Inverse-and-Edit: Effective and Fast Image Editing by Cycle Consistency Models

Aibek Alanov; Andrey Kuznetsov; Ilia Beletskii

REVIEW 4 major objections 6 minor 31 references

Inverse-and-Edit: Effective and Fast Image Editing by Cycle Consistency Models

T0 review · 4 major / 6 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read A cycle-consistency loss lets four-step consistency models edit images as well as 50-step diffusion models.

desk verdict The cycle-consistency fine-tuning is a real improvement over iCD, but the 'state-of-the-art' claim and hand-tuned evaluation protocol need substantial revision before the results are trustworthy. read the letter →

arxiv 2506.19103 v1 pith:4I56I3KR submitted 2025-06-23 cs.CV

Ilia Beletskii , Andrey Kuznetsov , Aibek Alanov This is my paper · ORCID

classification cs.CV

keywords imageeditingconsistencymodelsdiffusioninversioncycle-consistencylossdistilledzero-shotself-guidanceLPIPS

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

The paper tries to close the gap between fast and full-step diffusion editing. It argues that the bottleneck is inversion quality, not generation speed: distilled four-step models reconstruct source images poorly, so edits drift. The fix is a cycle-consistency fine-tuning of the forward consistency model, backpropagating a perceptual loss through the full noising-then-denoising loop. With that inversion improvement plus a self-guidance mechanism, the method claims editing results that match or surpass full-step diffusion models while using only four steps.

What carries the argument

The central mechanism is the cycle-consistency loss, a perceptual reconstruction objective evaluated after the complete inversion-and-generation cycle: encode the image, noise it through four forward consistency-model steps, denoise it through the frozen backward model, decode, and compare with LPIPS against the original. Because the loop is only eight model calls, end-to-end backpropagation is feasible, so the forward model is directly optimized for the reconstruction task, while consistency-distillation and forward-preservation losses keep its trajectory aligned with the original model.

What would settle it

Run the released method on a new, out-of-distribution set of 300 images and prompts using only the hyperparameters printed in the paper; if reconstruction LPIPS or DINOv2 similarity no longer beats the iCD baseline by the reported margin, the cycle-consistency gain is configuration-dependent rather than a property of the loss itself.

Watch

Extended reading notes

Core claim

On its own terms, the paper claims that fine-tuning the forward consistency model with the cycle-consistency loss $L_{\mathrm{rec}} = \mathrm{LPIPS}(G_{\theta_+}(F_{\theta_-}(x_0)), x_0)$ substantially improves image inversion and content preservation, measured by LPIPS, MSE, and ImageReward. Combined with a dynamic classifier-free guidance schedule and noise-rescaled self-attention and feature guidance, the resulting four-step editing pipeline matches the editing quality of full-step diffusion baselines such as Guide-and-Rescale, while outperforming other few-step methods, without needing blend words or Prompt-to-Prompt.

Load-bearing premise

The reported editing advantage relies on a hand-picked inference configuration (dynamic CFG schedule, guider weights, and noise-rescale bounds) that the paper says is varied during experiments, with no held-out validation described; if those settings are tuned to the evaluation benchmark, the gains may not generalize to new images and prompts.

Editorial extensions

If this is right

Four-step editing becomes a practical alternative to 50-step diffusion editing, cutting compute by roughly an order of magnitude without a visible quality drop.
Inversion quality in distilled consistency models can be improved without retraining the generation (backward) model, so existing fast-generation checkpoints can be reused.
Editing no longer depends on attention-manipulation tricks or hand-picked blend words; switching the prompt after a high-fidelity inversion is enough for many edits.
The trade-off between editability and content preservation can be controlled explicitly through guidance strength and noise-rescale bounds.
If the cycle-consistency gains transfer, other four-step editing methods built on the same distilled backbone could adopt the same fine-tuning and improve immediately.

