pith. sign in

arxiv: 2605.25801 · v1 · pith:YYIMVXRZnew · submitted 2026-05-25 · 💻 cs.CV

PixelWizard: Towards Efficient High-Fidelity Video Generation at Ultra-Large Spatial Resolution

Wenxue Li , Jingjing Ren , Peng Zhang , Tian Ye , Daiguo Zhou , Jian Luan , Lei Zhu This is my paper

Pith reviewed 2026-06-29 22:22 UTC · model grok-4.3

classification 💻 cs.CV
keywords video generationhigh resolutiondiffusion modelsefficient inferencespatiotemporal anchorshortcut trainingnoise scheduling
0
0 comments X

The pith

PixelWizard decouples global structure from local details to enable over 10x faster generation of native 2K and 4K videos.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper aims to overcome the instability and high costs of generating high-resolution videos by separating the modeling of overall structure from the synthesis of fine details. It first builds a compact spatiotemporal anchor that captures dense structural information to steer the high-resolution generation process. This prevents the optimization from favoring local textures over global coherence. The approach then uses a shortcut training method with aligned sampling to allow the model to skip many steps during inference. Readers would care because it promises high-quality video at large scales with much lower computational demands.

Core claim

PixelWizard hierarchically decouples global structure modeling from fine-grained detail synthesis. It establishes a compact spatiotemporal anchor to concentrate dense structural priors which then guides fine-grained generation at high resolution to mitigate local optimization bias. Leveraging this, Noise-Span Aligned Shortcut Training with Exponential Index-Biased Sampling and Adaptive Noise-Span Calibration enables the model to traverse the generation trajectory with large steps for robust few-step inference, resulting in superior visual quality and over 10x acceleration for native 2K/4K videos.

What carries the argument

Compact spatiotemporal anchor for concentrating structural priors to guide high-resolution generation, combined with Noise-Span Aligned Shortcut Training for large-step inference.

If this is right

  • High-resolution video generation avoids structural collapse from local texture bias.
  • Generative sampling for 2K and 4K videos accelerates by over 10 times.
  • Few-step inference becomes robust without the need for heavy distillation techniques.
  • High-frequency details remain preserved while maintaining structural stability.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • This hierarchical decoupling could be tested on other high-dimensional generation tasks like 3D scenes.
  • The sampling alignment methods may reduce training time in related diffusion models.
  • Applications in real-time video synthesis become more feasible if the speedup holds across datasets.

Load-bearing premise

That a compact spatiotemporal anchor can reliably concentrate dense structural priors to guide fine-grained high-resolution generation without introducing new instabilities or losing high-frequency details.

What would settle it

Conducting inference on 4K video generation and finding that the speedup falls below 10x or that quality metrics are lower than standard methods.

Figures

Figures reproduced from arXiv: 2605.25801 by Daiguo Zhou, Jian Luan, Jingjing Ren, Lei Zhu, Peng Zhang, Tian Ye, Wenxue Li.

Figure 9
Figure 9. Figure 9 [PITH_FULL_IMAGE:figures/full_fig_p020_9.png] view at source ↗
Figure 9
Figure 9. Figure 9 [PITH_FULL_IMAGE:figures/full_fig_p021_9.png] view at source ↗
Figure 9
Figure 9. Figure 9 [PITH_FULL_IMAGE:figures/full_fig_p022_9.png] view at source ↗
Figure 9
Figure 9. Figure 9 [PITH_FULL_IMAGE:figures/full_fig_p023_9.png] view at source ↗
Figure 9
Figure 9. Figure 9 [PITH_FULL_IMAGE:figures/full_fig_p024_9.png] view at source ↗
Figure 9
Figure 9. Figure 9 [PITH_FULL_IMAGE:figures/full_fig_p025_9.png] view at source ↗
Figure 9
Figure 9. Figure 9 [PITH_FULL_IMAGE:figures/full_fig_p026_9.png] view at source ↗
Figure 9
Figure 9. Figure 9 [PITH_FULL_IMAGE:figures/full_fig_p027_9.png] view at source ↗
Figure 9
Figure 9. Figure 9 [PITH_FULL_IMAGE:figures/full_fig_p028_9.png] view at source ↗
Figure 9
Figure 9. Figure 9 [PITH_FULL_IMAGE:figures/full_fig_p029_9.png] view at source ↗
Figure 9
Figure 9. Figure 9 [PITH_FULL_IMAGE:figures/full_fig_p030_9.png] view at source ↗
Figure 9
Figure 9. Figure 9 [PITH_FULL_IMAGE:figures/full_fig_p031_9.png] view at source ↗
Figure 9
Figure 9. Figure 9 [PITH_FULL_IMAGE:figures/full_fig_p032_9.png] view at source ↗
Figure 9
Figure 9. Figure 9 [PITH_FULL_IMAGE:figures/full_fig_p033_9.png] view at source ↗
Figure 9
Figure 9. Figure 9 [PITH_FULL_IMAGE:figures/full_fig_p034_9.png] view at source ↗
read the original abstract

