REVIEW 4 major objections 6 minor 1 cited by
Normalized Attention Guidance: Universal Negative Guidance for Diffusion Models
T0 review · 4 major / 6 minor · reviewed 2026-08-07 · deepseek-v4-flash
Pith's one-line read Attention-space trick revives negative guidance in few-step diffusion
desk verdict A genuinely useful stabilization of attention-space negative guidance, but the paper's central claim that it provides semantic negative guidance is not yet isolated from generic quality enhancement. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The central object is the NAG feature transform applied inside cross-attention layers: $\tilde{Z} = Z^{+} + \phi (Z^{+} - Z^{-})$, followed by token-wise L1-norm ratio clipping, $\hat{Z}[i] = \frac{\min(R[i], \tau)}{R[i]} \tilde{Z}[i]$ with $R[i] = \frac{\| \tilde{Z}[i] \|_1}{\| Z^{+}[i] \|_1}$, and final refinement $Z_{\mathrm{NAG}} = \alpha \hat{Z} + (1-\alpha) Z^{+}$. This transform replaces output-space CFG extrapolation with a bounded, direction-preserving move in attention space; the two stabilization operations work as a magnitude guardrail and a manifold anchor, respectively, and are what let the guidance remain stable at high scale.
What would settle it
Reproduce Table 1 on 4-step Flux-Schnell with the universal negative prompt 'Low resolution, blurry'. The claim would be falsified if (a) the reported CLIP, FID, and ImageReward improvements do not reproduce, (b) ablations without normalization or without blending match or beat the full method, or (c) at a fixed scale the extrapolated attention features $\hat{Z}$ land measurably far outside the distribution of unmodified $Z^{+}$ in a way that predicts the observed artifacts.
Extended reading notes
Core claim
The paper's central claim is that effective negative guidance can be recovered in few-step diffusion models by moving extrapolation out of the predicted-noise/output space and into the attention feature space, with two stabilizing operations that prior attention-manipulation methods like NASA lack. Specifically, NAG computes attention outputs $Z^{+}$ and $Z^{-}$ from the positive and negative prompts, forms an extrapolated feature $\tilde{Z} = Z^{+} + \phi (Z^{+} - Z^{-})$, then rescales each token's L1 norm to a bounded threshold $\tau$ and blends the result with $Z^{+}$ by a factor $\alpha$. The authors argue the L1 normalization preserves low-magnitude activations that carry subtle semantics while constraining extremes, and that the blending pulls features back toward the stable positive manifold; together these prevent the out-of-manifold drift and feature collapse they observe in prior methods. They claim this makes NAG a universal plug-in: it works where CFG collapses (few-step UNet and DiT models such as Flux-Schnell, SDXL-Lightning, and DMD2-SDXL), improves multi-step models when added on top of CFG and PAG, and transfers to video diffusion (Wan2.1) for both content suppression and motion enhancement.
Load-bearing premise
The whole method rests on the heuristic that L1-norm clipping plus alpha-blending keeps extrapolated attention features close enough to the true feature manifold, and that this holds across architectures and modalities without per-model tuning; the paper does not derive this from theory, only from the demonstrated experiments.
Editorial extensions
If this is right
- Few-step models that ship without CFG support (Flux-Schnell, SANA-Sprint, SD3.5-Large-Turbo, NitroSD-Realism, DMD2-SDXL, SDXL-Lightning) gain a working negative-prompting interface without retraining.
- When combined with CFG and PAG in 25-step models, NAG improves CLIP score and ImageReward while leaving FID and PFID roughly flat or slightly better, giving a way to add negative control on top of existing guidance.
- Applying NAG only for the first fraction of denoising steps (e.g., $\theta = 0.25$) keeps most of the measured benefit while cutting the added latency, since the guidance effect is concentrated in early timesteps.
- NAG extends to video generation on Wan2.1, suppressing undesired content such as human faces or static motion and improving temporal dynamics.
- The method requires no retraining and adds less latency than CFG on most model families (e.g., +13% on SANA and +12% on Wan2.1 versus +90% to +100% for CFG), making it a practical drop-in at inference time.
Reading between the lines
- The consistent behavior under a fixed L1 clip and blend across very different architectures hints that attention features in these models share a common scale structure; one could test this directly by measuring the distribution of $\|Z\|_1$ across layers and models and checking whether the default $\tau = 2.5$ tracks a consistent quantile.
- Because NAG operates on frozen attention features rather than on training, the same mechanism could be aimed at other steering objectives—concept erasure, style transfer, or prompt-based debiasing—provided a negative description can be written; the paper demonstrates the prompting use, not the full space of guidance tasks.