Reading between the lines

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

The fine-tuning is done on MS-COCO only; a natural test is whether the inversion gain persists on out-of-domain images (e.g., paintings, medical scans, faces), where both the VAE and LPIPS behave differently.
A similar cycle-consistency objective could be applied to other distilled backbones (e.g., SDXL-Turbo or latent consistency models), where inversion quality is currently a bottleneck; the paper does not test this.
Because the guidance weights and CFG schedule are varied during experiments, a fair cross-method comparison would require fixing them on a validation set disjoint from the test set; the paper does not describe such a split.
The reported DINOv2-vs-LPIPS trade-off suggests the method favors semantic structure over pixel-level similarity; an interactive user study on target edits could confirm whether that bias matches human preference.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Desk Editor's Note

read the letter

Short version: this is a decent, useful paper. The central idea is simple—fine-tune only the forward consistency model so that noising followed by denoising reconstructs the input image, keeping the backward model frozen—and it works in their ablations. The gains in reconstruction and content preservation over the iCD baseline are clear, and the adaptation of Guide-and-Rescale to consistency models is a reasonable port. The code is public, which earns credit.

What's new: cycle-consistency fine-tuning of a forward CM for inversion is not in the prior work they cite. The contribution is incremental but not trivial: it makes 4-step editing much more usable.

Where it gets soft: the abstract says 'state-of-the-art' and 'matches or surpasses full-step diffusion models,' but Table 2 doesn't support that. TurboEdit beats them on ImageReward by a wide margin, and they don't uniformly win on LPIPS or CLIP either. The conclusion depends on which metric you emphasize. The bigger problem is the evaluation protocol. The CFG schedule (0, 7, 11, 19), guider weights (20000, 0.5), and rescale bounds are chosen during experiments on Pie-Bench, with no held-out validation split and no variance reporting. That means the comparison is against baselines at their defaults, at an operating point selected on the test set. That's a real threat to the generalization claim, and it's load-bearing for 'state-of-the-art.' Also, the reconstruction loss is LPIPS, and LPIPS is a headline metric, evaluated on COCO val after training on COCO train—so part of that improvement is by construction. The editing gains on Pie-Bench are not directly optimizing the editing metrics, so the circularity is partial, not fatal.

None of this is a dealbreaker. The method is plausible, the ablations support the core effect, and the limitations section is honest. But the paper needs a held-out validation procedure for the hyperparameters, error bars or seed variance, and a more careful claim. As is, I'd accept it for review, but I'd expect revision.

Who it's for: people working on efficient image editing with distilled diffusion models. I'd bring it to a reading group as a useful case study in how evaluation choices shape claims. I wouldn't cite it as a definitive result until the evaluation is tightened. Send it to a serious referee.

Referee Report

4 major / 6 minor

Summary. The paper proposes Inverse-and-Edit, a framework for four-step text-guided image editing built on invertible consistency distillation (iCD). The main idea is to fine-tune only the forward consistency model by backpropagating a cycle-consistency LPIPS reconstruction loss through the full four-step inversion-and-generation pipeline, while freezing the backward model. The authors also adapt the Guide-and-Rescale self-attention and feature guidance mechanism to guidance-distilled consistency models, together with a dynamic CFG schedule and noise rescaling bounds. The paper reports improved reconstruction quality on MS-COCO and improved content-preservation metrics on Pie-Bench relative to several few-step baselines, claiming performance comparable to or better than full-step diffusion methods while being faster.

