Pith. sign in

REVIEW 5 major objections 5 minor 1 cited by

CineScale: Free Lunch in High-Resolution Cinematic Visual Generation

T0 review · 5 major / 5 minor · reviewed 2026-08-05 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read CineScale claims pretrained diffusion models can generate 8k images with no fine-tuning and 4k video with minimal LoRA, by fusing global and local attention through a frequency split.

desk verdict A useful engineering recipe for high-res image/video generation, but the frequency-split mechanism is asserted, not validated, and the quantitative claims need more support. read the letter →

arxiv 2508.15774 v1 pith:DJR733ZO submitted 2025-08-21 cs.CV

Haonan Qiu , Ning Yu , Ziqi Huang , Paul Debevec , Ziwei Liu This is my paper
classification cs.CV
keywords diffusionmodelshigh-resolutiongenerationtuning-freeinferencescalefusionfrequencyseparationDiTLoRAfine-tuningtext-to-video
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

CineScale claims that pre-trained diffusion models can generate far beyond their training resolution without retraining, if the inference recipe compensates for the extra high-frequency information that appears when the model sees more pixels. The paper proposes an inference-time pipeline—self-cascade upscaling, restrained dilated convolution, and a Scale Fusion step that combines global and local attention through a frequency split—and reports that it removes the repetitive-object artifacts that plague naive upscaling. On image models it claims 8192×8192 generation with no fine-tuning, and on transformer-based video models it claims 4K (2176×3840) generation after only a small LoRA adaptation. The paper also claims the recipe extends the task scope beyond text-to-image and text-to-video to image-to-video and video-to-video, where prior tuning-free methods did not operate.

What carries the argument

The load-bearing object is Scale Fusion, the identity h_fusion = (h_global − G(h_global)) + G(h_local), where G is a Gaussian low-pass filter applied inside the self-attention layer. It fuses a global self-attention pass, which gathers object semantics across the whole latent, with shifted local self-attention patches, which preserve local detail, by keeping the global high frequencies and the local low frequencies. The paper also uses a tailored self-cascade upscaling schedule that re-noises and denoises progressively, a restrained dilated convolution that only widens receptive fields in down/mid blocks and early timesteps, and—for the DiT variant—NTK-RoPE positional encoding, attention tem

What would settle it

Generate a 4096×4096 image of a periodic texture (e.g., a brick wall or herringbone fabric) with a model trained at 1024×1024 under CineScale, and inspect the spectrum of the intermediate self-attention outputs at the highest scale. If the Gaussian blurred low-frequency component already contains the duplicated-object artifacts, or if removing the high-frequency residual fails to eliminate repetition, then Eq. (7)'s premise is violated. A simpler version: compare CineScale to a variant that replaces the Gaussian blur with an ideal low-pass filter at the same cut-off; large quality differences

Watch

Extended reading notes

Core claim

At the center of the paper is a claim about why high-resolution generation fails and how to fix it without training. The paper argues that when a diffusion model generates at a resolution beyond its training data, the latent representations carry an excess of high-frequency information, and that this excess accumulates into repetitive patterns. The proposed remedy, Scale Fusion, edits self-attention outputs with the identity h_fusion = high_freq(h_global) + low_freq(h_local): it takes the global attention output, removes its low-frequency content with a Gaussian blur to keep the high-frequency signal, and re-adds the Gaussian-blurred local attention output so the low-frequency semantic struc

Load-bearing premise

The load-bearing premise is that a Gaussian low-pass filter on attention outputs in latent space separates repetition-causing high-frequency signals from semantic structure; if repetition artifacts leak into low-frequency channels—or the blur destroys object-level structure—Scale Fusion loses its justification, and the paper reports no quantitative test of that frequency separation.

Editorial extensions

If this is right

  • Pre-trained image diffusion models trained at 1024×1024 can generate 8192×8192 outputs with no training, at quality the paper reports as better than prior tuning-free baselines on FID, KID, and IS.
  • Pre-trained video models can be pushed to 4K (2176×3840) with only a minimal LoRA adaptation, and the same pipeline extends to image-to-video and video-to-video, not just text-conditioned generation.
  • The method works across architectural families: UNet-based models (SDXL, VideoCrafter2) and DiT-based models (Wan) need different fixes, and the paper specifies which component addresses which failure mode—repetition for UNet, blur for DiT.
  • Users can control detail level spatially via a mask and can perform local semantic editing at high resolution by editing prompts during the cascade, which the paper demonstrates for both images and videos.
  • The recipe is compatible with distilled fast models such as SDXL-Turbo, so it is not restricted to slow full-step samplers.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • An implication the paper leaves implicit: the frequency-separation hypothesis predicts that Scale Fusion should be most effective when repetition artifacts are spectrally localized; a user could test this by generating a high-resolution image of a scene with a fine, periodic texture and checking whether the Gaussian-blur cut-off frequency needs to scale with the upsampling factor.
  • The success of NTK-RoPE and attention scaling in this setting suggests that long-context adaptation techniques from language models may transfer to any attention-based generative model with positional encodings, so the same recipe could apply to future non-DiT transformer generators, not merely the Wan backbone used here.
  • Because the cascade regenerates ambiguous low-resolution regions from the model's learned prior, CineScale is a generative upsampler rather than a faithful restorer; for applications that require pixel-level fidelity to an input frame, the paper's own comparisons with super-resolution methods indicate a user would still need an SR model.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

5 major / 5 minor

Summary. The paper proposes CineScale, an inference-time framework for generating images and videos at resolutions far above a model's training resolution. For UNet-based models it builds on the authors' prior FreeScale, combining self-cascade upscaling, restrained dilated convolution, and a Scale Fusion mechanism (Eq. 7) that mixes global and local attention outputs through a Gaussian-blur frequency split. For DiT-based models it adds NTK-RoPE, attention temperature scaling, noise shifting, and minimal LoRA fine-tuning. The method is evaluated on SDXL and VideoCrafter2 for text-to-image/text-to-video, and on Wan and ReCamMaster for text-to-video, image-to-video, and video-to-video, with claims of 8k image generation without fine-tuning and 4k video generation with minimal LoRA. Quantitative comparisons use FID/KID/FIDc/KIDc/IS, VBench metrics, user studies, and ablations.

