusion: Asynchronous video generation with auto-regressive diffusion. In CVPR, 2025. find
  77. Sergey Tulyakov, Ming-Yu Liu, Xiaodong Yang, and Jan Kautz. Mocogan: Decomposing motion and content for video generation. In CVPR, 2018. find
  78. Dani Valevski, Yaniv Leviathan, Moab Arar, and Shlomi Fruchter. Diffusion models are real-time game engines. In ICLR, 2025. find
  79. Aaron Van Den Oord, Oriol Vinyals, et al. Neural discrete representation learning. In NeurIPS, 2017. find
  80. R Villegas, H Moraldo, S Castro, M Babaeizadeh, H Zhang, J Kunze, PJ Kindermans, MT Saffar, and D Erhan. Phenaki: Variable length video generation from open domain textual descriptions. In ICLR, 2023. find
  81. Ruben Villegas, Jimei Yang, Seunghoon Hong, Xunyu Lin, and Honglak Lee. Decomposing motion and content for natural video sequence prediction. In ICLR, 2017. find
  82. Carl Vondrick, Hamed Pirsiavash, and Antonio Torralba. Generating videos with scene dynamics. NeurIPS, 2016. find
  83. Ang Wang, Baole Ai, Bin Wen, Chaojie Mao, Chen-Wei Xie, Di Chen, Feiwu Yu, Haiming Zhao, Jianxiao Yang, Jianyuan Zeng, et al. Wan: Open and advanced large-scale video generative models. arXiv preprint arXiv:2503.20314, 2025.
  84. Jing Wang, Fengzhuo Zhang, Xiaoli Li, Vincent YF Tan, Tianyu Pang, Chao Du, Aixin Sun, and Zhuoran Yang. Error analyses of auto-regressive video diffusion models: A unified framework. arXiv preprint arXiv:2503.10704, 2025. — [arXiv:2503.10704](https://arxiv.org/abs/2503.10704) find
  85. Wenhao Wang and Yi Yang. Vidprom: A million-scale real prompt-gallery dataset for text-to-video diffusion models. In NeurIPS, 2024. find
  86. Yuqing Wang, Tianwei Xiong, Daquan Zhou, Zhijie Lin, Yang Zhao, Bingyi Kang, Jiashi Feng, and Xihui Liu. Loong: Generating minute-level long videos with autoregressive language models. arXiv preprint arXiv:2410.02757, 2024. — [arXiv:2410.02757](https://arxiv.org/abs/2410.02757) find
  87. Zhengyi Wang, Cheng Lu, Yikai Wang, Fan Bao, Chongxuan Li, Hang Su, and Jun Zhu. Prolificdreamer: High-fidelity and diverse text-to-3d generation with variational score distillation. NeurIPS, 2023. find
  88. Dirk Weissenborn, Oscar Täckström, and Jakob Uszkoreit. Scaling autoregressive video models. In ICLR, 2020. find
  89. Wenming Weng, Ruoyu Feng, Yanhui Wang, Qi Dai, Chunyu Wang, Dacheng Yin, Zhiyuan Zhao, Kai Qiu, Jianmin Bao, Yuhui Yuan, et al. Art-v: Auto-regressive text-to-video generation with diffusion models. In CVPR, 2024. find
  90. Tong Wu, Zhihao Fan, Xiao Liu, Hai-Tao Zheng, Yeyun Gong, Jian Jiao, Juntao Li, Jian Guo, Nan Duan, Weizhu Chen, et al. Ar-diffusion: Auto-regressive diffusion model for text generation. In NeurIPS, 2023. find
  91. Yushu Wu, Zhixing Zhang, Yanyu Li, Yanwu Xu, Anil Kag, Yang Sui, Huseyin Coskun, Ke Ma, Aleksei Lebedev, Ju Hu, et al. Snapgen-v: Generating a five-second video within five seconds on a mobile device. In CVPR, 2025. find
  92. Guangxuan Xiao, Yuandong Tian, Beidi Chen, Song Han, and Mike Lewis. Efficient streaming language models with attention sinks. In ICLR, 2024. find
  93. Desai Xie, Zhan Xu, Yicong Hong, Hao Tan, Difan Liu, Feng Liu, Arie Kaufman, and Yang Zhou. Progressive autoregressive video diffusion models. arXiv preprint arXiv:2410.08151, 2024. — [arXiv:2410.08151](https://arxiv.org/abs/2410.08151) find
  94. Wilson Yan, Yunzhi Zhang, Pieter Abbeel, and Aravind Srinivas. Videogpt: Video generation using vq-vae and transformers. arXiv preprint arXiv:2104.10157, 2021. — [arXiv:2104.10157](https://arxiv.org/abs/2104.10157) find
  95. An Yang, Baosong Yang, Beichen Zhang, Binyuan Hui, Bo Zheng, Bowen Yu, Chengyuan Li, Dayiheng Liu, Fei Huang, Haoran Wei, et al. Qwen2. 5 technical report. arXiv preprint arXiv:2412.15115, 2024. find
  96. Mengjiao Yang, Yilun Du, Kamyar Ghasemipour, Jonathan Tompson, Dale Schuurmans, and Pieter Abbeel. Learning interactive real-world simulators. In ICLR, 2024. find
  97. Zhuoyi Yang, Jiayan Teng, Wendi Zheng, Ming Ding, Shiyu Huang, Jiazheng Xu, Yuanming Yang, Wenyi Hong, Xiaohan Zhang, Guanyu Feng, et al. Cogvideox: Text-to-video diffusion models with an expert transformer. In ICLR, 2025.