Significance. If the central claims are supported, the main contribution is useful: the paper demonstrates that the short inference trajectory of distilled consistency models makes full-process cycle-consistency optimization feasible, and that this can improve inversion fidelity and editing quality without retraining the backward model. The method is clearly described, the code is promised to be public, and the ablation in Table 3 at least supports the qualitative direction that fine-tuning plus guidance changes the editability/preservation trade-off. The practical value is that the approach requires only four sampling steps and does not rely on blend words or Prompt-to-Prompt post-processing. However, the strength of the headline claims is not yet commensurate with the evidence: the evaluation protocol mixes multiple metrics without an explicit decision rule, the hyperparameters for the final operating point are described as chosen during experiments rather than on a held-out validation set, and some reported comparisons contradict the abstract's state-of-the-art wording.

major comments (4)

[§5.2, Table 2] The abstract and introduction claim that the method 'matches or surpasses full-step diffusion models' and achieves 'state-of-the-art performance', but Table 2 does not support a uniform superiority claim. For example, TurboEdit achieves ImageReward 0.536 versus 0.279 for Ours, and InfEdit achieves DINOv2 0.781 and LPIPS 0.236 versus 0.747 and 0.296 for Ours. The conclusion that the method is 'comparable' to full-step methods may be defensible, but the conclusion that it outperforms other accelerated models is contradicted by at least one baseline on one of the four reported metrics. The authors should either restrict the claim to specific metrics or datasets, or provide a statistically grounded multi-metric comparison with error bars and a stated aggregation rule.
[§5.2 and Appendix A.2] The final inference configuration is selected from choices that the authors explicitly report varying during experiments: the dynamic CFG schedule (0, 7, 11, 19), self-attention guider weight 20000, feature guider weight 0.5, and noise rescale bounds r_lower=0 and r_upper=1.0. No held-out validation split is used to select these values, and no ablation or sensitivity analysis is reported for them. Since all baselines are run at their default settings, the comparison in Table 2 risks reflecting favorable tuning of the proposed method rather than a general efficiency-quality advantage. The authors should report the selection procedure, the range of values explored, and the performance variation over that range, or validate the chosen configuration on an independent split of Pie-Bench.
[§4.1, Eq. (4); §5.1, Table 1] The cycle-consistency loss in Eq. (4) directly optimizes LPIPS, and the reconstruction evaluation in Table 1 uses LPIPS as a headline metric, on the same domain used for fine-tuning (COCO train versus COCO validation). The reported LPIPS improvement over iCD is therefore partly by construction and does not by itself establish that inversion quality is generally better. The authors should complement the LPIPS comparison with held-out, non-latent-aware distortion measures (e.g., human preference ratings, or metrics not aligned with the training loss) and should report the behavior of MSE and ImageReward more carefully, since those metrics do not show a uniform improvement over iCD or DDIM.
[§5.3, Table 3] Table 3 shows that adding guidance to the proposed method improves DINOv2 and LPIPS but degrades ImageReward (0.313 to 0.279) and CLIPScore (0.304 to 0.302). The text states that guidance improves content preservation and calls this 'key to better visual quality', but the reported metrics move in opposite directions, so the claimed benefit is metric-selective. The authors should either present a user study or a clear preference-based evaluation to support the claim that the guided operating point is visually better, or explicitly frame the contribution as a trade-off controller rather than an unambiguous improvement.

minor comments (6)

[§3, 'Inversion in consistency models'] The sentence 'The consistency distillation loss from Equation (4) is adapted for both the fCM and CM training objectives' appears to refer to Eq. (3), since Eq. (4) is the reconstruction loss introduced later; please update the cross-reference.
[Appendix A.1] The phrase 'for fCD and the forward preservation loss' appears to contain a typo; it should likely read 'fCM'.
[Table 1] Several rows in Table 1 are typeset without separators between the metric value and the model name (e.g., '0.424iCD (4 steps)', '0.372ReNoise SDXL-Turbo (4 steps)', '0.444Ours (4 steps)'), making the table difficult to read.
[§5.1] The statement that 'most of the error attributed to approximation mismatches between adjacent timesteps' is presented without supporting evidence; a quantitative attribution or at least a diagnostic experiment would be needed to justify this claim.
[§5.2] The paper does not state the number of seeds or runs used to produce Table 2, and with 420 Pie-Bench images some reported metric differences may be within noise; please add a statement on variance or provide per-run results.
[References] The Null-text Inversion reference is duplicated in the bibliography (Mokady et al., 2022 and Mokady et al., 2023); the duplicate entry should be removed or cross-referenced properly.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the cycle-consistency loss is a direct training objective evaluated on held-out data, and the self-citation to Guide-and-Rescale is not load-bearing.