- The early-stopping result suggests the semantic direction of a negative prompt is mostly determined in the first denoising steps; if bias suppression is applied only there, downstream steps remain untouched, which might make NAG-based debiasing less likely to degrade fine texture.
- Nothing in the method requires the negative condition to be textual; using an image as the negative condition could test whether NAG's attention-space extrapolation works for reference-based editing.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper proposes Normalized Attention Guidance (NAG), a training-free inference-time method for negative prompting in diffusion models. NAG computes attention outputs for positive and negative prompts, extrapolates in attention feature space as eZ = Z+ + φ(Z+ - Z-), applies L1-ratio clipping with threshold τ, and blends with the positive output using α. The authors claim that NAG restores effective negative guidance where CFG collapses, particularly in few-step models, and that it generalizes across UNet/DiT architectures, image/video modalities, and few-/multi-step settings. Experiments on seven few-step models, several multi-step CFG models, and one video model report CLIP/FID/PFID/ImageReward gains, supplemented by ablations, latency measurements, and a user study.
Significance. If the causal attribution to negative-attribute suppression is established, NAG would be a practical and broadly applicable plug-in for diffusion models. The method is simple and clearly specified (Algorithm 1, Table 5), and the paper provides unusually broad model coverage, including recent DiT models and a 14B video model, along with latency measurements and a component ablation. The authors are also candid about failure cases in Section 6. However, the current evaluation does not isolate negative-attribute suppression from generic quality enhancement, so the central claim is not yet verified. The paper also lacks comparisons with standard negative-prompt baselines and statistical uncertainty estimates, which are necessary to support the 'universal' and 'consistent improvements' claims.
major comments (4)
- [§5.1, Table 1, §5.5] The quantitative evaluation relies on the single negative prompt 'Low resolution, blurry' and on aggregate quality metrics (CLIP, FID, PFID, ImageReward) that would also improve from generic sharpening or contrast enhancement. The paper does not report any direct measure of the presence of the suppressed attribute (e.g., classifier accuracy for 'glasses', 'tiger', or 'male', or CLIP similarity to the negative text), and the user study in Section 5.5 instructs participants to judge text alignment while considering both positive and negative prompts, so it cannot separate suppression from aesthetic improvement. Consequently, the central claim that NAG implements directional negative guidance rather than acting as a generic feature regularizer is not yet supported. I recommend adding attribute-presence metrics on the generated samples for several negative concepts, or a controlled experiment in which the negative prompt targets a semantic attribute while the positive prompt does not mention it.
- [§5.1–5.3] There is no baseline against standard text-level negative prompting. For multi-step models, CFG with a negative prompt is the default practice; for few-step models, straightforward baselines such as prompt inversion or simply appending the negative prompt to the positive prompt could be used. Without such comparisons, the gains in Tables 1 and 2 could stem from the additional negative conditioning per se rather than from the attention-space mechanism of NAG. Please add these baselines to the evaluation.
- [Table 5, Table 2, Table 6] The hyperparameters (ϕ, τ, α) in Table 5 are chosen per model family, and the paper does not state whether they were tuned on the same COCO-5K evaluation set used in Tables 1, 2, and 6. Moreover, no error bars or repeated runs are reported; several FID and PFID differences are within typical noise (e.g., SD3.5-Large CFG FID +0.35 in Table 2, SD1.5 FID +0.14 and PFID +1.29 in Table 6). This weakens the 'consistent improvements' and 'universal' claims. Please report the tuning protocol and provide confidence intervals or at least per-prompt variance.
- [§5.2, Appendix E] The comparison with NASA is limited to UNet models, and the user study against NASA (Table 8) is conducted on only one model (DMD2-SDXL). Since the paper motivates NAG partly by NASA's instability on DiT architectures, a quantitative DiT comparison (even if it shows NASA collapse) would strengthen the claim. Also, the attribute-specific suppression of NAG versus NASA is not compared directly, so it remains unclear whether NAG's advantage is in guidance direction or in overall stability.
minor comments (6)
- [Algorithm 1] The computed z_nag is never assigned to hidden_states; the function returns the unmodified hidden_states. Replace 'return hidden_states' with 'return z_nag' or add an assignment before the return.
- [Equations (8)–(9)] The symbol R is used both for the per-token ratio R[i] and for the vector in Eq. (9); please use a different symbol for the vector (e.g., lowercase r) and make the element-wise division explicit.