High-resolution video generation faces a coupled bottleneck of optimization instability and prohibitive computational costs. The massive expansion of the token sequence not only biases optimization toward local textures at the expense of global coherence, leading to structural collapse, but also imposes prohibitive training costs and severe inference latency. To address this, we propose PixelWizard, a framework that hierarchically decouples global structure modeling from fine-grained detail synthesis. PixelWizard first establishes a compact spatiotemporal anchor to concentrate dense structural priors, which then guides fine-grained generation at high resolution. This mitigates the local optimization bias to ensure structural stability without compromising high-frequency details. Leveraging this structural stability, we introduce Noise-Span Aligned Shortcut Training to break the inference bottleneck. By explicitly modeling the step size, this mechanism allows the model to traverse the generation trajectory with large steps. Crucially, we incorporate Exponential Index-Biased Sampling and Adaptive Noise-Span Calibration to align optimization with the shifted noise schedules of high-resolution grids, ensuring robust few-step inference without incurring the heavy overhead of distillation. Extensive experiments demonstrate that PixelWizard achieves superior visual quality while accelerating the generative sampling of native 2K/4K videos by over 10x.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 1 minor

Summary. The manuscript presents PixelWizard, a framework for efficient high-fidelity video generation at ultra-large spatial resolutions such as 2K and 4K. It hierarchically decouples global structure modeling from fine-grained detail synthesis by establishing a compact spatiotemporal anchor to concentrate structural priors, which guides the high-resolution generation. This is combined with Noise-Span Aligned Shortcut Training, Exponential Index-Biased Sampling, and Adaptive Noise-Span Calibration to enable robust few-step inference, claiming to achieve superior visual quality and over 10x acceleration in generative sampling without distillation.

Significance. If the results hold, the work would be significant for the field of generative AI, particularly in video synthesis, by addressing key bottlenecks in optimization stability and computational efficiency at high resolutions. The approach of using shortcut training without distillation could offer a more efficient alternative to existing methods for few-step generation.

major comments (2)
  1. [Abstract] Abstract: The central empirical claim that PixelWizard accelerates generative sampling of native 2K/4K videos by over 10x while achieving superior visual quality is not supported by any quantitative results, ablation tables, or experimental details in the provided manuscript text, which prevents verification of the claims.
  2. [Abstract] Abstract: The description of the compact spatiotemporal anchor states that it 'concentrate[s] dense structural priors' and 'mitigates the local optimization bias,' but provides no mechanism details, scaling analysis, or evidence that it remains stable at ultra-large resolutions without introducing instabilities or losing high-frequency details, which is critical for the hierarchical decoupling to succeed.
minor comments (1)
  1. [Abstract] Abstract: The abstract mentions 'Extensive experiments demonstrate' but without referencing specific sections or figures where these results are presented.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive comments on our manuscript. We address each major point below and outline revisions to improve clarity and verifiability of the claims.

read point-by-point responses

  1. Referee: [Abstract] Abstract: The central empirical claim that PixelWizard accelerates generative sampling of native 2K/4K videos by over 10x while achieving superior visual quality is not supported by any quantitative results, ablation tables, or experimental details in the provided manuscript text, which prevents verification of the claims.