full rationale

The paper's central derivation is an explicit optimization: Eq. (4) fine-tunes the forward consistency model by minimizing LPIPS(G_theta+(F_theta-(x0)), x0). The reconstruction evaluation in Table 1 reports LPIPS, MSE, and ImageReward on the held-out MS-COCO validation split, which is disjoint from the COCO training split used for fine-tuning, so the reported LPIPS gain is a generalization measurement rather than an identity. The editing evaluation on Pie-Bench uses target-prompt edits and metrics (DINOv2, CLIP, ImageReward, LPIPS) that are not directly optimized by Eq. (4), giving independent content. The guidance mechanism is adapted from Guide-and-Rescale, a published method with one overlapping author, but the paper does not invoke it as an unverified self-citation or uniqueness theorem. The dynamic CFG schedule and guider weights were selected during experiments, which is an evaluation-protocol robustness concern, not a definitional circularity. No equation reduces to its own inputs and no fitted parameter is renamed as a prediction.

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

The method depends on the iCD forward/backward consistency-model framework and on several manually chosen inference hyperparameters. The training setup itself follows iCD defaults, but the guider weights and CFG schedule used for the final reported numbers are hand-tuned without a separate validation split. No new physical or architectural entities are introduced.

free parameters (4)

Dynamic CFG schedule = [0, 7, 11, 19] across four steps
Hand-tuned in experiments to balance editability and over-saturation; no separate validation set for selection (Section 5.2, Appendix A.2).
Self-attention guider weight = 20000
Chosen manually for editing experiments; not justified by a held-out set (Appendix A.2).
Feature guider weight = 0.5
Chosen manually alongside the attention weight (Appendix A.2).
Noise rescale bounds = r_lower=0, r_upper=1.0
Heuristics to limit guider influence on editability; introduced ad hoc (Appendix A.2).

assumptions (4)

domain assumption Pretrained iCD forward and backward consistency models provide a valid inversion-generation framework.
The method initializes from public iCD checkpoints and assumes their forward/backward architecture and preservation losses are sound (Section 4.1).
domain assumption LPIPS (VGG-16) is an adequate proxy for structural and perceptual fidelity in reconstruction and editing.
LPIPS is used as the fine-tuning loss and as a headline evaluation metric; the paper does not validate that LPIPS correlates with human judgments for this task (Sections 4.1, 5).
domain assumption Self-attention and feature guiders from Guide-and-Rescale transfer to consistency models.
The guidance mechanism is adapted from Titov et al. 2024, and the paper assumes the same energy functions work with a guidance-distilled consistency model (Section 4.2).
domain assumption Fine-tuning on MS-COCO training images generalizes to Pie-Bench and other editing distributions.
The forward model is fine-tuned on COCO training data and then evaluated on Pie-Bench and custom images; the paper provides no out-of-distribution analysis (Sections 5, 5.2).

how reviews work

0 comments

Cite this review

Pith. "Pith review of Inverse-and-Edit: Effective and Fast Image Editing by Cycle Consistency Models." pith.science (2026). https://pith.science/paper/4I56I3KR