- [References and Table 2] Table 2 and the main text cite PAG as [12], but [12] is 'Rectified diffusion guidance' whereas Perturbed Attention Guidance is [59]; please correct the citation.
- [Conclusion] There is a typo: 'the the limitations' should be 'the limitations,' and 'CFG ... fails' should agree in number with 'CFG'.
- [Figure 4] The terms 'Guidance Boundary' and 'Refinement Manifold' are introduced only in the caption; please define them in the text or make the figure self-contained.
- [Section 5.4] The video evaluation is qualitative only; the contribution statement in Section 1 about improving 'motion characteristics' would benefit from a quantitative metric on the generated videos.
Circularity Check
No circularity: NAG is an empirical, measured method; no claimed derivation reduces to its inputs.
full rationale
This paper proposes an inference-time heuristic (attention-space extrapolation with L1 normalization and α-blending) and evaluates it with standard generative metrics. There is no derivation chain in which a predicted quantity is defined in terms of a fitted quantity or in which an output is equivalent to an input by construction. The method's components—extrapolation (Eq. 7), normalization (Eqs. 8–9), and refinement (Eq. 10)—are stated as operations, not derived from the metrics they are judged against. The claims of restored negative guidance are supported by measured CLIP/FID/PFID/ImageReward changes and user studies; these are empirical observations rather than consequences of the method's definitions. The few self-citations (e.g., Nitrofusion as an evaluation backbone) are not load-bearing for the central claim. The choice of the universal negative prompt 'Low resolution, blurry' and per-model-family hyperparameters may raise evaluation-validity questions (e.g., in-sample tuning or quality-enhancement confounds), but those are not circular reasoning: the reported numbers are not forced by the equations. No circular step can be quoted or exhibited, so the appropriate finding is no significant circularity.
Assumptions & free parameters
free parameters (3)
- Guidance scale phi =
Per model family: Flux 4, SD3.5-Large 4, SANA 4, PixArt-Sigma 4, Wan2.1 4, SDXL 2, Playground 2, SD1.5 2
- Refinement factor alpha =
Per model family: Flux 0.25, SD3.5-Large 0.125, SANA 0.375, PixArt-Sigma 0.375, Wan2.1 0.25, SDXL 0.5, Playground 0.5…
- Norm threshold tau =
2.5 for all model families
assumptions (4)
- domain assumption Cross-attention features encode semantic information in the direction of the prompt.
- domain assumption L1 norm preserves low-magnitude activations that encode subtle semantics better than L2.
- domain assumption Blending with the positive baseline pulls features back to the data manifold.
- domain assumption The pre-trained diffusion model is well-calibrated enough to be guided by attention feature perturbations.
Cite this review
Pith. "Pith review of Normalized Attention Guidance: Universal Negative Guidance for Diffusion Models." pith.science (2026). https://pith.science/paper/3H4R47TT
@misc{pith2026250521179,
author = {Pith},
title = {Pith review of: Normalized Attention Guidance: Universal Negative Guidance for Diffusion Models},
year = {2026},
howpublished = {\url{https://pith.science/paper/3H4R47TT}},
note = {Machine review of arXiv:2505.21179}
}
read the original abstract
Negative guidance -- explicitly suppressing unwanted attributes -- remains a fundamental challenge in diffusion models, particularly in few-step sampling regimes. While Classifier-Free Guidance (CFG) works well in standard settings, it fails under aggressive sampling step compression due to divergent predictions between positive and negative branches. We present Normalized Attention Guidance (NAG), an efficient, training-free mechanism that applies extrapolation in attention space with L1-based normalization and refinement. NAG restores effective negative guidance where CFG collapses while maintaining fidelity. Unlike existing approaches, NAG generalizes across architectures (UNet, DiT), sampling regimes (few-step, multi-step), and modalities (image, video), functioning as a \textit{universal} plug-in with minimal computational overhead. Through extensive experimentation, we demonstrate consistent improvements in text alignment (CLIP Score), fidelity (FID, PFID), and human-perceived quality (ImageReward). Our ablation studies validate each design component, while user studies confirm significant preference for NAG-guided outputs. As a model-agnostic inference-time approach requiring no retraining, NAG provides effortless negative guidance for all modern diffusion frameworks -- pseudocode in the Appendix!
Figures
Figures from the paper (19 more)
Forward citations
Cited by 1 Pith paper
-
MissingBench-Verified: Probing Vision-Language Models' Inability to Detect Missing Object Parts
Ten leading VLMs mostly fail to report removed essential object parts as missing, and simulated detector evidence, image tools, longer reasoning, and an easier fine-tune barely improve accuracy.
