Pith. sign in

REVIEW 5 major objections 5 minor 59 references

StyleRWKV: High-Quality and High-Efficiency Style Transfer with RWKV-like Architecture

T0 review · 5 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read The paper claims that an RWKV-style backbone, StyleRWKV, delivers state-of-the-art arbitrary style transfer while keeping inference linear in sequence length.

desk verdict A plausible first RWKV-based style transfer model with well-ablated components, but the evaluation protocol needs fixing before the 'outperforms SOTA' claim is credible. read the letter →

arxiv 2412.19535 v1 pith:UZ6XBZGY submitted 2024-12-27 cs.CV

Miaomiao Dai , Qianyu Zhou , Lizhuang Ma This is my paper · ORCID
classification cs.CV
keywords styletransferRWKVlinearcomplexityattentionmechanismdeformableconvolutionefficientinferenceencoder-decoderarchitecture
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

StyleRWKV aims to show that an RWKV-style sequence model—previously used for long-context language modeling—can match or beat specialized Transformer-, diffusion-, and state-space-based style transfer models on both quality and speed. The paper introduces three components: Recurrent WKV (Re-WKV) attention for global context with linear complexity, Deformable Shifting (Deform-Shifting) for adaptive local token fusion, and Skip Scanning (S-Scanning) for long-range dependencies. On a 5,000-pair COCO-WikiArt test set it reports the best ArtFID, FID, and LPIPS among all compared methods, while keeping inference time at 1.53 seconds at 512×512 resolution and scaling linearly with resolution. If correct, it establishes RWKV-like architectures as a practical backbone for efficient high-quality arbitrary style transfer, especially at high resolutions.

What carries the argument

The load-bearing mechanism is the Recurrent WKV (Re-WKV) attention: it takes the bidirectional WKV attention from Vision-RWKV and applies it recurrently q times along the S-Scanning paths, with a channel-wise spatial decay vector w and a current-token bonus u, to compute a global attention output in linear time. This is paired with two complementary modules: Deform-Shifting, which learns convolution offsets to move tokens from regions of interest and enhance local dependencies, and S-Scanning, which slices the feature map with a step size p to create sparse scanning directions that connect distant patches. Together they give the ST-RWKV block both global and local receptive fields while avoiding quadratic cost.

What would settle it

A re-run that selects q and p on a separate validation split and then evaluates on the untouched 5,000-pair test set—plus an independent timing of StyTR2 at 512×512 on the same hardware—would determine whether the reported quality and efficiency margins hold.

Watch

Extended reading notes

Core claim

The paper claims that StyleRWKV, a 4-level U-shaped encoder-decoder built from ST-RWKV blocks, outperforms state-of-the-art arbitrary style transfer methods in stylization quality, model complexity, and inference efficiency. The core of the model is the Re-WKV attention, which runs Bi-WKV attention recurrently q=2 times along S-Scanning directions, giving a global receptive field at linear cost; local details are handled by Deform-Shifting, which uses deformable convolution to shift tokens from regions of interest; and S-Scanning samples tokens with a stride p=2 to build long-range context. In experiments the model achieves ArtFID 26.370, FID 16.362, LPIPS 0.451, and inference times of 0.266s at 128 and 3.670s at 1024 sequence lengths, better than all compared Transformer-, diffusion-, and Mamba-based baselines.

Load-bearing premise

The central claim depends on the reported comparisons being fair—that the 5,000-pair test set was not used to select hyperparameters (q=2, p=2) and that the baseline inference times were measured under comparable conditions.

Editorial extensions

If this is right

  • Transformer-based style transfer methods incur quadratic cost as resolution grows, so a linear-complexity alternative like StyleRWKV can scale to high-resolution images without that penalty.
  • The reported inference times—0.266s at 128, 1.530s at 512, and 3.670s at 1024 sequence lengths—make the method fast enough for practical use on standard hardware.
  • With 28.8M parameters, the model is far smaller than diffusion-based style transfer models, which reduces memory usage and training overhead.
  • The user study reports that 70% of responses favored StyleRWKV over ten competing methods, suggesting the quality gains are perceptually visible, not just metric-based.