    Authors: The abstract summarizes the key empirical outcomes from our experiments. The full manuscript contains the supporting quantitative results, including speedup measurements and quality metrics, in the Experiments section along with corresponding tables. To directly address the concern and enable verification from the abstract itself, we will revise the abstract to incorporate specific quantitative highlights (e.g., reported speedup factors and quality metrics) drawn from those experiments. revision: yes

  2. Referee: [Abstract] Abstract: The description of the compact spatiotemporal anchor states that it 'concentrate[s] dense structural priors' and 'mitigates the local optimization bias,' but provides no mechanism details, scaling analysis, or evidence that it remains stable at ultra-large resolutions without introducing instabilities or losing high-frequency details, which is critical for the hierarchical decoupling to succeed.

    Authors: The abstract provides a high-level overview of the anchor's role. The full manuscript details the mechanism in Section 3, including how the compact spatiotemporal anchor is constructed and its effect on optimization. Scaling behavior and stability at 2K/4K resolutions are examined through experiments and ablations in the main text and appendix. We agree that the abstract would benefit from a concise reference to these elements; we will add a brief clause on the mechanism and stability evidence to the abstract in revision. revision: yes

Circularity Check

0 steps flagged

No circularity; claims rest on empirical validation of proposed methods

full rationale

The paper advances a hierarchical decoupling architecture (compact spatiotemporal anchor + Noise-Span Aligned Shortcut Training + calibration) whose performance claims—superior 2K/4K quality and >10x few-step sampling—are presented strictly as outcomes of experiments rather than any closed-form derivation, fitted parameter renamed as prediction, or self-citation chain. No equations appear in the abstract or description; the anchor's role is described as an engineering choice whose stability is asserted via ablation and scaling results, not by construction. The derivation chain is therefore self-contained against external benchmarks and does not reduce to its own inputs.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract-only review supplies no explicit free parameters, axioms, or invented entities; the central claim rests on the unstated assumption that the hierarchical anchor and shortcut training can be implemented without hidden fitting costs or post-hoc tuning.

pith-pipeline@v0.9.1-grok · 5753 in / 1193 out tokens · 24108 ms · 2026-06-29T22:22:08.375667+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

52 extracted references · 23 canonical work pages · 10 internal anchors

  1. [1]

    Scalable diffusion models with transformers,

    W. Peebles and S. Xie, “Scalable diffusion models with transformers,” inProceedings of the IEEE/CVF international conference on computer vision, 2023, pp. 4195–4205. 11 PixelWizard

    2023

  2. [2]

    Kling 2.5 Turbo,

    Kuaishou Technology, “Kling 2.5 Turbo,” https://app.klingai.com/cn/release-notes/, 2025, accessed: 2025-09-19

    2025

  3. [3]

    Veo 3.1,

    Google DeepMind, “Veo 3.1,” https://deepmind.google/technologies/veo/, 2025

    2025

  4. [4]

    OpenAI, “Sora 2,” https://openai.com/sora, 2025

    2025

  5. [5]

    HunyuanVideo 1.5 Technical Report

    B. Wu, C. Zou, C. Li, D. Huang, F. Yang, H. Tan, J. Peng, J. Wu, J. Xiong, J. Jianget al., “Hunyuanvideo 1.5 technical report,”arXiv preprint arXiv:2511.18870, 2025

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  6. [6]

    LTX-2: Efficient Joint Audio-Visual Foundation Model

    Y. HaCohen, B. Brazowski, N. Chiprut, Y. Bitterman, A. Kvochko, A. Berkowitz, D. Shalem, D. Lifschitz, D. Moshe, E. Poratet al., “Ltx-2: Efficient joint audio-visual foundation model,” arXiv preprint arXiv:2601.03233, 2026

    work page internal anchor Pith review Pith/arXiv arXiv 2026

  7. [7]

    Waver: Wave your way to lifelike video generation,

    Y. Zhang, H. Yang, Y. Zhang, Y. Hu, F. Zhu, C. Lin, X. Mei, Y. Jiang, B. Peng, and Z. Yuan, “Waver: Wave your way to lifelike video generation,”arXiv preprint arXiv:2508.15761, 2025

    work page arXiv 2025

  8. [8]

    LTX-Video: Realtime Video Latent Diffusion

    Y. HaCohen, N. Chiprut, B. Brazowski, D. Shalem, D. Moshe, E. Richardson, E. Levin, G. Shiran, N. Zabari, O. Gordonet al., “Ltx-video: Realtime video latent diffusion,”arXiv preprint arXiv:2501.00103, 2025