Reference graph
Works this paper leans on
-
[1]
Black Forest Labs. Flux. https://github.com/black-forest-labs/flux , 2024
2024
-
[2]
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. In CVPR, 2022
2022
-
[3]
Denoising diffusion probabilistic models
Jonathan Ho, Ajay Jain, and Pieter Abbeel. Denoising diffusion probabilistic models. In NeurIPS, 2020
2020
-
[4]
Photorealistic text-to- image diffusion models with deep language understanding
Chitwan Saharia, William Chan, Saurabh Saxena, Lala Li, Jay Whang, Emily L Denton, Kamyar Ghasemipour, Raphael Gontijo Lopes, Burcu Karagol Ayan, Tim Salimans, et al. Photorealistic text-to- image diffusion models with deep language understanding. In NeurIPS, 2022
2022
-
[5]
Kingma, Ben Poole, Mohammad Norouzi, David J
Jonathan Ho, William Chan, Chitwan Saharia, Jay Whang, Ruiqi Gao, Alexey Gritsenko, Diederik P. Kingma, Ben Poole, Mohammad Norouzi, David J. Fleet, and Tim Salimans. Imagen video: High definition video generation with diffusion models. arXiv preprint arXiv:2210.02303, 2022
-
[6]
Wan: Open and advanced large-scale video generative models
Team Wan, Ang Wang, Baole Ai, Bin Wen, Chaojie Mao, Chen-Wei Xie, Di Chen, Feiwu Yu, Haiming Zhao, Jianxiao Yang, Jianyuan Zeng, Jiayu Wang, Jingfeng Zhang, Jingren Zhou, Jinkai Wang, Jixuan Chen, Kai Zhu, Kang Zhao, Keyu Yan, Lianghua Huang, Mengyang Feng, Ningyi Zhang, Pandeng Li, Pingyu Wu, Ruihang Chu, Ruili Feng, Shiwei Zhang, Siyang Sun, Tao Fang, T...
arXiv 2025
-
[7]
Sdxl-lightning: Progressive adversarial diffusion distillation
Shanchuan Lin, Anran Wang, and Xiao Yang. Sdxl-lightning: Progressive adversarial diffusion distillation. arXiv preprint arXiv:2402.13929, 2024
arXiv 2024
-
[8]
Nitrofusion: High-fidelity single- step diffusion through dynamic adversarial training
Dar-Yen Chen, Hmrishav Bandyopadhyay, Kai Zou, and Yi-Zhe Song. Nitrofusion: High-fidelity single- step diffusion through dynamic adversarial training. arXiv preprint arXiv:2412.02030, 2024
arXiv 2024
Show all 87 references
-
[9]
Classifier-free diffusion guidance
Jonathan Ho and Tim Salimans. Classifier-free diffusion guidance. arXiv preprint arXiv:2207.12598 , 2022
2022 arXiv
-
[10]
Eliminating oversaturation and artifacts of high guidance scales in diffusion models
Seyedmorteza Sadat, Otmar Hilliges, and Romann M Weber. Eliminating oversaturation and artifacts of high guidance scales in diffusion models. In ICLR, 2024
2024
-
[11]
Characteristic guidance: Non-linear correction for diffusion model at large guidance scale
Candi Zheng and Yuan Lan. Characteristic guidance: Non-linear correction for diffusion model at large guidance scale. arXiv preprint arXiv:2312.07586, 2023
2023 arXiv
-
[12]
Rectified diffusion guidance for conditional generation
Mengfei Xia, Nan Xue, Yujun Shen, Ran Yi, Tieliang Gong, and Yong-Jin Liu. Rectified diffusion guidance for conditional generation. arXiv preprint arXiv:2410.18737, 2024
2024
-
[13]
Snoopi: Supercharged one-step diffusion distillation with proper guidance
Viet Nguyen, Anh Nguyen, Trung Dao, Khoi Nguyen, Cuong Pham, Toan Tran, and Anh Tran. Snoopi: Supercharged one-step diffusion distillation with proper guidance. arXiv preprint arXiv:2412.02687, 2024
2024
-
[14]
Scalable diffusion models with transformers
William Peebles and Saining Xie. Scalable diffusion models with transformers. In ICCV, 2023
2023
-
[15]
Scaling rectified flow transformers for high-resolution image synthesis
Patrick Esser, Sumith Kulal, Andreas Blattmann, Rahim Entezari, Jonas Müller, Harry Saini, Yam Levi, Dominik Lorenz, Axel Sauer, Frederic Boesel, et al. Scaling rectified flow transformers for high-resolution image synthesis. In ICML, 2024