Significance. If the central claims hold, the paper has clear practical significance: it offers a largely tuning-free recipe for extending open diffusion models to resolutions two orders of magnitude above training, and it expands high-resolution generation to I2V and V2V tasks that prior baselines do not cover. The paper's strengths are its breadth (multiple architectures, multiple tasks, several baselines, user studies, and ablations), its transparency about reusing the prior FreeScale method, and the reported time-cost comparisons. However, the quantitative support is weakened by the choice of reference distributions, the absence of quantitative evaluation at the headline 8k/4k resolutions, and the lack of direct validation of the frequency-separation hypothesis behind Eq. 7. The empirical claims are plausible but not yet fully established.

major comments (5)
  1. [§3.4, Eq. (7)] The core mechanism for eliminating repetition is the frequency split h_fusion = (h_global - G(h_global)) + G(h_local), where G is a Gaussian blur. The paper asserts that repetition artifacts are 'dispersed high-frequency signals' and that the low-pass-filtered local attention output contains the semantic structure. This is not demonstrated. Repeated objects (e.g., extra eyes or faces) contain substantial low-frequency content such as skin tone and coarse shape, so the split is not guaranteed to isolate repetition. The Gaussian blur operates on latent-space deep features, not on the decoded image spectrum, so the frequency interpretation in Eq. (7) needs direct evidence. Table 4 ablates Scale Fusion only as a whole and does not isolate the frequency-split hypothesis. Please add a spectral analysis, visualizations of G(h_local) and of the high-pass residual, or a controlled variant that sw
  2. [§5 (Evaluation Metrics), Tables 1, 2, 5] The quantitative protocol is non-standard: FID/KID/FVD are computed between the originally generated low-resolution outputs and the corresponding high-resolution outputs, not against a real high-resolution reference distribution. This measures the distributional similarity of high-res outputs to their own low-res counterparts, which rewards methods that preserve the low-res content and penalizes legitimate novel high-frequency detail. Consequently, the scores in Tables 1, 2, and 5 do not support an absolute claim of high-resolution quality. Please report metrics against a real high-resolution image/video dataset or otherwise clearly reframe the tables as measuring consistency with the low-resolution output, and justify why this is the appropriate criterion for 'superiority'.
  3. [Abstract, §5, Tables 1-4] The headline claim is '8k image generation without any fine-tuning,' yet all image-quantitative tables (Tables 1-4) evaluate at 2048x2048 and 4096x4096. The 8k result appears only as a single qualitative example (Fig. 4). Since 8k is a load-bearing claim, the paper should provide quantitative support at 8k, even if limited (e.g., a small set of prompts with FID/FIDc, or a dedicated user study), or at minimum explicitly state that the 8k claim is qualitative-only and discuss variability across prompts.
  4. [Table 1 (SDXL-DI row)] In Table 1, SDXL-DI achieves the best FIDc and KIDc at both 2048^2 and 4096^2. The text explains this by saying SDXL-DI tends to generate duplicated objects, so its crops may be closer to the reference images. That is a post hoc reinterpretation and undermines the validity of FIDc as a quality metric in this setup: a method that visibly produces repetitive artifacts is rewarded. If FIDc and KIDc are retained, the paper must either justify why these metrics are meaningful when the worst-behaving method wins, or replace/annotate them with a metric that better agrees with the qualitative and user-study outcomes.
  5. [§6.1, Fig. 18, Table 10] The 4k text-to-video and image-to-video results are presented only qualitatively (Figs. 18, 19). The quantitative VBench tables for DiT models are at 960x1664 (Tables 9, 10, 12) or 1088x1920; the 1920x3328 table (Table 11) reports only three metrics. The LoRA fine-tuning recipe (about 20,000 videos, rank/alpha, steps) is not ablated, and no comparison is made against full fine-tuning or against a high-resolution-trained model of comparable size. Please add quantitative 4k evaluation (e.g., a subsampled VBench run or a user study at 4k) and ablate the LoRA configuration, or explicitly state that the 4k claim is qualitative-only.
minor comments (5)
  1. [§4.2] There are several presentation issues: 'destroy their signa [64]' should read 'signal'; the symbol 'λβ' is introduced but never defined; and Eq. (9) is malformed in the text ('softmax( qTmkn t p |D| )'). These need correction.
  2. [Figures 12, 15, 16 captions] The captions contain sentence fragments: 'Although all variants can generate rough results. Our full method performs the best.' Please rewrite as complete sentences.
  3. [§5.4, Tables 6, 7] The user study reports 23 participants for images and an unspecified number for video, with no confidence intervals or statistical significance tests. Given the large claimed margins, adding per-aspect vote counts and a basic significance test would strengthen the conclusions.
  4. [§4.2, §4.3] Several hyperparameters are not specified: the attention temperature t in Eq. (9), the noise shift schedule, the LoRA rank/alpha and number of training steps, and the exact data-filtering procedure for the 20,000 Pexels videos. Please include these for reproducibility.
  5. [Table 11] The statement that 'only three metrics in VBench can be measured due to memory limitations' is vague. Specify which metrics, why they could not be computed, and the subsampling procedure used for the reported metrics.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: Eq. 7 is an empirical frequency-separation hypothesis validated by ablations, and the reused FreeScale method is disclosed rather than disguised as a derivation.