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  9. [9]

    Ultravideo: High-quality uhd video dataset with comprehensive captions,

    Z. Xue, J. Zhang, T. Hu, H. He, Y. Chen, Y. Cai, Y. Wang, C. Wang, Y. Liu, X. Li, and D. Tao, “Ultravideo: High-quality uhd video dataset with comprehensive captions,” inNeurIPS, 2025

    2025

  10. [10]

    Ultragen: High-resolution video generation with hierar- chical attention,

    T. Hu, J. Zhang, Z. Su, and R. Yi, “Ultragen: High-resolution video generation with hierar- chical attention,”arXiv preprint arXiv:2510.18775, 2025

    work page arXiv 2025

  11. [11]

    Freeswim: Revisiting sliding-window atten- tion mechanisms for training-free ultra-high-resolution video generation,

    Y. Wu, J. Song, Z. Tan, Z. He, and S. Liu, “Freeswim: Revisiting sliding-window atten- tion mechanisms for training-free ultra-high-resolution video generation,”arXiv preprint arXiv:2511.14712, 2025

    work page arXiv 2025

  12. [12]

    Freescale: Unleashing the resolution of diffusion models via tuning-free scale fusion,

    H. Qiu, S. Zhang, Y. Wei, R. Chu, H. Yuan, X. Wang, Y. Zhang, and Z. Liu, “Freescale: Unleashing the resolution of diffusion models via tuning-free scale fusion,” inProceedings of the IEEE/CVF International Conference on Computer Vision, 2025, pp. 16893–16903

    2025

  13. [13]

    Flashvideo: Flowing fidelity to detail for efficient high-resolution video generation,

    S. Zhang, W. Li, S. Chen, C. Ge, P. Sun, Y. Zhang, Y. Jiang, Z. Yuan, B. Peng, and P. Luo, “Flashvideo: Flowing fidelity to detail for efficient high-resolution video generation,” inAAAI, 2026

    2026

  14. [14]

    Turbo2k: Towards ultra-efficient and high-quality 2k video synthesis,

    J. Ren, W. Li, Z. Wang, H. Sun, B. Liu, H. Chen, J. Xu, A. Li, S. Zhang, B. Shao, Y. Guo, and L. Zhu, “Turbo2k: Towards ultra-efficient and high-quality 2k video synthesis,” in Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), 2025, pp. 18155–18165

    2025

  15. [15]

    Cinescale: Free lunch in high-resolution cinematic visual generation,

    H. Qiu, N. Yu, Z. Huang, P. Debevec, and Z. Liu, “Cinescale: Free lunch in high-resolution cinematic visual generation,”arXiv preprint arXiv:2508.15774, 2025

    work page arXiv 2025

  16. [16]

    Scalecrafter: Tuning-free higher-resolution visual generation with diffusion models,

    Y. He, S. Yang, H. Chen, X. Cun, M. Xia, Y. Zhang, X. Wang, R. He, Q. Chen, and Y. Shan, “Scalecrafter: Tuning-free higher-resolution visual generation with diffusion models,” inThe Twelfth International Conference on Learning Representations, 2023

    2023

  17. [17]

    Ultra-resolution adaptation with ease,

    R. Yu, S. Liu, Z. Tan, and X. Wang, “Ultra-resolution adaptation with ease,”International Conference on Machine Learning, 2025

    2025

  18. [18]

    Ultraflux: Data-model co-design for high-quality native 4k text-to- image generation across diverse aspect ratios,

    T. Ye, S. Fei, and L. Zhu, “Ultraflux: Data-model co-design for high-quality native 4k text-to- image generation across diverse aspect ratios,”arXiv preprint arXiv:2511.18050, 2025. 12 PixelWizard

    work page arXiv 2025

  19. [19]

    Star: Spatial-temporal augmentation with text-to-video models for real-world video super- resolution,

    R. Xie, Y. Liu, P. Zhou, C. Zhao, J. Zhou, K. Zhang, Z. Zhang, J. Yang, Z. Yang, and Y. Tai, “Star: Spatial-temporal augmentation with text-to-video models for real-world video super- resolution,” inICCV, 2025

    2025

  20. [20]

    Seedvr: Seeding infinity in diffusion transformer towards generic video restoration,