2024
-
[16]
Learning transferable visual models from natural language supervision
Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, et al. Learning transferable visual models from natural language supervision. In ICML, 2021
2021
-
[17]
Gans trained by a two time-scale update rule converge to a local nash equilibrium
Martin Heusel, Hubert Ramsauer, Thomas Unterthiner, Bernhard Nessler, and Sepp Hochreiter. Gans trained by a two time-scale update rule converge to a local nash equilibrium. In NeurIPS, 2017
2017
-
[18]
Imagereward: Learning and evaluating human preferences for text-to-image generation
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. In NeurIPS, 2024
2024
-
[19]
Diffusion models beat gans on image synthesis
Prafulla Dhariwal and Alexander Nichol. Diffusion models beat gans on image synthesis. In NeurIPS, 2021
2021
-
[20]
Hierarchical text-conditional image generation with clip latents
Aditya Ramesh, Prafulla Dhariwal, Alex Nichol, Casey Chu, and Mark Chen. Hierarchical text-conditional image generation with clip latents. arXiv preprint arXiv:2204.06125, 2022
2022 arXiv
-
[21]
ediff-i: Text-to-image diffusion models with an ensemble of expert denoisers
Yogesh Balaji, Seungjun Nah, Xun Huang, Arash Vahdat, Jiaming Song, Qinsheng Zhang, Karsten Kreis, Miika Aittala, Timo Aila, Samuli Laine, Bryan Catanzaro, Tero Karras, and Ming-Yu Liu. ediff-i: Text-to-image diffusion models with an ensemble of expert denoisers. arXiv preprin...
2022 arXiv
-
[22]
Make-a-video: Text-to-video generation without text-video data
Uriel Singer, Adam Polyak, Thomas Hayes, Xi Yin, Jie An, Songyang Zhang, Qiyuan Hu, Harry Yang, Oron Ashual, Oran Gafni, Devi Parikh, Sonal Gupta, and Yaniv Taigman. Make-a-video: Text-to-video generation without text-video data. arXiv preprint arXiv:2209.14792, 2022. 10
2022 arXiv
-
[23]
Structure and content-guided video synthesis with diffusion models
Patrick Esser, Johnathan Chiu, Parmida Atighehchian, Jonathan Granskog, and Anastasis Germanidis. Structure and content-guided video synthesis with diffusion models. In ICCV, 2023
2023
-
[24]
Align your latents: High-resolution video synthesis with latent diffusion models
Andreas Blattmann, Robin Rombach, Huan Ling, Tim Dockhorn, Seung Wook Kim, Sanja Fidler, and Karsten Kreis. Align your latents: High-resolution video synthesis with latent diffusion models. In CVPR, 2023
2023
-
[25]
Photorealistic video generation with diffusion models
Agrim Gupta, Lijun Yu, Kihyuk Sohn, Xiuye Gu, Meera Hahn, Li Fei-Fei, Irfan Essa, Lu Jiang, and José Lezama. Photorealistic video generation with diffusion models. In ECCV, 2024
2024
-
[26]
Score-based generative modeling through stochastic differential equations
Yang Song, Jascha Sohl-Dickstein, Diederik P Kingma, Abhishek Kumar, Stefano Ermon, and Ben Poole. Score-based generative modeling through stochastic differential equations. In ICLR, 2021
2021
-
[27]
Elucidating the design space of diffusion-based generative models
Tero Karras, Miika Aittala, Timo Aila, and Samuli Laine. Elucidating the design space of diffusion-based generative models. In NeurIPS, 2022
2022
-
[28]
Building normalizing flows with stochastic interpolants
Michael Samuel Albergo and Eric Vanden-Eijnden. Building normalizing flows with stochastic interpolants. In ICLR, 2023
2023
-
[29]
Flow straight and fast: Learning to generate and transfer data with rectified flow
Xingchao Liu, Chengyue Gong, and Qiang Liu. Flow straight and fast: Learning to generate and transfer data with rectified flow. In ICLR, 2023
2023
-