@misc{pith2026250619103,
  author       = {Pith},
  title        = {Pith review of: Inverse-and-Edit: Effective and Fast Image Editing by Cycle Consistency Models},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/4I56I3KR}},
  note         = {Machine review of arXiv:2506.19103}
}

read the original abstract

Recent advances in image editing with diffusion models have achieved impressive results, offering fine-grained control over the generation process. However, these methods are computationally intensive because of their iterative nature. While distilled diffusion models enable faster inference, their editing capabilities remain limited, primarily because of poor inversion quality. High-fidelity inversion and reconstruction are essential for precise image editing, as they preserve the structural and semantic integrity of the source image. In this work, we propose a novel framework that enhances image inversion using consistency models, enabling high-quality editing in just four steps. Our method introduces a cycle-consistency optimization strategy that significantly improves reconstruction accuracy and enables a controllable trade-off between editability and content preservation. We achieve state-of-the-art performance across various image editing tasks and datasets, demonstrating that our method matches or surpasses full-step diffusion models while being substantially more efficient. The code of our method is available on GitHub at https://github.com/ControlGenAI/Inverse-and-Edit.

Figures

Figures reproduced from arXiv: 2506.19103 by the authors.

**Figure 2.** (a): Visual comparison of editing results produced by our fine-tuned model and the baseline. [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗

**Figure 4.** Schematic illustration of the Cycle-Consistency method with guidance. Image editing is [PITH_FULL_IMAGE:figures/full_fig_p006_4.png] view at source ↗

Figures from the paper (8 more)

**Figure 3.** Figure 3: Visualization of editing results produced by prompt switching (second column) and by our method with guidance (third column). Noise rescale Strong guidance from energy functions can significantly reduce editability. Guide-and-Rescale Titov et al. [2024] proposes resc…

**Figure 5.** Figure 5: Examples of image reconstruction obtained from our method and from other approaches. [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]

**Figure 6.** Figure 6: Quantitative evaluation of our method and other approaches on the image reconstruction task on the MSCOCO validation set. 5.2 Text-guided image editing To validate our approach, we compare it with leading fast methods (iCD [Starodubcev et al., 2024], TurboEdit [Deutch…

**Figure 7.** Figure 7: Examples of image editing results obtained using our method with guidance and other [PITH_FULL_IMAGE:figures/full_fig_p009_7.png]

**Figure 8.** Figure 8: Quantitative evaluation of editing results obtained by our method and other baselines on the Pie-Bench dataset 6 Conclusion We propose a novel approach for forward consistency model optimization over the entire process of image reconstruction, including inversion and …

**Figure 9.** Figure 9: Diagram of our fine-tune method. We optimize the forward consistency model by back [PITH_FULL_IMAGE:figures/full_fig_p012_9.png]

**Figure 10.** Figure 10: Examples of image reconstruction obtained using our method and from other approaches. [PITH_FULL_IMAGE:figures/full_fig_p013_10.png]

**Figure 11.** Figure 11: Examples of image editing results obtained using our method with guidance and other [PITH_FULL_IMAGE:figures/full_fig_p014_11.png]

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

31 extracted references · 1 canonical work pages

[1]

Universal guidance for diffusion models, 2023

Arpit Bansal, Hong-Min Chu, Avi Schwarzschild, Soumyadip Sengupta, Micah Goldblum, Jonas Geiping, and Tom Goldstein. Universal guidance for diffusion models, 2023. URL https://arxiv.org/abs/2302.07121

arXiv 2023
[2]

Masactrl: Tuning-free mutual self-attention control for consistent image synthesis and editing, 2023

Mingdeng Cao, Xintao Wang, Zhongang Qi, Ying Shan, Xiaohu Qie, and Yinqiang Zheng. Masactrl: Tuning-free mutual self-attention control for consistent image synthesis and editing, 2023. URL https://arxiv.org/abs/2304.08465

arXiv 2023
[3]

Turboedit: Text-based image editing using few-step diffusion models, 2024

Gilad Deutch, Rinon Gal, Daniel Garibi, Or Patashnik, and Daniel Cohen-Or. Turboedit: Text-based image editing using few-step diffusion models, 2024. URL https://arxiv.org/abs/2408.00735

arXiv 2024
[4]