    J. Wang, Z. Lin, M. Wei, Y. Zhao, C. Yang, C. C. Loy, and L. Jiang, “Seedvr: Seeding infinity in diffusion transformer towards generic video restoration,” inCVPR, 2025

    2025

  21. [21]

    Seedvr2: One-step video restoration via diffusion adversarial post-training,

    J. Wang, S. Lin, Z. Lin, Y. Ren, M. Wei, Z. Yue, S. Zhou, H. Chen, Y. Zhao, C. Yang, X. Xiao, C. C. Loy, and L. Jiang, “Seedvr2: One-step video restoration via diffusion adversarial post-training,”arXiv preprint arXiv:2506.05301, 2025

    work page arXiv 2025

  22. [22]

    Dove: Efficient one-step diffusion model for real-world video super-resolution,

    Z. Chen, Z. Zou, K. Zhang, X. Su, X. Yuan, Y. Guo, and Y. Zhang, “Dove: Efficient one-step diffusion model for real-world video super-resolution,” inNeurIPS, 2025

    2025

  23. [23]

    Flashvsr: Towards real-time diffusion-based streaming video super-resolution,

    J. Zhuang, S. Guo, X. Cai, X. Li, Y. Liu, C. Yuan, and T. Xue, “Flashvsr: Towards real-time diffusion-based streaming video super-resolution,”arXiv preprint arXiv:2510.12747, 2025

    work page arXiv 2025

  24. [24]

    Simplegvr: A simple baseline for latent-cascaded video super-resolution,

    L. Xie, Y. Li, S. Du, M. Xia, X. Wang, F. Yu, Z. Chen, P. Wan, J. Zhou, and C. Dong, “Simplegvr: A simple baseline for latent-cascaded video super-resolution,”arXiv preprint arXiv:2506.19838, 2025

    work page arXiv 2025

  25. [25]

    Vivid-vr: Distilling concepts from text-to-video diffusion transformer for photorealistic video restoration,

    H. Bai, X. Chen, C. Yang, Z. He, S. Deng, and Y. Chen, “Vivid-vr: Distilling concepts from text-to-video diffusion transformer for photorealistic video restoration,”arXiv preprint arXiv:2508.14483, 2025. [Online]. Available: https://arxiv.org/abs/2508.14483

    work page internal anchor Pith review arXiv 2025

  26. [26]

    Histream: Efficient high-resolution video generation via redundancy-eliminated streaming,

    H. Qiu, S. Liu, Z. Zhou, Z. An, W. Ren, Z. Liu, J. Schult, S. He, S. Chen, Y. Conget al., “Histream: Efficient high-resolution video generation via redundancy-eliminated streaming,” arXiv preprint arXiv:2512.21338, 2025

    work page arXiv 2025

  27. [27]

    One-step diffusion with distribution matching distillation,

    T. Yin, M. Gharbi, R. Zhang, E. Shechtman, F. Durand, W. T. Freeman, and T. Park, “One-step diffusion with distribution matching distillation,” inProceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2024, pp. 6613–6623

    2024

  28. [28]

    Dual-expertconsistencymodel for efficient and high-quality video generation,

    Z.Lv, C.Si, T.Pan, Z.Chen, K.-Y.K.Wong, Y.Qiao, andZ.Liu, “Dual-expertconsistencymodel for efficient and high-quality video generation,” inProceedings of the IEEE/CVF International Conference on Computer Vision, 2025, pp. 14983–14993

    2025

  29. [29]

    Osv: One step is enough for high-quality image to video generation,

    X. Mao, Z. Jiang, F.-Y. Wang, J. Zhang, H. Chen, M. Chi, Y. Wang, and W. Luo, “Osv: One step is enough for high-quality image to video generation,” inProceedings of the Computer Vision and Pattern Recognition Conference, 2025, pp. 12585–12594

    2025

  30. [30]

    Self forcing: Bridging the train-test gap in autoregressive video diffusion,

    X. Huang, Z. Li, G. He, M. Zhou, and E. Shechtman, “Self forcing: Bridging the train-test gap in autoregressive video diffusion,” inThe Thirty-ninth Annual Conference on Neural Information Processing Systems, 2025

    2025

  31. [31]

    Magicdistilla- tion: Weak-to-strong video distillation for large-scale few-step synthesis,