[30]
Albergo, Nicholas M
Nanye Ma, Mark Goldstein, Michael S. Albergo, Nicholas M. Boffi, Eric Vanden-Eijnden, and Saining Xie. Sit: Exploring flow and diffusion-based generative models with scalable interpolant transformers. In ECCV, 2024
2024
-
[31]
Improving and generalizing flow-based generative models with minibatch optimal transport
Alexander Tong, Kilian FATRAS, Nikolay Malkin, Guillaume Huguet, Yanlei Zhang, Jarrid Rector-Brooks, Guy Wolf, and Yoshua Bengio. Improving and generalizing flow-based generative models with minibatch optimal transport. TMLR, 2024
2024
-
[32]
Aram-Alexandre Pooladian, Heli Ben-Hamu, Carles Domingo-Enrich, Brandon Amos, Yaron Lipman, and Ricky T. Q. Chen. Multisample flow matching: Straightening flows with minibatch couplings. arXiv preprint arXiv:2304.14772, 2023
2023 arXiv
-
[33]
Flow matching in latent space
Quan Dao, Hao Phung, Binh Nguyen, and Anh Tran. Flow matching in latent space. arXiv preprint arXiv:2307.08698, 2023
2023 arXiv
-
[34]
Instaflow: One step is enough for high-quality diffusion-based text-to-image generation
Xingchao Liu, Xiwen Zhang, Jianzhu Ma, Jian Peng, and qiang liu. Instaflow: One step is enough for high-quality diffusion-based text-to-image generation. In ICLR, 2024
2024
-
[35]
Yaron Lipman, Ricky T. Q. Chen, Heli Ben-Hamu, Maximilian Nickel, and Matthew Le. Flow matching for generative modeling. In ICLR, 2023
2023
-
[36]
Pixart-$\alpha$: Fast training of diffusion transformer for photorealistic text-to-image synthesis
Junsong Chen, Jincheng YU, Chongjian GE, Lewei Yao, Enze Xie, Zhongdao Wang, James Kwok, Ping Luo, Huchuan Lu, and Zhenguo Li. Pixart-$\alpha$: Fast training of diffusion transformer for photorealistic text-to-image synthesis. In ICLR, 2024
2024
-
[37]
Pixart- \sigma: Weak-to-strong training of diffusion transformer for 4k text-to-image generation
Junsong Chen, Chongjian Ge, Enze Xie, Yue Wu, Lewei Yao, Xiaozhe Ren, Zhongdao Wang, Ping Luo, Huchuan Lu, and Zhenguo Li. Pixart- \sigma: Weak-to-strong training of diffusion transformer for 4k text-to-image generation. arXiv preprint arXiv:2403.04692, 2024
2024 arXiv
-
[38]
Playground v3: Improving text-to-image alignment with deep-fusion large language models
Bingchen Liu, Ehsan Akhgari, Alexander Visheratin, Aleks Kamko, Linmiao Xu, Shivam Shrirao, Joao Souza, Suhail Doshi, and Daiqing Li. Playground v3: Improving text-to-image alignment with deep-fusion large language models. arXiv preprint arXiv:2409.10695, 2024
2024 arXiv
-
[39]
Lumina-next : Making lumina-t2x stronger and faster with next-dit
Le Zhuo, Ruoyi Du, Han Xiao, Yangguang Li, Dongyang Liu, Rongjie Huang, Wenze Liu, Xiangyang Zhu, Fu-Yun Wang, Zhanyu Ma, Xu Luo, Zehan Wang, Kaipeng Zhang, Lirui Zhao, Si Liu, Xiangyu Yue, Wanli Ouyang, Yu Qiao, Hongsheng Li, and Peng Gao. Lumina-next : Making lumina-t2x stro...
2024
-
[40]
Lumina-video: Efficient and flexible video generation with multi-scale Next-DiT
Dongyang Liu, Shicheng Li, Yutong Liu, Zhen Li, Kai Wang, Xinyue Li, Qi Qin, Yufei Liu, Yi Xin, Zhongyu Li, Bin Fu, Chenyang Si, Yuewen Cao, Conghui He, Ziwei Liu, Yu Qiao, Qibin Hou, Hongsheng Li, and Peng Gao. Lumina-video: Efficient and flexible video generation with multi-...
2025 arXiv
-
[41]
Sana 1.5: Efficient scaling of training-time and inference-time compute in linear diffusion transformer
Enze Xie, Junsong Chen, Yuyang Zhao, Jincheng Yu, Ligeng Zhu, Chengyue Wu, Yujun Lin, Zhekai Zhang, Muyang Li, Junyu Chen, Han Cai, Bingchen Liu, Daquan Zhou, and Song Han. Sana 1.5: Efficient scaling of training-time and inference-time compute in linear diffusion transformer....