full rationale

I find no step in which a predicted quantity is defined in terms of the input or in which a fitted parameter is relabeled as a prediction. Eq. 7 defines h_fusion as (h_global - G(h_global)) + G(h_local), where G is a Gaussian low-pass filter; the paper's claim that repetition artifacts arise from dispersed high-frequency signals is an explicit empirical hypothesis, not a consequence of the definition. The paper tests this hypothesis by ablating Scale Fusion (Table 4) and by comparing against external baselines (ScaleCrafter, DemoFusion, FouriScale, Wan-DI, LTX, SeedVR2, Upscale-A-Video) on standard metrics. The NTK-RoPE and attention-scaling modifications are imported from external LLM work (YaRN), and the LoRA fine-tuning stage is disclosed and trained on 2K-resolution videos, not on the 4K evaluation outputs. The paper does reuse the authors' prior FreeScale [16] as the UNet backbone, but this reuse is explicitly disclosed in the Introduction and Section 4, and the central new claim—DiT-based high-resolution generation—is evaluated against external baselines with VBench metrics. The main weakness of the paper is that the frequency-separation premise is under-validated (a correctness/robustness concern, not circularity), and the self-citation to FreeScale is a disclosure/attribution matter rather than a load-bearing circular derivation.

Assumptions & free parameters 8 free parameters · 4 assumptions · 0 invented entities

The method depends on a web of unstated hyperparameters and heuristic frequency-domain assumptions. Most design choices (alpha, K, patch size, blur sigma, attention temperature, RoPE base, noise shift, LoRA configuration) are not quantified in the paper. There are no invented entities. The only 'input' beyond the pretrained backbones is the choice of these knobs and the 20k-video LoRA set.

free parameters (8)
  • Detail control exponent alpha
    Eq. 4; scaled cosine decay; alpha (scalar or 2D tensor) is user-chosen, no selection rule or value reported.
  • Cascade noise timestep K
    Eq. 3; level of noise injected at each upscale stage; value not specified.
  • Local attention patch size and stride (h,w,dh,dw)
    Section 3.4; defines local crops for scale fusion; no values reported.
  • Gaussian blur sigma (Scale Fusion low-pass G)
    Eq. 7; bandwidth of frequency split; no value reported.
  • Attention temperature t
    Eq. 9; 'moderate value' chosen, not quantified.
  • NTK-RoPE base scaling factor lambda_beta
    Eq. 8; chosen for positional encoding extrapolation; not reported.
  • Noise shift amounts per resolution
    Section 4.2; gradually increasing noise shift; exact schedule absent.
  • LoRA rank, alpha, training steps, data filtering
    Section 4.3; 'minimal' LoRA on roughly 20,000 Pexels videos; rank, alpha, steps, and filtering criteria not stated.
assumptions (4)
  • domain assumption Pre-trained models (SDXL, VideoCrafter2, Wan) generate coherent outputs at their native training resolution.
    Sections 5 and 6; the entire cascade starts from a native-resolution generation and assumes this output is structurally sound.
  • domain assumption High-frequency content increase is the cause of repetition artifacts in UNet models; token dilution and positional errors are the cause of blur in DiT models.
    Introduction and Figure 2; asserted without measurement; motivates dilated convolution and frequency fusion.
  • ad hoc to paper Gaussian blur separation in deep attention features cleanly separates repetition-causing high frequencies from semantic low frequencies (Eq. 7).
    Section 3.4; this is the load-bearing heuristic of Scale Fusion.
  • domain assumption NTK-RoPE and attention temperature correction transfer from LLM long-context generation to DiT diffusion models.
    Section 4.2; borrowed from YaRN [62] and applied to Wan without re-derivation.

how reviews work

0 comments
Cite this review

Pith. "Pith review of CineScale: Free Lunch in High-Resolution Cinematic Visual Generation." pith.science (2026). https://pith.science/paper/DJR733ZO

@misc{pith2026250815774,
  author       = {Pith},
  title        = {Pith review of: CineScale: Free Lunch in High-Resolution Cinematic Visual Generation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/DJR733ZO}},
  note         = {Machine review of arXiv:2508.15774}
}
read the original abstract

Visual diffusion models achieve remarkable progress, yet they are typically trained at limited resolutions due to the lack of high-resolution data and constrained computation resources, hampering their ability to generate high-fidelity images or videos at higher resolutions. Recent efforts have explored tuning-free strategies to exhibit the untapped potential higher-resolution visual generation of pre-trained models. However, these methods are still prone to producing low-quality visual content with repetitive patterns. The key obstacle lies in the inevitable increase in high-frequency information when the model generates visual content exceeding its training resolution, leading to undesirable repetitive patterns deriving from the accumulated errors. In this work, we propose CineScale, a novel inference paradigm to enable higher-resolution visual generation. To tackle the various issues introduced by the two types of video generation architectures, we propose dedicated variants tailored to each. Unlike existing baseline methods that are confined to high-resolution T2I and T2V generation, CineScale broadens the scope by enabling high-resolution I2V and V2V synthesis, built atop state-of-the-art open-source video generation frameworks. Extensive experiments validate the superiority of our paradigm in extending the capabilities of higher-resolution visual generation for both image and video models. Remarkably, our approach enables 8k image generation without any fine-tuning, and achieves 4k video generation with only minimal LoRA fine-tuning. Generated video samples are available at our website: https://eyeline-labs.github.io/CineScale/.

Figures

Figures reproduced from arXiv: 2508.15774 by the authors.