    S. Shao, H. Yi, H. Guo, T. Ye, D. Zhou, M. Lingelbach, Z. Xu, and Z. Xie, “Magicdistilla- tion: Weak-to-strong video distillation for large-scale few-step synthesis,”arXiv preprint arXiv:2503.13319, 2025

    work page arXiv 2025

  32. [32]

    Timestep embedding tells: It’s time to cache for video diffusion model,

    F. Liu, S. Zhang, X. Wang, Y. Wei, H. Qiu, Y. Zhao, Y. Zhang, Q. Ye, and F. Wan, “Timestep embedding tells: It’s time to cache for video diffusion model,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024. 13 PixelWizard

    2024

  33. [33]

    Magcache: Fast video generation with magnitude-aware cache,

    Z. Ma, L. Wei, F. Wang, S. Zhang, and Q. Tian, “Magcache: Fast video generation with magnitude-aware cache,” inThe Thirty-ninth Annual Conference on Neural Information Processing Systems, 2025

    2025

  34. [34]

    Less is enough: Training-free video diffusion acceleration via runtime-adaptive caching,

    X. Zhou, D. Liang, K. Chen, , T. Feng, X. Chen, H. Lin, Y. Ding, F. Tan, H. Zhao, and X. Bai, “Less is enough: Training-free video diffusion acceleration via runtime-adaptive caching,” arXiv preprint arXiv:2507.02860, 2025

    work page arXiv 2025

  35. [35]

    Sparsevideogen: Accelerating video diffusion transformers with spatial-temporal sparsity,

    H.Xi,S.Yang,Y.Zhao,C.Xu,M.Li,X.Li,Y.Lin,H.Cai,J.Zhang,D.Lietal.,“Sparsevideogen: Accelerating video diffusion transformers with spatial-temporal sparsity,”International Conference on Machine Learning, 2025

    2025

  36. [36]

    Sparse videogen2: Accelerate video generation with sparse attention via semantic-aware permu- tation,

    S. Yang, H. Xi, Y. Zhao, M. Li, J. Zhang, H. Cai, Y. Lin, X. Li, C. Xu, K. Penget al., “Sparse videogen2: Accelerate video generation with sparse attention via semantic-aware permu- tation,” inThe Thirty-ninth Annual Conference on Neural Information Processing Systems, 2025

    2025

  37. [37]

    Radial attention:O (𝑛log𝑛) sparse attention with energy decayforlongvideogeneration,

    X. Li*, M. Li*, T. Cai, H. Xi, S. Yang, Y. Lin, L. Zhang, S. Yang, J. Hu, K. Peng, M. Agrawala, I. Stoica, K. Keutzer, and S. Han, “Radial attention:O (𝑛log𝑛) sparse attention with energy decayforlongvideogeneration,” inTheThirty-ninthAnnualConferenceonNeuralInformation Processing Systems, 2025

    2025

  38. [38]

    Dc- videogen: Efficient video generation with deep compression video autoencoder,

    J. Chen, W. He, Y. Gu, Y. Zhao, J. Yu, J. Chen, D. Zou, Y. Lin, Z. Zhang, M. Liet al., “Dc- videogen: Efficient video generation with deep compression video autoencoder,”arXiv preprint arXiv:2509.25182, 2025

    work page arXiv 2025

  39. [39]

    One Step Diffusion via Shortcut Models

    K. Frans, D. Hafner, S. Levine, and P. Abbeel, “One step diffusion via shortcut models,”arXiv preprint arXiv:2410.12557, 2024

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  40. [40]

    Scaling rectified flow transformers for high-resolution image synthesis,

    P. Esser, S. Kulal, A. Blattmann, R. Entezari, J. Müller, H. Saini, Y. Levi, D. Lorenz, A. Sauer, F. Boeselet al., “Scaling rectified flow transformers for high-resolution image synthesis,” in Forty-first international conference on machine learning, 2024

    2024

  41. [41]

    Wan: Open and Advanced Large-Scale Video Generative Models

    T. Wan, A. Wang, B. Ai, B. Wen, C. Mao, C.-W. Xie, D. Chen, F. Yu, H. Zhao, J. Yang, J. Zeng, J. Wang, J. Zhang, J. Zhou, J. Wang, J. Chen, K. Zhu, K. Zhao, K. Yan, L. Huang, M. Feng, N. Zhang, P. Li, P. Wu, R. Chu, R. Feng, S. Zhang, S. Sun, T. Fang, T. Wang, T. Gui, T. Weng, T. Shen, W. Lin, W. Wang, W. Wang, W. Zhou, W. Wang, W. Shen, W. Yu, X. Shi, X....