Reading between the lines

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

  • Because the hyperparameters q=2 and p=2 were selected using test-set metrics, re-running the ablation with a separate validation split would verify whether the claimed margins hold on an untouched test set.
  • The same recurrent bidirectional WKV + skip-scanning design could transfer to other high-resolution dense prediction tasks, such as super-resolution or semantic segmentation, where global context and local detail both matter.
  • Deform-Shifting's learned offsets could be visualized to reveal whether the model implicitly discovers object boundaries, which would explain its strong preservation of local details like the duck's beak in the qualitative results.
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 StyleRWKV, a hierarchical encoder-decoder architecture for arbitrary neural style transfer that replaces quadratic-complexity attention with RWKV-like linear-complexity components. Three components are introduced: a Recurrent WKV (Re-WKV) attention with bidirectional exponential decay applied recurrently, a Deformable Shifting (Deform-Shifting) layer using deformable convolution for local token interaction, and a Skip Scanning (S-Scanning) mechanism that groups tokens with a step size to build global dependencies. The model is trained with VGG-based content, style, and identity losses, and evaluated on 5,000 COCO-WikiArt pairs. The paper reports the best ArtFID, FID, and LPIPS among the compared methods, lower parameter counts than transformer and diffusion baselines, and faster inference than Mamba-based and transformer-based competitors at long sequence lengths, with ablations over recurrence count, shifting mechanism, and scanning method.

Significance. If the empirical claims hold, StyleRWKV is a meaningful step toward efficient arbitrary style transfer: it combines the linear-scaling property of RWKV with a hierarchical, multi-scale design, and the ablation study suggests that each proposed module contributes positively. The architecture itself is plausible and the complexity analysis is credible. The paper also responds to a real gap, since most existing style transfer models are either quadratic in complexity or require iterative diffusion sampling. However, the comparative claims in the abstract and Table 1 rest on an evaluation protocol that is currently not fully verifiable: hyperparameters appear to be selected on the same test set used for final numbers, baseline timings lack a measurement protocol, and the user-preference row is ambiguous. The contribution is of interest to the style-transfer and efficient-sequence-modeling communities, but the evidence needs to be tightened before the central claim can be accepted.

major comments (5)
  1. [Section 4.3, Tables 2 and 4] The hyperparameters q=2 and p=2 are selected using the same metrics (ArtFID, FID, LPIPS) that are later reported as the final results in Table 1, but Section 4.1 describes only a single 5,000-pair COCO-WikiArt test set with no separate validation split. This creates a risk of test-set tuning that would bias the reported ArtFID 26.370, FID 16.362, and LPIPS 0.451 optimistically. Notably, Table 2 shows q=3 improves all three quality metrics over q=2, so the choice of q=2 is explicitly a trade-off between quality and efficiency, not a quality-maximizing selection. The paper should either introduce a held-out validation set for hyperparameter selection or clearly state that the ablation tables are exploratory comparisons and that the chosen configuration is a speed-quality compromise.
  2. [Table 1, inference time rows] No measurement protocol is reported for the inference times: no GPU model, batch size, framework version, warm-up procedure, or number of repeated runs. Some numbers are implausible or at least unexplained: AesPA-Net goes from 0.668 s at 128 to 0.695 s at 1024 (an almost flat curve for a 4x increase in sequence length), while StyTR2 grows to 43,319 s at 1024 (over 12 hours) despite having only 35.39M parameters. Without a standardized protocol, the efficiency claim in the abstract and the comparison in Table 1 are not independently verifiable. The authors should specify the hardware and timing methodology and provide error bars or at least explain the anomalously flat AesPA-Net scaling.
  3. [Table 1, user preference row] The preference row is ambiguous: the 'Ours' column contains '0.7*' while every other method has a pair such as '0.66/0.34'. The text states that most participants favored the stylized results, but the colon-separated pairs are not defined. If the pair is 'ours vs. competitor', the table should say so explicitly; if the values are the competitor's win rate, the claim is contradicted. In addition, no statistical significance test or inter-subject agreement is reported for the 2,250 responses, so the reader cannot tell whether the preference differences are meaningful.
  4. [Section 3.3, Eq. (5)] The definition of Re-WKV is internally inconsistent. The recursion states wkv^{(j)} = Bi-WKV^{(j)}(K, wkv^{(j-1)}), which uses the previous output as the value input, but the closed-form Bi-WKV(K, V) immediately below uses V as the value input. The paper should clarify whether the first iteration uses V = V_{ms}^{n} and subsequent iterations use V = wkv^{(j-1)}, or whether Eq. (5) contains a typo. This is load-bearing for reproducibility because the recurrent mechanism is the paper's central architectural novelty. Additionally, 'According to Eq. 4' in the same section appears to reference the wrong equation; Eq. (5) or the Bi-WKV definition is meant.
  5. [Section 3.4, Eq. (7)] The S-Scanning operation is not defined precisely enough to be reproduced. The slicing notation K[..., a::p, b::p] with (a,b) given by Eq. (8) is not a standard slicing expression, and the arrow-based notation for the split and merge steps is garbled in the text. The authors should state exactly how the four groups O_i are formed, how the scanning order is fed into Re-WKV, and how the groups are merged back into the full feature map. Without this, the S-Scanning mechanism cannot be implemented from the paper.
minor comments (5)
  1. [Table 1 note] The note '128×128, 256×256, etc. refer to sequence lengths rather than resolution' is confusing because the headers read 128×128, 256×256. If the sequence length is 128 (rather than 128×128 patches), the headers should be '128', '256', etc., or the text should specify the patch size used to derive the sequence length.
  2. [Section 3.3] The sentence 'According to Eq. 4, Bi-WKV is partially determined by the relative position bias' references Eq. (4), which is the channel-mix formula; the relevant formula is Eq. (5) or the definition of Bi-WKV.
  3. [General] There are typographical issues throughout the paper, e.g., 'DMs-based methods ... are hindered by their time- consuming generation process' in the introduction, and reference [27] contains an apparent page-number artifact ('arXiv:2303.15403, 3, 2023'). A thorough proofread is needed.
  4. [Tables 2-4] The 'Time/s' columns in the ablation tables do not state the input sequence length at which the timings were measured, making it hard to compare them with the timing rows of Table 1.
  5. [Section 4.1] The paper does not report random seeds, the number of runs per evaluation, or any error bars for the main metrics. At minimum, the authors should state that metrics are single-run and report seeds for reproducibility.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity; the architecture's complexity claims follow from its stated equations and the empirical comparison is against external baselines.