Figure 1
Figure 1. Overall framework of CineScale. (a) Tailored Self-Cascade Upscaling. CineScale first upsamples a generated image or video from the training resolution, then gradually adds noise to the high-resolution latent, and finally denoises it to achieve detail reconstruction. Part of the clean latent is reintroduced during denoising to stabilize generation and control detail. (b) Scale Fusion. For the UNet structure, we modif… view at source ↗
Figure 2
Figure 2. Structure gap. UNet-based LDMs and DiT-based LDMs will face different challenges in the higher-resolution generation task. UNet￾based LDMs face repetition problems while DiT-based LDMs face blur problems. generation but help to suppress redundant over-frequency information in the image generation. Therefore, we adopt upsampling in RGB space for higher-resolution image gen￾eration and latent space upsampling in highe… view at source ↗
Figure 3
Figure 3. Image qualitative comparisons with other baselines. Our method generates both 20482 and 40962 vivid images with better content coherence and local details. signals in the local representations with those from the global level h global out : h fusion out = h global out − G  h global out  | {z } high frequency + G  h local out  | {z } low frequency , (7) where G is a low-pass filter implemented as a Gaussian blur,… view at source ↗
Figures from the paper (16 more)
Figure 4
Figure 4. Figure 4: Zoomed in details for the 8k image. FreeScale may regenerate the original blurred areas at low resolution based on the prior knowledge that the model has learned. As shown in the bottom row, two originally chaotic and blurry faces are clearly outlined at 8k resolution.…
Figure 5
Figure 5. Figure 5: Flexible aspect ratio generation. FreeScale can directly achieve a flexible aspect ratio (the resolution must be a multiple of 512) without any adaptation. replaces β in the original RoPE:  cos  n β 0  ,sin  n β 0  , · · · , cos  n β d/2−1  ,sin  n β d/2−1  (…
Figure 6
Figure 6. Figure 6: Image qualitative comparisons with super-resolution. FreeScale is not inferior to SDXL+Real-ESRGAN in visual quality, and adds more details. In addition, SR methods will faithfully follow the low-resolution input while FreeScale can regenerate the original blurred area…
Figure 7
Figure 7. Figure 7: Results of flexible control for detail level. A better result will be generated by adding the coefficient weight in the area of Griffons and reducing the coefficient weight in the other regions. 1× Result No Editing Hair Editing Face Editing [PITH_FULL_IMAGE:figures/f…
Figure 8
Figure 8. Figure 8: Local semantic editing of images. FreeScale makes the hair purple or edits the face to make this person look more Japanese in the higher-resolution (40962 ). the model has learned. In [PITH_FULL_IMAGE:figures/full_fig_p008_8.png]
Figure 9
Figure 9. Figure 9: UNet-based video qualitative comparisons with other baselines. While other baselines fail in video generation, FreeScale effectively generates higher-resolution videos with high fidelity. Best viewed ZOOMED-IN. SDXL (50 steps) SDXL-Turbo (4 steps) SDXL-Turbo (2 steps) …
Figure 10
Figure 10. Figure 10: Fast generation with SDXL-Turbo. FreeScale can help SDXL￾Turbo generate results at 20482 resolution with even 2 timesteps. petitive performance in quantitative metrics. As reported in most previously published related works [8], [10], higher￾resolution generation meth…
Figure 11
Figure 11. Figure 11: Qualitative image comparisons with ablations. Our full method performs the best. The resolution of results is 40962 for better visualizing the difference between the various strategies. In addition, this mechanism can even be extended to local semantic editing. Utiliz…
Figure 12
Figure 12. Figure 12: Video ablations in Wan T2V without tuning in resolution 960 × 1664. Although all variants can generate rough results. Our full method performs the best. Best viewed ZOOMED-IN. TABLE 4 Image quantitative comparisons with other ablations. Our final FreeScale achieves be…
Figure 13
Figure 13. Figure 13: Video comparison with DiT-based models in resolution 960 × 1664. Although other baselines can produce reasonable results at moderately higher resolutions, they still suffer from varying degrees of blurriness. In contrast, CineScale generates high-quality videos with r…
Figure 14
Figure 14. Figure 14: Video comparison with DiT-based models in resolution 1920 × 3328. At the resolution several times higher than those used during training, LTX and Wan-DI tend to fail completely. While UAV, a video super-resolution approach, can still produce visually reasonable result…
Figure 15
Figure 15. Figure 15: Video ablations in Wan I2V without tuning in resolution 960 × 1664. Although all variants can generate rough results. Our full method performs the best. Best viewed ZOOMED-IN. TABLE 11 Video comparison with DiT-based models in resolution 1920 × 3328. Only three metric…
Figure 16
Figure 16. Figure 16: Video ablations in ReCamMaster V2V without tuning in resolution 960 × 1664. Without NTK-RoPE, repeated patterns are prone to occur due to errors in positional encoding. Although all variants can generate rough results. Our full method performs the best. TABLE 12 Video…
Figure 17
Figure 17. Figure 17: Local semantic editing for video generation. CineScale supports efficient editing by allowing users to preview results at low resolution while modifying high-resolution local semantics via prompts. 6.2 Image/Video-to-Video Generation CineScale also supports image-to-v…
Figure 18
Figure 18. Figure 18: 4k text-to-video generation. With minimal LoRA fine-tuning, CineScale can achieve 4k (2176 × 3840) text-to-video generation. We observe that at 4k resolution, faces can be generated with remarkable clarity even when they occupy only a small portion of the frame, and t…
Figure 19
Figure 19. Figure 19: 4k image-to-video generation. With minimal LoRA fine-tuning, CineScale can achieve 4k (2176 × 3840) image-to-video generation. eral examples provided in the original paper, and [PITH_FULL_IMAGE:figures/full_fig_p014_19.png]

Discussion (0). Sign in to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Phase-Aligned RoPE for Mixed-Resolution Diffusion Transformer

    cs.CV 2025-11 conditional novelty 6.0 of 10

    Expressing all RoPE positions on the query's grid ('one attention, one scale') plus a small boundary content-exchange step restores mixed-resolution diffusion generation that naive position interpolation destroys.