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  42. [42]

    Vbench: Comprehensive benchmark suite for video generative models,

    Z. Huang, Y. He, J. Yu, F. Zhang, C. Si, Y. Jiang, Y. Zhang, T. Wu, Q. Jin, N. Chanpaisitet al., “Vbench: Comprehensive benchmark suite for video generative models,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024, pp. 21807–21818

    2024

  43. [43]

    Exploring video quality assessment on user generated contents from aesthetic and technical perspectives,

    H. Wu, E. Zhang, L. Liao, C. Chen, J. H. Hou, A. Wang, W. S. Sun, Q. Yan, and W. Lin, “Exploring video quality assessment on user generated contents from aesthetic and technical perspectives,” inInternational Conference on Computer Vision (ICCV), 2023

    2023

  44. [44]

    Musiq: Multi-scale image quality transformer,

    J. Ke, Q. Wang, Y. Wang, P. Milanfar, and F. Yang, “Musiq: Multi-scale image quality transformer,” inProceedings of the IEEE/CVF international conference on computer vision, 2021, pp. 5148–5157. 14 PixelWizard

    2021

  45. [45]

    Making a “completely blind

    A. Mittal, R. Soundararajan, and A. C. Bovik, “Making a “completely blind” image quality analyzer,”IEEE Signal processing letters, vol. 20, no. 3, pp. 209–212, 2012

    2012

  46. [46]

    Flow straight and fast: Learning to generate and transfer data with rectified flow,

    X. Liu, C. Gong, and Q. Liu, “Flow straight and fast: Learning to generate and transfer data with rectified flow,”International Conference on Learning Representations, 2023

    2023

  47. [47]

    Vchitect-2.0: Parallel transformer for scaling up video diffusion models,

    W. Fan, C. Si, J. Song, Z. Yang, Y. He, L. Zhuo, Z. Huang, Z. Dong, J. He, D. Panet al., “Vchitect-2.0: Parallel transformer for scaling up video diffusion models,”arXiv preprint arXiv:2501.08453, 2025

    work page arXiv 2025

  48. [48]

    CogVideoX: Text-to-Video Diffusion Models with An Expert Transformer

    Z. Yang, J. Teng, W. Zheng, M. Ding, S. Huang, J. Xu, Y. Yang, W. Hong, X. Zhang, G. Feng et al., “Cogvideox: Text-to-video diffusion models with an expert transformer,”arXiv preprint arXiv:2408.06072, 2024

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  49. [49]

    Mochi 1,

    G. Team, “Mochi 1,” https://github.com/genmoai/models, 2024

    2024

  50. [50]

    Open-Sora: Democratizing Efficient Video Production for All

    Z. Zheng, X. Peng, T. Yang, C. Shen, S. Li, H. Liu, Y. Zhou, T. Li, and Y. You, “Open-sora: Democratizing efficient video production for all,”arXiv preprint arXiv:2412.20404, 2024

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  51. [51]

    Open-Sora Plan: Open-Source Large Video Generation Model

    B. Lin, Y. Ge, X. Cheng, Z. Li, B. Zhu, S. Wang, X. He, Y. Ye, S. Yuan, L. Chenet al., “Open-sora plan: Open-source large video generation model,”arXiv preprint arXiv:2412.00131, 2024

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  52. [52]

    HunyuanVideo: A Systematic Framework For Large Video Generative Models

    W. Kong, Q. Tian, Z. Zhang, R. Min, Z. Dai, J. Zhou, J. Xiong, X. Li, B. Wu, J. Zhanget al., “Hunyuanvideo: A systematic framework for large video generative models,”arXiv preprint arXiv:2412.03603, 2024. 15 PixelWizard This is supplementary material forPixelWizard: Towards Efficient High-Fidelity Video Genera- tion at Ultra-Large Spatial Resolutions. 6 O...

    work page internal anchor Pith review Pith/arXiv arXiv 2024