full rationale

StyleRWKV is an empirical architecture paper rather than a derivation from first principles. The claimed linear-time complexity follows from the stated operations: Re-WKV recurses a fixed q=2 times (Eq. 6) with q<<T, S-Scanning subsamples by step p (Eq. 7), and Deform-Shifting is a fixed-size deformable convolution (Eq. 9), so the O(T) claim is supported by the equations and is not equivalent to an input. The quality/efficiency comparisons are against external baselines (Table 1) using standard metrics (ArtFID, FID, LPIPS). The overlap between VGG-based training losses (Eqs. 10-11) and LPIPS is common field practice and is not a self-referential reduction; no metric value is defined in terms of the model's own output. The selection of q=2 and p=2 from ablations (Section 4.3) raises an evaluation-protocol question because the paper does not state a separate validation split, but the text does not explicitly say the final test set was used for hyperparameter selection, so this is a reproducibility/correctness concern rather than a demonstrated circularity. No load-bearing self-citation chain is present; citations to prior RWKV variants are external sources for components. Overall, the central claim is not forced by construction.

Assumptions & free parameters 5 free parameters · 5 assumptions · 0 invented entities

The method relies on several hand-chosen hyperparameters and on borrowed architectures. No new physical entities are introduced. The central efficiency claim depends on the prior RWKV complexity result, and the global receptive field claim depends on an unverified property of the recurrent scanning scheme.

free parameters (5)
  • Loss weights λc, λs, λid1, λid2 = 8, 15, 100, 1
    Set by hand (Section 4.1) to balance loss magnitudes; no validation-based search reported. They directly shape the stylization quality.
  • Recurrence count q = 2
    Chosen in Section 4.3 from Table 2 as a trade-off; q=3 gives better metrics (ArtFID 25.639) but slower inference. Selection appears to use the evaluation metrics.
  • Skip step p = 2
    Chosen in Section 4.3 from Table 4; p=2 gives best ArtFID among p=1,2,3. Selection appears to use the evaluation metrics.
  • ST-RWKV block counts N1-N4 = 4, 6, 6, 8
    Architecture widths/heights chosen without ablation in Section 4.1.
  • Channel width C = 48
    Input projection channel count chosen in Section 4.1.
assumptions (5)
  • standard math RWKV WKV attention has linear computational complexity in sequence length
    Assumed from cited prior works [15,37]; underpins the linear complexity claim for StyleRWKV.
  • domain assumption Applying Bi-WKV recurrently q times along S-Scanning paths yields a global receptive field
    Stated in Section 3.3 without proof; the actual pairwise coverage after q=2 and p=2 is not analyzed.
  • domain assumption Deformable convolution's learned offsets improve token shifting in image style transfer
    Borrowed from [8]; the transfer of deformable convolution to a token shifting layer is asserted, not derived.
  • domain assumption Perceptual VGG losses are appropriate for optimizing style transfer and correlate with human judgment
    Standard in NST; used in Section 3.5 for the optimization objective.
  • domain assumption LPIPS, FID, and ArtFID are valid evaluation metrics for style transfer
    Used as the sole quantitative benchmarks in Section 4.1; ArtFID is taken from [48].