Reference graph

Works this paper leans on

76 extracted references · 37 canonical work pages · cited by 1 Pith paper

  1. [1]

    Sdxl: Improving latent dif- fusion models for high-resolution image synthesis,

    D. Podell, Z. English, K. Lacey, A. Blattmann, T. Dockhorn, J. M ¨uller, J. Penna, and R. Rombach, “Sdxl: Improving latent dif- fusion models for high-resolution image synthesis,” arXiv preprint arXiv:2307.01952, 2023

    arXiv 2023

  2. [2]

    Pixart- α: Fast training of diffusion transformer for photorealistic text-to-image synthesis,

    J. Chen, J. Yu, C. Ge, L. Yao, E. Xie, Y. Wu, Z. Wang, J. Kwok, P . Luo, H. Lu, and Z. Li, “Pixart- α: Fast training of diffusion transformer for photorealistic text-to-image synthesis,” 2023

    work page 2023

  3. [3]

    Modelscope text-to-video technical report,

    J. Wang, H. Yuan, D. Chen, Y. Zhang, X. Wang, and S. Zhang, “Modelscope text-to-video technical report,” 2023

    work page 2023

  4. [4]

    Videocrafter2: Overcoming data limitations for high- quality video diffusion models,

    H. Chen, Y. Zhang, X. Cun, M. Xia, X. Wang, C. Weng, and Y. Shan, “Videocrafter2: Overcoming data limitations for high- quality video diffusion models,” 2024

    work page 2024

  5. [5]

    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

    arXiv 2024

  6. [6]

    Adding conditional control to text-to-image diffusion models,

    L. Zhang, A. Rao, and M. Agrawala, “Adding conditional control to text-to-image diffusion models,” in Proceedings of the IEEE/CVF international conference on computer vision, 2023, pp. 3836–3847

    work page 2023

  7. [7]

    High-resolution image synthesis with latent diffusion models,

    R. Rombach, A. Blattmann, D. Lorenz, P . Esser, and B. Ommer, “High-resolution image synthesis with latent diffusion models,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2022, pp. 10 684–10 695

    work page 2022

  8. [8]

    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,” in The Twelfth International Conference on Learning Representations, 2024

    work page 2024

Show all 76 references
  1. [9]

    Multidiffusion: Fusing diffusion paths for controlled image generation,

    O. Bar-Tal, L. Yariv, Y. Lipman, and T. Dekel, “Multidiffusion: Fusing diffusion paths for controlled image generation,” arXiv preprint arXiv:2302.08113, 2023

    2023 arXiv

  2. [10]

    Demo- fusion: Democratising high-resolution image generation with no $$$,

    R. Du, D. Chang, T. Hospedales, Y.-Z. Song, and Z. Ma, “Demo- fusion: Democratising high-resolution image generation with no $$$,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024, pp. 6159–6168

    2024

  3. [11]

    Fouriscale: A frequency perspective on training-free high- resolution image synthesis,

    L. Huang, R. Fang, A. Zhang, G. Song, S. Liu, Y. Liu, and H. Li, “Fouriscale: A frequency perspective on training-free high- resolution image synthesis,” arXiv preprint arXiv:2403.12963, 2024

    2024 arXiv

  4. [12]

    Make a cheap scaling: A self- cascade diffusion model for higher-resolution adaptation,

    L. Guo, Y. He, H. Chen, M. Xia, X. Cun, Y. Wang, S. Huang, Y. Zhang, X. Wang, Q. Chen et al., “Make a cheap scaling: A self- cascade diffusion model for higher-resolution adaptation,” arXiv preprint arXiv:2402.10491, 2024

    2024 arXiv

  5. [13]

    Extending context window of large language models via positional interpolation,

    S. Chen, S. Wong, L. Chen, and Y. Tian, “Extending context window of large language models via positional interpolation,” arXiv preprint arXiv:2306.15595, 2023

    2023 arXiv

  6. [14]

    YaRN: Efficient context window extension of large language models,

    B. Peng, J. Quesnelle, H. Fan, and E. Shippole, “YaRN: Efficient context window extension of large language models,” in The Twelfth International Conference on Learning Representations ,

  7. [15]

    Lora: Low-rank adaptation of large language models

    E. J. Hu, Y. Shen, P . Wallis, Z. Allen-Zhu, Y. Li, S. Wang, L. Wang, W. Chen et al. , “Lora: Low-rank adaptation of large language models.” ICLR, vol. 1, no. 2, p. 3, 2022

    2022

  8. [16]

    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,” arXiv preprint arXiv:2412.09626, 2024

    2024 arXiv

  9. [17]

    Instructvideo: Instructing video diffusion models with human feedback,

    H. Yuan, S. Zhang, X. Wang, Y. Wei, T. Feng, Y. Pan, Y. Zhang, Z. Liu, S. Albanie, and D. Ni, “Instructvideo: Instructing video diffusion models with human feedback,” in CVPR, 2024

    2024

  10. [18]

    Freeu: Free lunch in diffusion u-net,

    C. Si, Z. Huang, Y. Jiang, and Z. Liu, “Freeu: Free lunch in diffusion u-net,” in CVPR, 2024

    2024

  11. [19]

    Dreamvideo: Composing your dream videos with customized subject and motion,

    Y. Wei, S. Zhang, Z. Qing, H. Yuan, Z. Liu, Y. Liu, Y. Zhang, J. Zhou, and H. Shan, “Dreamvideo: Composing your dream videos with customized subject and motion,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2024, pp. 6537–6549

    2024

  12. [20]

    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. Pan et al., “Vchitect-2.0: Parallel transformer for scaling up video diffusion models,” arXiv preprint arXiv:2501.08453, 2025

    2025 arXiv

  13. [21]