Renoise: Real image inversion through iterative noising, 2024

Daniel Garibi, Or Patashnik, Andrey Voynov, Hadar Averbuch-Elor, and Daniel Cohen-Or. Renoise: Real image inversion through iterative noising, 2024. URL https://arxiv.org/abs/2403.14602

arXiv 2024
[5]

Multistep consistency models, 2024

Jonathan Heek, Emiel Hoogeboom, and Tim Salimans. Multistep consistency models, 2024. URL https://arxiv.org/abs/2403.06807

arXiv 2024
[6]

Prompt-to-prompt image editing with cross attention control, 2022

Amir Hertz, Ron Mokady, Jay Tenenbaum, Kfir Aberman, Yael Pritch, and Daniel Cohen-Or. Prompt-to-prompt image editing with cross attention control, 2022. URL https://arxiv.org/abs/2208.01626

arXiv 2022
[7]

Clipscore: A reference-free evaluation metric for image captioning, 2022

Jack Hessel, Ari Holtzman, Maxwell Forbes, Ronan Le Bras, and Yejin Choi. Clipscore: A reference-free evaluation metric for image captioning, 2022. URL https://arxiv.org/abs/2104.08718

arXiv 2022
[8]

Classifier-free diffusion guidance, 2022

Jonathan Ho and Tim Salimans. Classifier-free diffusion guidance, 2022. URL https://arxiv.org/abs/2207.12598

arXiv 2022

Show all 31 references

[9]

Denoising diffusion probabilistic models, 2020

Jonathan Ho, Ajay Jain, and Pieter Abbeel. Denoising diffusion probabilistic models, 2020. URL https://arxiv.org/abs/2006.11239

2020 arXiv
[10]

Direct inversion: Boosting diffusion-based editing with 3 lines of code, 2023

Xuan Ju, Ailing Zeng, Yuxuan Bian, Shaoteng Liu, and Qiang Xu. Direct inversion: Boosting diffusion-based editing with 3 lines of code, 2023. URL https://arxiv.org/abs/2310.01506

2023 arXiv
[11]

Lawrence Zitnick, and Piotr Dollár

Tsung-Yi Lin, Michael Maire, Serge Belongie, Lubomir Bourdev, Ross Girshick, James Hays, Pietro Perona, Deva Ramanan, C. Lawrence Zitnick, and Piotr Dollár. Microsoft coco: Common objects in context, 2015. URL https://arxiv.org/abs/1405.0312

2015 arXiv
[12]

Kingma, Stefano Ermon, Jonathan Ho, and Tim Salimans

Chenlin Meng, Robin Rombach, Ruiqi Gao, Diederik P. Kingma, Stefano Ermon, Jonathan Ho, and Tim Salimans. On distillation of guided diffusion models, 2023. URL https://arxiv.org/abs/2210.03142

2023 arXiv
[13]

Negative-prompt inversion: Fast image inversion for editing with text-guided diffusion models, 2024

Daiki Miyake, Akihiro Iohara, Yu Saito, and Toshiyuki Tanaka. Negative-prompt inversion: Fast image inversion for editing with text-guided diffusion models, 2024. URL https://arxiv.org/abs/2305.16807

2024 arXiv
[14]

Null-text inversion for editing real images using guided diffusion models, 2022

Ron Mokady, Amir Hertz, Kfir Aberman, Yael Pritch, and Daniel Cohen-Or. Null-text inversion for editing real images using guided diffusion models, 2022. URL https://arxiv.org/abs/2211.09794

2022 arXiv
[15]

Null-text inversion for editing real images using guided diffusion models

Ron Mokady, Amir Hertz, Kfir Aberman, Yael Pritch, and Daniel Cohen-Or. Null-text inversion for editing real images using guided diffusion models. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 6038--6047, June 2023

2023
[16]

Dinov2: Learning robust visual features without supervision, 2024

Maxime Oquab, Timothée Darcet, Théo Moutakanni, Huy Vo, Marc Szafraniec, Vasil Khalidov, Pierre Fernandez, Daniel Haziza, Francisco Massa, Alaaeldin El-Nouby, Mahmoud Assran, Nicolas Ballas, Wojciech Galuba, Russell Howes, Po-Yao Huang, Shang-Wen Li, Ishan Misra, Michael Rabba...