how reviews work

0 comments
Cite this review

Pith. "Pith review of StyleRWKV: High-Quality and High-Efficiency Style Transfer with RWKV-like Architecture." pith.science (2026). https://pith.science/paper/UZ6XBZGY

@misc{pith2026241219535,
  author       = {Pith},
  title        = {Pith review of: StyleRWKV: High-Quality and High-Efficiency Style Transfer with RWKV-like Architecture},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/UZ6XBZGY}},
  note         = {Machine review of arXiv:2412.19535}
}
read the original abstract

Style transfer aims to generate a new image preserving the content but with the artistic representation of the style source. Most of the existing methods are based on Transformers or diffusion models, however, they suffer from quadratic computational complexity and high inference time. RWKV, as an emerging deep sequence models, has shown immense potential for long-context sequence modeling in NLP tasks. In this work, we present a novel framework StyleRWKV, to achieve high-quality style transfer with limited memory usage and linear time complexity. Specifically, we propose a Recurrent WKV (Re-WKV) attention mechanism, which incorporates bidirectional attention to establish a global receptive field. Additionally, we develop a Deformable Shifting (Deform-Shifting) layer that introduces learnable offsets to the sampling grid of the convolution kernel, allowing tokens to shift flexibly and adaptively from the region of interest, thereby enhancing the model's ability to capture local dependencies. Finally, we propose a Skip Scanning (S-Scanning) method that effectively establishes global contextual dependencies. Extensive experiments with analysis including qualitative and quantitative evaluations demonstrate that our approach outperforms state-of-the-art methods in terms of stylization quality, model complexity, and inference efficiency.

Figures

Figures reproduced from arXiv: 2412.19535 by the authors.

Figure 1
Figure 1. Comparison of existing neural style transfer methods. [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. (a) Style-RWKV Architecture. (b) Our ST-RWKV block incorporates a Re-WKV attention mechanism to model global [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Illustrations of different token shifting mechanisms. [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Qualitative comparison with state-of-the-art neural style transfer methods, [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: Qualitative ablations on different recurrence number. [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]
Figure 6
Figure 6. Figure 6: Comparisons of Style, Content, and Stylization Result. [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

59 extracted references · 39 canonical work pages

  1. [1]

    Artflow: Unbiased image style transfer via reversible neural flows

    Jie An, Siyu Huang, Yibing Song, Dejing Dou, Wei Liu, and Jiebo Luo. Artflow: Unbiased image style transfer via reversible neural flows. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 862–871, 2021. 1, 2

    work page 2021

  2. [3]

    Mamba-ST: State Space Model for Efficient Style Transfer

    Filippo Botti et al. Mamba-st: State space model for efficient style transfer. arXiv preprint arXiv:2409.10385, 2024. 1

    work page Pith review arXiv 2024

  3. [4]

    Coco- stuff: Thing and stuff classes in context

    Holger Caesar, Jasper Uijlings, and Vittorio Ferrari. Coco- stuff: Thing and stuff classes in context. In CVPR, pages 1209–1218, 2018. 5

    work page 2018

  4. [5]

    Artistic style transfer with internal-external learning and contrastive learning

    Haibo Chen, Zhizhong Wang, Huiming Zhang, Zhiwen Zuo, Ailin Li, Wei Xing, Dongming Lu, et al. Artistic style transfer with internal-external learning and contrastive learning. Ad- vances in Neural Information Processing Systems, 34:26561– 26573, 2021. 1, 2

    work page 2021

  5. [6]