    Lavie: High-quality video gen- eration with cascaded latent diffusion models,

    Y. Wang, X. Chen, X. Ma, S. Zhou, Z. Huang, Y. Wang, C. Yang, Y. He, J. Yu, P . Yang et al. , “Lavie: High-quality video gen- eration with cascaded latent diffusion models,” arXiv preprint arXiv:2309.15103, 2023

    2023 arXiv

  14. [22]

    Repvideo: Rethinking cross-layer representation for video generation,

    C. Si, W. Fan, Z. Lv, Z. Huang, Y. Qiao, and Z. Liu, “Repvideo: Rethinking cross-layer representation for video generation,” arXiv 2501.08994, 2025

    2025 arXiv

  15. [23]

    Denoising diffusion probabilistic models,

    J. Ho, A. Jain, and P . Abbeel, “Denoising diffusion probabilistic models,” Advances in Neural Information Processing Systems, vol. 33, pp. 6840–6851, 2020

    2020

  16. [24]

    Diffusion models beat gans on image synthesis,

    P . Dhariwal and A. Nichol, “Diffusion models beat gans on image synthesis,” Advances in neural information processing systems, vol. 34, pp. 8780–8794, 2021

    2021

  17. [25]

    Scalable diffusion models with trans- formers,

    W. Peebles and S. Xie, “Scalable diffusion models with trans- formers,” in Proceedings of the IEEE/CVF international conference on computer vision, 2023, pp. 4195–4205

    2023

  18. [26]

    Video diffusion models,

    J. Ho, T. Salimans, A. Gritsenko, W. Chan, M. Norouzi, and D. J. Fleet, “Video diffusion models,” Advances in Neural Information Processing Systems, vol. 35, pp. 8633–8646, 2022

    2022

  19. [27]

    Latent video diffusion models for high-fidelity video generation with arbitrary lengths,

    Y. He, T. Yang, Y. Zhang, Y. Shan, and Q. Chen, “Latent video diffusion models for high-fidelity video generation with arbitrary lengths,” arXiv preprint arXiv:2211.13221, 2022

    2022 arXiv

  20. [28]

    Align your latents: High-resolution video synthesis with latent diffusion models,

    A. Blattmann, R. Rombach, H. Ling, T. Dockhorn, S. W. Kim, S. Fidler, and K. Kreis, “Align your latents: High-resolution video synthesis with latent diffusion models,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2023, pp. 22 563–22 575

    2023

  21. [29]

    Animatediff: Animate your personalized text-to-image diffusion models without specific tuning,

    Y. Guo, C. Yang, A. Rao, Y. Wang, Y. Qiao, D. Lin, and B. Dai, “Animatediff: Animate your personalized text-to-image diffusion models without specific tuning,” arXiv preprint arXiv:2307.04725 , 2023

    2023 arXiv

  22. [30]

    Videocomposer: Compositional video synthesis with motion controllability,

    X. Wang, H. Yuan, S. Zhang, D. Chen, J. Wang, Y. Zhang, Y. Shen, D. Zhao, and J. Zhou, “Videocomposer: Compositional video synthesis with motion controllability,” NeurIPS, 2023

    2023

  23. [31]

    Videocrafter1: Open diffusion models for high-quality video generation,

    H. Chen, M. Xia, Y. He, Y. Zhang, X. Cun, S. Yang, J. Xing, Y. Liu, Q. Chen, X. Wang et al. , “Videocrafter1: Open diffusion models for high-quality video generation,”arXiv preprint arXiv:2310.19512, 2023

    2023 arXiv

  24. [32]

    Stable video diffusion: Scaling latent video diffusion models to large datasets,

    A. Blattmann, T. Dockhorn, S. Kulal, D. Mendelevitch, M. Kilian, D. Lorenz, Y. Levi, Z. English, V . Voleti, A. Lettset al., “Stable video diffusion: Scaling latent video diffusion models to large datasets,” arXiv preprint arXiv:2311.15127, 2023

    2023 arXiv

  25. [33]

    Lumiere: A space-time diffusion model for video generation,

    O. Bar-Tal, H. Chefer, O. Tov, C. Herrmann, R. Paiss, S. Zada, A. Ephrat, J. Hur, Y. Li, T. Michaeli et al. , “Lumiere: A space-time diffusion model for video generation,” arXiv preprint arXiv:2401.12945, 2024

    2024 arXiv

  26. [34]

    Pyramidal flow matching for efficient video generative modeling,

    Y. Jin, Z. Sun, N. Li, K. Xu, K. Xu, H. Jiang, N. Zhuang, Q. Huang, Y. Song, Y. Mu, and Z. Lin, “Pyramidal flow matching for efficient video generative modeling,” 2024

    2024

  27. [35]

    Mochi 1,

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

    2024

  28. [36]

    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. Gordon, P . Panet, S. Weissbuch, V . Kulikov, Y. Bitterman, Z. Melumian, and JOURNAL OF LATEX CLASS FILES, VOL. 14, NO. 8, AUGUST 2021 16 O. Bibi, “Ltx-video: Realtime ...

    2021 arXiv

  29. [37]

    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. Zhang et al. , “Hunyuanvideo: A systematic framework for large video generative models,” arXiv preprint arXiv:2412.03603, 2024

    2024 arXiv

  30. [38]

    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. Wen...