2025 arXiv
-
[42]
Scaling vision transformers
Xiaohua Zhai, Alexander Kolesnikov, Neil Houlsby, and Lucas Beyer. Scaling vision transformers. In CVPR, 2022
2022
-
[43]
Scaling laws for diffusion transformers
Zhengyang Liang, Hao He, Ceyuan Yang, and Bo Dai. Scaling laws for diffusion transformers. arXiv preprint arXiv:2410.08184, 2024
2024
-
[44]
Progressive distillation for fast sampling of diffusion models
Tim Salimans and Jonathan Ho. Progressive distillation for fast sampling of diffusion models. In ICLR, 2022. 11
2022
-
[45]
Adversarial diffusion distillation
Axel Sauer, Dominik Lorenz, Andreas Blattmann, and Robin Rombach. Adversarial diffusion distillation. In ECCV, 2024
2024
-
[46]
Latent consistency models: Synthesizing high-resolution images with few-step inference
Simian Luo, Yiqin Tan, Longbo Huang, Jian Li, and Hang Zhao. Latent consistency models: Synthesizing high-resolution images with few-step inference. arXiv preprint arXiv:2310.04378, 2023
2023 arXiv
-
[47]
Trajectory consistency distillation
Jianbin Zheng, Minghui Hu, Zhongyi Fan, Chaoyue Wang, Changxing Ding, Dacheng Tao, and Tat-Jen Cham. Trajectory consistency distillation. arXiv preprint arXiv:2402.19159, 2024
2024 arXiv
-
[48]
Hyper-SD: Trajectory segmented consistency model for efficient image synthesis
Yuxi Ren, Xin Xia, Yanzuo Lu, Jiacheng Zhang, Jie Wu, Pan Xie, XING WANG, and Xuefeng Xiao. Hyper-SD: Trajectory segmented consistency model for efficient image synthesis. In NeurIPS, 2024
2024
-
[49]
One-step diffusion with distribution matching distillation
Tianwei Yin, Michaël Gharbi, Richard Zhang, Eli Shechtman, Frédo Durand, William T Freeman, and Taesung Park. One-step diffusion with distribution matching distillation. In CVPR, 2024
2024
-
[50]
Improved distribution matching distillation for fast image synthesis
Tianwei Yin, Michaël Gharbi, Taesung Park, Richard Zhang, Eli Shechtman, Fredo Durand, and William T Freeman. Improved distribution matching distillation for fast image synthesis. In NeurIPS, 2024
2024
-
[51]
Swiftbrush v2: Make your one-step diffusion model better than its teacher
Trung Dao, Thuan Hoang Nguyen, Thanh Le, Duc Vu, Khoi Nguyen, Cuong Pham, and Anh Tran. Swiftbrush v2: Make your one-step diffusion model better than its teacher. In ECCV, 2024
2024
-
[52]
Fast high-resolution image synthesis with latent adversarial diffusion distillation
Axel Sauer, Frederic Boesel, Tim Dockhorn, Andreas Blattmann, Patrick Esser, and Robin Rom- bach. Fast high-resolution image synthesis with latent adversarial diffusion distillation. arXiv preprint arXiv:2403.12015, 2024
2024 arXiv
-
[53]
Sana-sprint: One-step diffusion with continuous-time consistency distillation
Junsong Chen, Shuchen Xue, Yuyang Zhao, Jincheng Yu, Sayak Paul, Junyu Chen, Han Cai, Enze Xie, and Song Han. Sana-sprint: One-step diffusion with continuous-time consistency distillation. arXiv preprint arXiv: 2503.09641, 2025
2025
-
[54]
Diffusion adversarial post-training for one-step video generation
Shanchuan Lin, Xin Xia, Yuxi Ren, Ceyuan Yang, Xuefeng Xiao, and Lu Jiang. Diffusion adversarial post-training for one-step video generation. arXiv preprint arXiv:2501.08316, 2025
2025
-
[55]
Analysis of classifier-free guidance weight schedulers
WANG Xi, Nicolas Dufour, Nefeli Andreou, CANI Marie-Paule, Victoria Fernandez Abrevaya, David Picard, and Vicky Kalogeiton. Analysis of classifier-free guidance weight schedulers. TMLR, 2024
2024
-
[56]
CFG++: Manifold- constrained classifier free guidance for diffusion models
Hyungjin Chung, Jeongsol Kim, Geon Yeong Park, Hyelin Nam, and Jong Chul Ye. CFG++: Manifold- constrained classifier free guidance for diffusion models. In ICLR, 2025
2025
-
[57]
Yeh, and Ziwei Liu
Weichen Fan, Amber Yijia Zheng, Raymond A. Yeh, and Ziwei Liu. Cfg-zero*: Improved classifier-free guidance for flow matching models. arXiv preprint arXiv:2503.18886, 2025
2025 arXiv
-
[58]