    Artistic style transfer with internal-external learning and contrastive learning

    Haibo Chen, Zhizhong Wang, Huiming Zhang, Zhiwen Zuo, Ailin Li, Wei Xing, Dongming Lu, et al. Artistic style transfer with internal-external learning and contrastive learning. Ad- vances in Neural Information Processing Systems, 34:26561– 26573, 2021

    work page 2021

  6. [7]

    Style injection in diffusion: A training-free approach for adapting large-scale diffusion models for style transfer

    Jiwoo Chung, Sangeek Hyun, and Jae-Pil Heo. Style injection in diffusion: A training-free approach for adapting large-scale diffusion models for style transfer. In CVPR, pages 8795– 8805, 2024. 1, 2, 6

    work page 2024

  7. [8]

    Deformable convolutional networks

    Jifeng Dai, Haozhi Qi, Yuwen Xiong, Yi Li, Guodong Zhang, Han Hu, and Yichen Wei. Deformable convolutional networks. In ICCV, pages 764–773, 2017. 5

    work page 2017

  8. [9]

    Diffusefist: A fast image-guided style transfer method for adapting large-scale diffusion models

    Miaomiao Dai, Qianyu Zhou, Ran Yi, and Lizhuang Ma. Diffusefist: A fast image-guided style transfer method for adapting large-scale diffusion models. In IEEE Interna- tional Conference on Acoustics, Speech and Signal Process- ing (ICASSP), 2025. 1, 2

    work page 2025

Show all 59 references
  1. [10]

    Style transformer: Unpaired text style transfer without disentangled latent representation

    Ning Dai, Jianze Liang, Xipeng Qiu, and Xuanjing Huang. Style transformer: Unpaired text style transfer without disentangled latent representation. arXiv preprint arXiv:1905.05621, 2019. 1

    1905 arXiv

  2. [11]

    Arbitrary style transfer via multi-adaptation network

    Yingying Deng, Fan Tang, Weiming Dong, Wen Sun, Feiyue Huang, and Changsheng Xu. Arbitrary style transfer via multi-adaptation network. In Proceedings of the 28th ACM international conference on multimedia, pages 2719–2727,

  3. [12]

    Arbitrary video style transfer via multi-channel correlation

    Yingying Deng, Fan Tang, Weiming Dong, Haibin Huang, Chongyang Ma, and Changsheng Xu. Arbitrary video style transfer via multi-channel correlation. In Proceedings of the AAAI Conference on Artificial Intelligence, pages 1210–1217, 2021

    2021

  4. [13]

    Stytr2: Image style transfer with transformers

    Yingying Deng, Fan Tang, Weiming Dong, Chongyang Ma, Xingjia Pan, Lei Wang, and Changsheng Xu. Stytr2: Image style transfer with transformers. In CVPR, pages 11326– 11336, 2022. 1, 2, 6

    2022

  5. [14]

    An image is worth 16x16 words: Trans- formers for image recognition at scale

    Alexey Dosovitskiy. An image is worth 16x16 words: Trans- formers for image recognition at scale. arXiv preprint arXiv:2010.11929, 2020. 2

    2010 arXiv

  6. [15]

    Vision-rwkv: Efficient and scalable visual perception with rwkv-like architectures

    Yuchen Duan, Weiyun Wang, Zhe Chen, Xizhou Zhu, Lewei Lu, Tong Lu, Yu Qiao, Hongsheng Li, Jifeng Dai, and Wenhai Wang. Vision-rwkv: Efficient and scalable visual perception with rwkv-like architectures. arXiv preprint arXiv:2403.02308, 2024. 1, 2, 4

    2024 arXiv

  7. [16]

    Diffusion-rwkv: Scaling rwkv- like architectures for diffusion models

    Zhengcong Fei, Mingyuan Fan, Changqian Yu, Debang Li, and Junshi Huang. Diffusion-rwkv: Scaling rwkv- like architectures for diffusion models. arXiv preprint arXiv:2404.04478, 2024. 1, 2

    2024 arXiv

  8. [17]

    Image style transfer using convolutional neural networks

    Leon A Gatys, Alexander S Ecker, and Matthias Bethge. Image style transfer using convolutional neural networks. In CVPR, pages 2414–2423, 2016. 1, 2

    2016

  9. [18]