2024 arXiv
[17]

High-resolution image synthesis with latent diffusion models

Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser, and Bj\"orn Ommer. High-resolution image synthesis with latent diffusion models. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 10684--10695, June 2022

2022
[18]

Progressive distillation for fast sampling of diffusion models, 2022

Tim Salimans and Jonathan Ho. Progressive distillation for fast sampling of diffusion models, 2022. URL https://arxiv.org/abs/2202.00512

2022 arXiv
[19]

Lightning-fast image inversion and editing for text-to-image diffusion models, 2025

Dvir Samuel, Barak Meiri, Haggai Maron, Yoad Tewel, Nir Darshan, Shai Avidan, Gal Chechik, and Rami Ben-Ari. Lightning-fast image inversion and editing for text-to-image diffusion models, 2025. URL https://arxiv.org/abs/2312.12540

2025 arXiv
[20]

Fast high-resolution image synthesis with latent adversarial diffusion distillation, 2024

Axel Sauer, Frederic Boesel, Tim Dockhorn, Andreas Blattmann, Patrick Esser, and Robin Rombach. Fast high-resolution image synthesis with latent adversarial diffusion distillation, 2024. URL https://arxiv.org/abs/2403.12015

2024 arXiv
[21]

Very deep convolutional networks for large-scale image recognition, 2015

Karen Simonyan and Andrew Zisserman. Very deep convolutional networks for large-scale image recognition, 2015. URL https://arxiv.org/abs/1409.1556

2015 arXiv
[22]

Denoising diffusion implicit models, 2022

Jiaming Song, Chenlin Meng, and Stefano Ermon. Denoising diffusion implicit models, 2022. URL https://arxiv.org/abs/2010.02502

2022 arXiv
[23]

Kingma, Abhishek Kumar, Stefano Ermon, and Ben Poole

Yang Song, Jascha Sohl-Dickstein, Diederik P. Kingma, Abhishek Kumar, Stefano Ermon, and Ben Poole. Score-based generative modeling through stochastic differential equations, 2021. URL https://arxiv.org/abs/2011.13456

2021 arXiv
[24]

Consistency models, 2023

Yang Song, Prafulla Dhariwal, Mark Chen, and Ilya Sutskever. Consistency models, 2023. URL https://arxiv.org/abs/2303.01469

2023 arXiv
[25]

Invertible consistency distillation for text-guided image editing in around 7 steps, 2024

Nikita Starodubcev, Mikhail Khoroshikh, Artem Babenko, and Dmitry Baranchuk. Invertible consistency distillation for text-guided image editing in around 7 steps, 2024. URL https://arxiv.org/abs/2406.14539

2024 arXiv
[26]

Postedit: Posterior sampling for efficient zero-shot image editing, 2025

Feng Tian, Yixuan Li, Yichao Yan, Shanyan Guan, Yanhao Ge, and Xiaokang Yang. Postedit: Posterior sampling for efficient zero-shot image editing, 2025. URL https://arxiv.org/abs/2410.04844

2025 arXiv
[27]

Guide-and-rescale: Self-guidance mechanism for effective tuning-free real image editing, 2024

Vadim Titov, Madina Khalmatova, Alexandra Ivanova, Dmitry Vetrov, and Aibek Alanov. Guide-and-rescale: Self-guidance mechanism for effective tuning-free real image editing, 2024. URL https://arxiv.org/abs/2409.01322

2024 arXiv
[28]

Imagereward: Learning and evaluating human preferences for text-to-image generation, 2023 a