    2025 arXiv

  31. [39]

    Relay diffusion: Unifying diffusion process across reso- lutions for image synthesis,

    J. Teng, W. Zheng, M. Ding, W. Hong, J. Wangni, Z. Yang, and J. Tang, “Relay diffusion: Unifying diffusion process across reso- lutions for image synthesis,” arXiv preprint arXiv:2309.03350, 2023

    2023 arXiv

  32. [40]

    simple diffusion: End- to-end diffusion for high resolution images,

    E. Hoogeboom, J. Heek, and T. Salimans, “simple diffusion: End- to-end diffusion for high resolution images,” in International Con- ference on Machine Learning. PMLR, 2023, pp. 13 213–13 232

    2023

  33. [41]

    Ultrapixel: Advancing ultra-high-resolution image synthesis to new peaks,

    J. Ren, W. Li, H. Chen, R. Pei, B. Shao, Y. Guo, L. Peng, F. Song, and L. Zhu, “Ultrapixel: Advancing ultra-high-resolution image synthesis to new peaks,” arXiv preprint arXiv:2407.02158, 2024

    2024 arXiv

  34. [42]

    Linfusion: 1 gpu, 1 minute, 16k image,

    S. Liu, W. Yu, Z. Tan, and X. Wang, “Linfusion: 1 gpu, 1 minute, 16k image,” 2024

    2024

  35. [43]

    Any- size-diffusion: Toward efficient text-driven synthesis for any-size hd images,

    Q. Zheng, Y. Guo, J. Deng, J. Han, Y. Li, S. Xu, and H. Xu, “Any- size-diffusion: Toward efficient text-driven synthesis for any-size hd images,” in Proceedings of the AAAI Conference on Artificial Intelligence, vol. 38, no. 7, 2024, pp. 7571–7578

    2024

  36. [44]

    Resadapter: Domain consistent resolution adapter for diffusion models,

    J. Cheng, P . Xie, X. Xia, J. Li, J. Wu, Y. Ren, H. Li, X. Xiao, M. Zheng, and L. Fu, “Resadapter: Domain consistent resolution adapter for diffusion models,” 2024

    2024

  37. [45]

    Elasticdiffusion: Training-free arbitrary size image generation through global-local content separation,

    M. Haji-Ali, G. Balakrishnan, and V . Ordonez, “Elasticdiffusion: Training-free arbitrary size image generation through global-local content separation,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024, pp. 6603–6612

    2024

  38. [46]

    Cutdiffusion: A simple, fast, cheap, and strong diffusion extrapolation method,

    M. Lin, Z. Lin, W. Zhan, L. Cao, and R. Ji, “Cutdiffusion: A simple, fast, cheap, and strong diffusion extrapolation method,” arXiv preprint arXiv:2404.15141, 2024

    2024 arXiv

  39. [47]

    Syncdiffusion: Coherent montage via synchronized joint diffusions,

    Y. Lee, K. Kim, H. Kim, and M. Sung, “Syncdiffusion: Coherent montage via synchronized joint diffusions,” Advances in Neural Information Processing Systems, vol. 36, pp. 50 648–50 660, 2023

    2023

  40. [48]

    Training-free diffusion model adaptation for variable-sized text-to-image synthesis,

    Z. Jin, X. Shen, B. Li, and X. Xue, “Training-free diffusion model adaptation for variable-sized text-to-image synthesis,” Advances in Neural Information Processing Systems , vol. 36, pp. 70 847–70 860, 2023

    2023

  41. [49]

    Upsample guidance: Scale up dif- fusion models without training,

    J. Hwang, Y.-H. Park, and J. Jo, “Upsample guidance: Scale up dif- fusion models without training,” arXiv preprint arXiv:2404.01709 , 2024

    2024 arXiv

  42. [50]

    Ap-ldm: Attentive and progres- sive latent diffusion model for training-free high-resolution image generation,

    B. Cao, J. Ye, Y. Wei, and H. Shan, “Ap-ldm: Attentive and progres- sive latent diffusion model for training-free high-resolution image generation,” arXiv preprint arXiv:2410.06055, 2024

    2024

  43. [51]

    Hid- iffusion: Unlocking higher-resolution creativity and efficiency in pretrained diffusion models,

    S. Zhang, Z. Chen, Z. Zhao, Y. Chen, Y. Tang, and J. Liang, “Hid- iffusion: Unlocking higher-resolution creativity and efficiency in pretrained diffusion models,” in European Conference on Computer Vision. Springer, 2024, pp. 145–161

    2024

  44. [52]

    Diffusehigh: Training- free progressive high-resolution image synthesis through structure guidance,

    Y. Kim, G. Hwang, J. Zhang, and E. Park, “Diffusehigh: Training- free progressive high-resolution image synthesis through structure guidance,” arXiv preprint arXiv:2406.18459, 2024

    2024 arXiv

  45. [53]

    Accdiffusion: An accurate method for higher-resolution image generation,

    Z. Lin, M. Lin, M. Zhao, and R. Ji, “Accdiffusion: An accurate method for higher-resolution image generation,” arXiv preprint arXiv:2407.10738, 2024

    2024 arXiv

  46. [54]

    Hiprompt: Tuning-free higher-resolution generation with hierarchical mllm prompts,

    X. Liu, Y. He, L. Guo, X. Li, B. Jin, P . Li, Y. Li, C.-M. Chan, Q. Chen, W. Xue et al., “Hiprompt: Tuning-free higher-resolution generation with hierarchical mllm prompts,” arXiv preprint arXiv:2409.02919 , 2024

    2024 arXiv

  47. [55]

    Beyondscene: Higher-resolution human-centric scene generation with pretrained diffusion,

    G. Kim, H. Kim, H. Seo, D. U. Kang, and S. Y. Chun, “Beyondscene: Higher-resolution human-centric scene generation with pretrained diffusion,” in European Conference on Computer Vision . Springer, 2024, pp. 126–142

    2024

  48. [56]

    Real-esrgan: Training real-world blind super-resolution with pure synthetic data,

    X. Wang, L. Xie, C. Dong, and Y. Shan, “Real-esrgan: Training real-world blind super-resolution with pure synthetic data,” in Proceedings of the IEEE/CVF international conference on computer vision, 2021, pp. 1905–1914

    2021

  49. [57]

    Upscale-a-video: Temporal-consistent diffusion model for real-world video super- resolution,