    Mamba: Linear-time sequence modeling with selective state spaces

    Albert Gu and Tri Dao. Mamba: Linear-time sequence modeling with selective state spaces. arXiv preprint arXiv:2312.00752, 2023. 1

    2023 arXiv

  10. [19]

    Rwkv-clip: A robust vision-language rep- resentation learner

    Tiancheng Gu et al. Rwkv-clip: A robust vision-language rep- resentation learner. arXiv preprint arXiv:2406.06973, 2024. 1, 2

    2024 arXiv

  11. [20]

    Label-free regional consistency for image-to-image translation

    Shaohua Guo, Qianyu Zhou, Ye Zhou, Qiqi Gu, Junshu Tang, Zhengyang Feng, and Lizhuang Ma. Label-free regional consistency for image-to-image translation. In IEEE Interna- tional Conference on Multimedia and Expo, pages 1–6. IEEE,

  12. [21]

    Pointrwkv: Efficient rwkv-like model for hierarchical point cloud learning

    Qingdong He et al. Pointrwkv: Efficient rwkv-like model for hierarchical point cloud learning. arXiv preprint arXiv:2405.15214, 2024. 1, 2

    2024 arXiv

  13. [22]

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

    Martin Heusel, Hubert Ramsauer, Thomas Unterthiner, Bern- hard Nessler, and Sepp Hochreiter. Gans trained by a two time-scale update rule converge to a local nash equilibrium. NeurIPS, 30, 2017. 5

    2017

  14. [23]

    Denoising dif- fusion probabilistic models

    Jonathan Ho, Ajay Jain, and Pieter Abbeel. Denoising dif- fusion probabilistic models. NeurIPS, 33:6840–6851, 2020. 1

    2020

  15. [24]

    Aespa-net: Aesthetic pattern-aware style transfer networks

    Kibeom Hong et al. Aespa-net: Aesthetic pattern-aware style transfer networks. In ICCV, pages 22758–22767, 2023. 1, 2, 6

    2023

  16. [25]

    Zigma: Zigzag mamba diffusion model

    Vincent Tao Hu, Stefan Andreas Baumann, Ming Gui, Olga Grebenkova, Pingchuan Ma, Johannes Fischer, and Bjorn Ommer. Zigma: Zigzag mamba diffusion model. arXiv preprint arXiv:2403.13802, 2024. 4

    2024 arXiv

  17. [26]

    Arbitrary style transfer in real-time with adaptive instance normalization

    Xun Huang and Serge Belongie. Arbitrary style transfer in real-time with adaptive instance normalization. In ICCV, pages 1501–1510, 2017. 1, 2, 3, 6

    2017

  18. [27]

    Training- free style transfer emerges from h-space in diffusion models

    Jaeseok Jeong, Mingi Kwon, and Youngjung Uh. Training- free style transfer emerges from h-space in diffusion models. arXiv preprint arXiv:2303.15403, 3, 2023. 1, 2

    2023 arXiv

  19. [28]

    Perceptual losses for real-time style transfer and super-resolution

    Justin Johnson, Alexandre Alahi, and Li Fei-Fei. Perceptual losses for real-time style transfer and super-resolution. In Computer Vision–ECCV 2016: 14th European Conference, Amsterdam, The Netherlands, October 11-14, 2016, Proceed- ings, Part II 14, pages 694–711. Springer, 2016. 1, 2

    2016

  20. [29]

    Style transfer by relaxed optimal transport and self-similarity

    Nicholas Kolkin, Jason Salavon, and Gregory Shakhnarovich. Style transfer by relaxed optimal transport and self-similarity. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 10051–10060, 2019

    2019

  21. [30]

    Diffusion-based image translation using disentangled style and content representa- tion

    Gihyun Kwon and Jong Chul Ye. Diffusion-based image translation using disentangled style and content representa- tion. arXiv preprint arXiv:2209.15264, 2022. 1, 2, 6

    2022 arXiv

  22. [31]

    Combining markov random fields and convolutional neural networks for image synthesis

    Chuan Li and Michael Wand. Combining markov random fields and convolutional neural networks for image synthesis. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 2479–2486, 2016

    2016

  23. [32]

    Learning linear transformations for fast image and video style transfer

    Xueting Li, Sifei Liu, Jan Kautz, and Ming-Hsuan Yang. Learning linear transformations for fast image and video style transfer. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 3809–3817, 2019

    2019

  24. [33]