Jiazheng Xu, Xiao Liu, Yuchen Wu, Yuxuan Tong, Qinkai Li, Ming Ding, Jie Tang, and Yuxiao Dong. Imagereward: Learning and evaluating human preferences for text-to-image generation, 2023 a . URL https://arxiv.org/abs/2304.05977

2023 arXiv
[29]

Inversion-free image editing with natural language, 2023 b

Sihan Xu, Yidong Huang, Jiayi Pan, Ziqiao Ma, and Joyce Chai. Inversion-free image editing with natural language, 2023 b . URL https://arxiv.org/abs/2312.04965

2023 arXiv
[30]

Freeman, and Taesung Park

Tianwei Yin, Michaël Gharbi, Richard Zhang, Eli Shechtman, Fredo Durand, William T. Freeman, and Taesung Park. One-step diffusion with distribution matching distillation, 2024. URL https://arxiv.org/abs/2311.18828

2024 arXiv
[31]

Efros, Eli Shechtman, and Oliver Wang

Richard Zhang, Phillip Isola, Alexei A. Efros, Eli Shechtman, and Oliver Wang. The unreasonable effectiveness of deep features as a perceptual metric, 2018. URL https://arxiv.org/abs/1801.03924

2018 arXiv

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.

[1] [1]

Universal guidance for diffusion models, 2023

Arpit Bansal, Hong-Min Chu, Avi Schwarzschild, Soumyadip Sengupta, Micah Goldblum, Jonas Geiping, and Tom Goldstein. Universal guidance for diffusion models, 2023. URL https://arxiv.org/abs/2302.07121

arXiv 2023

[2] [2]

Masactrl: Tuning-free mutual self-attention control for consistent image synthesis and editing, 2023

Mingdeng Cao, Xintao Wang, Zhongang Qi, Ying Shan, Xiaohu Qie, and Yinqiang Zheng. Masactrl: Tuning-free mutual self-attention control for consistent image synthesis and editing, 2023. URL https://arxiv.org/abs/2304.08465

arXiv 2023

[3] [3]

Turboedit: Text-based image editing using few-step diffusion models, 2024

Gilad Deutch, Rinon Gal, Daniel Garibi, Or Patashnik, and Daniel Cohen-Or. Turboedit: Text-based image editing using few-step diffusion models, 2024. URL https://arxiv.org/abs/2408.00735

arXiv 2024

[4] [4]

Renoise: Real image inversion through iterative noising, 2024

Daniel Garibi, Or Patashnik, Andrey Voynov, Hadar Averbuch-Elor, and Daniel Cohen-Or. Renoise: Real image inversion through iterative noising, 2024. URL https://arxiv.org/abs/2403.14602

arXiv 2024

[5] [5]

Multistep consistency models, 2024

Jonathan Heek, Emiel Hoogeboom, and Tim Salimans. Multistep consistency models, 2024. URL https://arxiv.org/abs/2403.06807

arXiv 2024

[6] [6]

Prompt-to-prompt image editing with cross attention control, 2022

Amir Hertz, Ron Mokady, Jay Tenenbaum, Kfir Aberman, Yael Pritch, and Daniel Cohen-Or. Prompt-to-prompt image editing with cross attention control, 2022. URL https://arxiv.org/abs/2208.01626

arXiv 2022

[7] [7]

Clipscore: A reference-free evaluation metric for image captioning, 2022

Jack Hessel, Ari Holtzman, Maxwell Forbes, Ronan Le Bras, and Yejin Choi. Clipscore: A reference-free evaluation metric for image captioning, 2022. URL https://arxiv.org/abs/2104.08718

arXiv 2022

[8] [8]

Classifier-free diffusion guidance, 2022

Jonathan Ho and Tim Salimans. Classifier-free diffusion guidance, 2022. URL https://arxiv.org/abs/2207.12598

arXiv 2022