  98. Tianwei Yin, Michaël Gharbi, Taesung Park, Richard Zhang, Eli Shechtman, Fredo Durand, and Bill Freeman. Improved distribution matching distillation for fast image synthesis. NeurIPS, 2024. find
  99. Tianwei Yin, Michaël Gharbi, Richard Zhang, Eli Shechtman, Fredo Durand, William T Freeman, and Taesung Park. One-step diffusion with distribution matching distillation. In CVPR, 2024. find
  100. Tianwei Yin, Qiang Zhang, Richard Zhang, William T Freeman, Fredo Durand, Eli Shechtman, and Xun Huang. From slow bidirectional to fast autoregressive video diffusion models. In CVPR, 2025.
  101. Alan Yu, Ge Yang, Ran Choi, Yajvan Ravan, John Leonard, and Phillip Isola. Learning visual parkour from generated images. In CoRL, 2024. find
  102. Jiwen Yu, Yiran Qin, Xintao Wang, Pengfei Wan, Di Zhang, and Xihui Liu. Gamefactory: Creating new games with generative interactive videos. arXiv preprint arXiv:2501.08325, 2025. — [arXiv:2501.08325](https://arxiv.org/abs/2501.08325) find
  103. Lantao Yu, Weinan Zhang, Jun Wang, and Yong Yu. Seqgan: Sequence generative adversarial nets with policy gradient. In AAAI, 2017. find
  104. Lijun Yu, José Lezama, Nitesh B Gundavarapu, Luca Versari, Kihyuk Sohn, David Minnen, Yong Cheng, Vighnesh Birodkar, Agrim Gupta, Xiuye Gu, et al. Language model beats diffusion–tokenizer is key to visual generation. In ICLR, 2024. find
  105. Lvmin Zhang and Maneesh Agrawala. Packing input frame context in next-frame prediction models for video generation. arXiv preprint arXiv:2504.12626, 2025. — [arXiv:2504.12626](https://arxiv.org/abs/2504.12626) find
  106. Tianyuan Zhang, Sai Bi, Yicong Hong, Kai Zhang, Fujun Luan, Songlin Yang, Kalyan Sunkavalli, William T Freeman, and Hao Tan. Test-time training done right. arXiv preprint arXiv:2505.23884, 2025. find
  107. Yuan Zhang, Jiacheng Jiang, Guoqing Ma, Zhiying Lu, Haoyang Huang, Jianlong Yuan, and Nan Duan. Generative pre-trained autoregressive diffusion transformer. arXiv preprint arXiv:2505.07344, 2025. — [arXiv:2505.07344](https://arxiv.org/abs/2505.07344) find
  108. Zhixing Zhang, Yanyu Li, Yushu Wu, Anil Kag, Ivan Skorokhodov, Willi Menapace, Aliaksandr Siarohin, Junli Cao, Dimitris Metaxas, Sergey Tulyakov, et al. Sf-v: Single forward video generation model. In NeurIPS, 2024. find
  109. Xuanlei Zhao, Xiaolong Jin, Kai Wang, and Yang You. Real-time video generation with pyramid attention broadcast. In ICLR, 2025. find
  110. Chunting Zhou, Lili Yu, Arun Babu, Kushal Tirumala, Michihiro Yasunaga, Leonid Shamis, Jacob Kahn, Xuezhe Ma, Luke Zettlemoyer, and Omer Levy. Transfusion: Predict the next token and diffuse images with one multi-modal model. In ICLR, 2025.
  111. Deyu Zhou, Quan Sun, Yuang Peng, Kun Yan, Runpei Dong, Duomin Wang, Zheng Ge, Nan Duan, Xiangyu Zhang, Lionel M Ni, et al. Taming teacher forcing for masked autoregressive video generation. In CVPR, 2025. find
  112. Mingyuan Zhou, Huangjie Zheng, Yi Gu, Zhendong Wang, and Hai Huang. Adversarial score identity distillation: Rapidly surpassing the teacher in one step. In ICLR, 2025. find
  113. Mingyuan Zhou, Huangjie Zheng, Zhendong Wang, Mingzhang Yin, and Hai Huang. Score identity distillation: Exponentially fast distillation of pretrained diffusion models for one-step generation. In ICML, 2024. find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/huang-2025-self-forcing-bridging-train-test/