    S. Zhou, P . Yang, J. Wang, Y. Luo, and C. C. Loy, “Upscale-a-video: Temporal-consistent diffusion model for real-world video super- resolution,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024, pp. 2535–2545

    2024

  50. [58]

    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. Shaoet al., “Turbo2k: Towards ultra-efficient and high- quality 2k video synthesis,” arXiv preprint arXiv:2504.14470, 2025

    2025 arXiv

  51. [59]

    Freenoise: Tuning-free longer video diffusion via noise reschedul- ing,

    H. Qiu, M. Xia, Y. Zhang, Y. He, X. Wang, Y. Shan, and Z. Liu, “Freenoise: Tuning-free longer video diffusion via noise reschedul- ing,” arXiv preprint arXiv:2310.15169, 2023

    2023 arXiv

  52. [60]

    Flux.1 : An advanced state-of-the-art generative deep learning model,

    B. F. Labs, “Flux.1 : An advanced state-of-the-art generative deep learning model,” Black Forest Labs, Tech. Rep., 2024. [Online]. Available: https://flux1.io/

    2024

  53. [61]

    I-max: Maximize the resolution potential of pre-trained rectified flow transformers with projected flow,

    R. Du, D. Liu, L. Zhuo, Q. Qi, H. Li, Z. Ma, and P . Gao, “I-max: Maximize the resolution potential of pre-trained rectified flow transformers with projected flow,” arXiv preprint arXiv:2410.07536, 2024

    2024 arXiv

  54. [62]

    Yarn: Efficient con- text window extension of large language models,

    B. Peng, J. Quesnelle, H. Fan, and E. Shippole, “Yarn: Efficient con- text window extension of large language models,” arXiv preprint arXiv:2309.00071, 2023

    2023 arXiv

  55. [63]

    Roformer: Enhanced transformer with rotary position embedding,

    J. Su, M. Ahmed, Y. Lu, S. Pan, W. Bo, and Y. Liu, “Roformer: Enhanced transformer with rotary position embedding,” Neuro- computing, vol. 568, p. 127063, 2024

    2024

  56. [64]

    Scaling rectified flow transformers for high-resolution image synthesis,

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

    2024

  57. [65]

    Gans trained by a two time-scale update rule converge to a local nash equilibrium,

    M. Heusel, H. Ramsauer, T. Unterthiner, B. Nessler, and S. Hochre- iter, “Gans trained by a two time-scale update rule converge to a local nash equilibrium,” Advances in neural information processing systems, vol. 30, 2017

    2017

  58. [66]

    De- mystifying mmd gans,

    M. Bi ´nkowski, D. J. Sutherland, M. Arbel, and A. Gretton, “De- mystifying mmd gans,” arXiv preprint arXiv:1801.01401, 2018

    2018 arXiv

  59. [67]

    Improved techniques for training gans,

    T. Salimans, I. Goodfellow, W. Zaremba, V . Cheung, A. Radford, and X. Chen, “Improved techniques for training gans,” Advances in neural information processing systems, vol. 29, 2016

    2016

  60. [68]

    Any- resolution training for high-resolution image synthesis,

    L. Chai, M. Gharbi, E. Shechtman, P . Isola, and R. Zhang, “Any- resolution training for high-resolution image synthesis,” in Euro- pean Conference on Computer Vision. Springer, 2022, pp. 170–188

    2022

  61. [69]

    Towards accurate generative mod- els of video: A new metric & challenges,

    T. Unterthiner, S. van Steenkiste, K. Kurach, R. Marinier, M. Michalski, and S. Gelly, “Towards accurate generative mod- els of video: A new metric & challenges,” arXiv preprint arXiv:1812.01717, 2018

    2018 arXiv

  62. [70]

    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. Chanpaisit et al., “Vbench: Comprehensive benchmark suite for video generative models,”arXiv preprint arXiv:2311.17982, 2023

    2023 arXiv

  63. [71]

    Segment anything,

    A. Kirillov, E. Mintun, N. Ravi, H. Mao, C. Rolland, L. Gustafson, T. Xiao, S. Whitehead, A. C. Berg, W.-Y. Lo et al. , “Segment anything,” in Proceedings of the IEEE/CVF International Conference on Computer Vision, 2023, pp. 4015–4026

    2023

  64. [72]

    Adversarial diffusion distillation,

    A. Sauer, D. Lorenz, A. Blattmann, and R. Rombach, “Adversarial diffusion distillation,” in European Conference on Computer Vision . Springer, 2024, pp. 87–103

    2024

  65. [73]

    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,” 2025

    2025

  66. [74]

    Vbench++: Comprehensive and versatile benchmark suite for video generative models,

    Z. Huang, F. Zhang, X. Xu, Y. He, J. Yu, Z. Dong, Q. Ma, N. Chanpaisit, C. Si, Y. Jiang, Y. Wang, X. Chen, Y.-C. Chen, L. Wang, D. Lin, Y. Qiao, and Z. Liu, “Vbench++: Comprehensive and versatile benchmark suite for video generative models,” arXiv preprint arXiv:2411.13503, 2024

    2024 arXiv

  67. [75]

    Recammaster: Camera-controlled generative rendering from a single video,

    J. Bai, M. Xia, X. Fu, X. Wang, L. Mu, J. Cao, Z. Liu, H. Hu, X. Bai, P . Wanet al., “Recammaster: Camera-controlled generative rendering from a single video,” arXiv preprint arXiv:2503.11647 , 2025. JOURNAL OF LATEX CLASS FILES, VOL. 14, NO. 8, AUGUST 2021 17 8 B IOGRAPHY SEC...

    2025 arXiv

  68. [2024]

    Available: https://openreview.net/forum?id= wHBfxhZu1u

    [Online]. Available: https://openreview.net/forum?id= wHBfxhZu1u

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.