    Universal style transfer via feature transforms

    Yijun Li, Chen Fang, Jimei Yang, Zhaowen Wang, Xin Lu, and Ming-Hsuan Yang. Universal style transfer via feature transforms. NeurIPS, 30, 2017

    2017

  25. [34]

    Adaattn: Revisit attention mechanism in arbitrary neural style transfer

    Songhua Liu et al. Adaattn: Revisit attention mechanism in arbitrary neural style transfer. In ICCV, pages 6649–6658,

  26. [35]

    Dgmamba: Domain generalization via generalized state space model

    Shaocong Long, Qianyu Zhou, Xiangtai Li, Xuequan Lu, Chenhao Ying, Yuan Luo, Lizhuang Ma, and Shuicheng Yan. Dgmamba: Domain generalization via generalized state space model. In Proceedings of the 32nd ACM International Con- ference on Multimedia, pages 3607–3616, 2024. 1

    2024

  27. [36]

    Arbitrary style trans- fer with style-attentional networks

    Dae Young Park and Kwang Hee Lee. Arbitrary style trans- fer with style-attentional networks. In proceedings of the IEEE/CVF conference on computer vision and pattern recog- nition, pages 5880–5888, 2019. 2

    2019

  28. [37]

    Rwkv: Reinventing rnns for the transformer era

    Bo Peng, Alcaide, et al. Rwkv: Reinventing rnns for the transformer era. arXiv preprint arXiv:2305.13048, 2023. 1, 2, 4

    2023 arXiv

  29. [38]

    High-resolution image synthesis with latent diffusion models

    Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser, and Björn Ommer. High-resolution image synthesis with latent diffusion models. InCVPR, pages 10684– 10695, 2022. 1

    2022

  30. [39]

    Large-scale classification of fine-art paintings: Learning the right metric on the right feature

    Babak Saleh and Ahmed Elgammal. Large-scale classification of fine-art paintings: Learning the right metric on the right feature. arXiv preprint arXiv:1505.00855, 2015. 5

    2015 arXiv

  31. [40]

    Avatar- net: Multi-scale zero-shot style transfer by feature decoration

    Lu Sheng, Ziyi Lin, Jing Shao, and Xiaogang Wang. Avatar- net: Multi-scale zero-shot style transfer by feature decoration. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 8242–8250, 2018. 1, 2

    2018

  32. [41]

    Denoising diffusion implicit models

    Jiaming Song, Chenlin Meng, and Stefano Ermon. Denoising diffusion implicit models. arXiv preprint arXiv:2010.02502,

    2010 arXiv

  33. [42]

    Maxvit: Multi- axis vision transformer

    Zhengzhong Tu, Hossein Talebi, Han Zhang, Feng Yang, Peyman Milanfar, Alan Bovik, and Yinxiao Li. Maxvit: Multi- axis vision transformer. In European conference on computer vision, pages 459–479, 2022. 4

    2022

  34. [43]

    Texture networks: Feed-forward synthesis of tex- tures and stylized images

    Dmitry Ulyanov, Vadim Lebedev, Andrea Vedaldi, and Victor Lempitsky. Texture networks: Feed-forward synthesis of tex- tures and stylized images. arXiv preprint arXiv:1603.03417,

    arXiv

  35. [44]

    Attention is all you need

    A Vaswani. Attention is all you need. NeurIPS, 2017. 1

    2017

  36. [45]

    Stylemamba: State space model for efficient text-driven image style transfer

    Zijia Wang and Zhi-Song Liu. Stylemamba: State space model for efficient text-driven image style transfer. arXiv preprint arXiv:2405.05027, 2024. 1

    2024 arXiv

  37. [46]

    Stylediffusion: Controllable disentangled style transfer via diffusion models

    Zhizhong Wang, Lei Zhao, and Wei Xing. Stylediffusion: Controllable disentangled style transfer via diffusion models. In ICCV, pages 7677–7689, 2023. 1, 2

    2023

  38. [47]

    A comparative study of cnn-and transformer- based visual style transfer

    Hua-Peng Wei, Ying-Ying Deng, Fan Tang, Xing-Jia Pan, and Wei-Ming Dong. A comparative study of cnn-and transformer- based visual style transfer. Journal of Computer Science and Technology, 37(3):601–614, 2022. 1

    2022

  39. [48]

    Artfid: Quantitative evaluation of neural style transfer

    Matthias Wright and Björn Ommer. Artfid: Quantitative evaluation of neural style transfer. In DAGM GCPR, pages 560–576, 2022. 5

    2022

  40. [49]

    Pointdg- mamba: Domain generalization of point cloud classification via generalized state space model

    Hao Yang, Qianyu Zhou, Haijia Sun, Xiangtai Li, Fengqi Liu, Xuequan Lu, Lizhuang Ma, and Shuicheng Yan. Pointdg- mamba: Domain generalization of point cloud classification via generalized state space model. In Proceedings of the AAAI Conference on Artificial Intelligence, 2025. 1

    2025

  41. [50]

    Attention-aware multi-stroke style transfer

    Yuan Yao, Jianqiang Ren, Xuansong Xie, Weidong Liu, Yong- Jin Liu, and Jun Wang. Attention-aware multi-stroke style transfer. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 1467–1475,

  42. [51]

    Attention-aware multi-stroke style transfer

    Yuan Yao, Jianqiang Ren, Xuansong Xie, Weidong Liu, Yong- Jin Liu, and Jun Wang. Attention-aware multi-stroke style transfer. In CVPR, pages 1467–1475, 2019. 1, 2

    2019

  43. [52]

    Video rwkv: Video action recognition based rwkv

    Zhuowen Yin, Chengru Li, and Xingbo Dong. Video rwkv: Video action recognition based rwkv. arXiv preprint arXiv:2411.05636, 2024. 1, 2

    2024 arXiv

  44. [53]

    Multi-scale context aggregation by dilated convolutions

    F Yu. Multi-scale context aggregation by dilated convolutions. arXiv preprint arXiv:1511.07122, 2015. 4

    2015 arXiv

  45. [54]

    Mamba or rwkv: Exploring high-quality and high-efficiency segment anything model

    Haobo Yuan, Xiangtai Li, Lu Qi, Tao Zhang, Ming-Hsuan Yang, Shuicheng Yan, and Chen Change Loy. Mamba or rwkv: Exploring high-quality and high-efficiency segment anything model. arXiv preprint arXiv:2406.19369, 2024. 1, 2

    2024 arXiv

  46. [55]

    The unreasonable effectiveness of deep features as a perceptual metric

    Richard Zhang, Phillip Isola, Alexei A Efros, Eli Shechtman, and Oliver Wang. The unreasonable effectiveness of deep features as a perceptual metric. In CVPR, pages 586–595,

  47. [56]

    Exact feature distribution matching for arbitrary style transfer and domain generalization

    Yabin Zhang, Minghan Li, Ruihuang Li, Kui Jia, and Lei Zhang. Exact feature distribution matching for arbitrary style transfer and domain generalization. In CVPR, pages 8035– 8045, 2022. 1, 2, 6

    2022

  48. [57]

    Do- main enhanced arbitrary image style transfer via contrastive learning

    Yuxin Zhang, Fan Tang, Weiming Dong, Haibin Huang, Chongyang Ma, Tong-Yee Lee, and Changsheng Xu. Do- main enhanced arbitrary image style transfer via contrastive learning. In ACM SIGGRAPH, pages 1–8, 2022. 6

    2022

  49. [58]

    Inversion-based style transfer with diffusion models

    Yuxin Zhang, Nisha Huang, Fan Tang, Haibin Huang, Chongyang Ma, Weiming Dong, and Changsheng Xu. Inversion-based style transfer with diffusion models. In CVPR, pages 10146–10156, 2023. 1, 2, 6

    2023

  50. [59]

    Transvod: end-to-end video object detection with spatial- temporal transformers

    Qianyu Zhou, Xiangtai Li, Lu He, Yibo Yang, Guangliang Cheng, Yunhai Tong, Lizhuang Ma, and Dacheng Tao. Transvod: end-to-end video object detection with spatial- temporal transformers. IEEE Transactions on Pattern Analy- sis and Machine Intelligence, 45(6):7853–7869, 2023. 1

    2023

  51. [60]

    Vision mamba: Efficient visual representation learning with bidirectional state space model

    Lianghui Zhu, Bencheng Liao, Qian Zhang, Xinlong Wang, Wenyu Liu, and Xinggang Wang. Vision mamba: Efficient visual representation learning with bidirectional state space model. arXiv preprint arXiv:2401.09417, 2024. 4

    2024 arXiv

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.