Improving sample quality of diffusion models using self-attention guidance
Susung Hong, Gyuseong Lee, Wooseok Jang, and Seungryong Kim. Improving sample quality of diffusion models using self-attention guidance. In ICCV, 2023
2023
-
[59]
Self-rectifying diffusion sampling with perturbed-attention guidance
Donghoon Ahn, Hyoungwon Cho, Jaewon Min, Wooseok Jang, Jungwoo Kim, SeonHwa Kim, Hyun Hee Park, Kyong Hwan Jin, and Seungryong Kim. Self-rectifying diffusion sampling with perturbed-attention guidance. In ECCV, 2024
2024
-
[60]
Microsoft coco: Common objects in context
Tsung-Yi Lin, Michael Maire, Serge Belongie, James Hays, Pietro Perona, Deva Ramanan, Piotr Dollár, and C Lawrence Zitnick. Microsoft coco: Common objects in context. In ECCV, 2014
2014
-
[61]
Sdxl: Improving latent diffusion models for high-resolution image synthesis
Dustin Podell, Zion English, Kyle Lacey, Andreas Blattmann, Tim Dockhorn, Jonas Müller, Joe Penna, and Robin Rombach. Sdxl: Improving latent diffusion models for high-resolution image synthesis. arXiv preprint arXiv:2307.01952, 2023
2023 arXiv
-
[62]
SANA: Efficient high-resolution text-to-image synthesis with linear diffusion transformers
Enze Xie, Junsong Chen, Junyu Chen, Han Cai, Haotian Tang, Yujun Lin, Zhekai Zhang, Muyang Li, Ligeng Zhu, Yao Lu, and Song Han. SANA: Efficient high-resolution text-to-image synthesis with linear diffusion transformers. In ICLR, 2025
2025
-
[63]
Flux.1-turbo-alpha
Alimama Smart Creative and AI Application Team. Flux.1-turbo-alpha. https://huggingface.co/ alimama-creative/FLUX.1-Turbo-Alpha , 2024
2024
-
[64]
Playground v2.5: Three insights towards enhancing aesthetic quality in text-to-image generation
Daiqing Li, Aleks Kamko, Ehsan Akhgari, Ali Sabet, Linmiao Xu, and Suhail Doshi. Playground v2.5: Three insights towards enhancing aesthetic quality in text-to-image generation. arXiv preprint arXiv:2402.17245, 2024
2024 arXiv
-
[65]
NAG” and the small words “Normalized Attention Guidance: Universal Negative Guidance for Diffusion Models
Yuval Kirstain, Adam Polyak, Uriel Singer, Shahbuland Matiana, Joe Penna, and Omer Levy. Pick-a-pic: An open dataset of user preferences for text-to-image generation. In NeurIPS, 2023. 12 Table of Contents 1 Introduction 2 2 Related Works 3 3 Background 3 3.1 Text-to-Image Dif...
2023
-
[66]
An anthropomorphic cat thoughtfully paints an oil self-portrait on canvas, capturing its likeness with delicate brushstrokes inside a warmly lit, artistically cluttered studio
-
[67]
The fox is made of polygons
An origami fox running in the forest. The fox is made of polygons. Speed and passion. Realistic. Figure 21 and Figure 22: First row:
-
[68]
A cabinet in which all the planets of the solar system are collected
-
[69]
Cyborg bird
a futuristic interpretation of a dodo bird. Cyborg bird. Amazing colorful. Artstation, hyperrealistic
-
[70]
a happy female wizard surrounded by pieces of paper flying in the air around her
-
[71]
a recruitment consultant, sitting before a screen full of analysis diagram, carrying mobile device, fuji film style, like moss in wandering earth Second row:
-
[72]
Large birthday cake for a cardiothoracic surgeon
-
[73]
an anthropomorphic white rabbit, male wizard face, dressed in black and white, fine art, award-winning, intricate, elegant, sharp focus, cinematic lighting, highly detailed, digital painting, 8 k concept art, art by guweiz and z. w. gu, masterpiece, trending on artstation, 8 k
-
[74]
an owl transforms into an eagle
-
[75]
a photorealistic dragon pup Third row:
-
[76]
in a room a MGb car smashing through hole in the wall ,sparks dust rubble bricks ,studio lighting,white walls, mg logo
-
[77]
a werewolf reading a book
-
[78]
Black and white 1905 year portrait of futuristic professional photographer with camera in hand sadly seating deep in a dark pit covered by splash of dust
1905
-
[79]
a panda riding a motorcycle Fourth row:
-
[80]
a wide angle photo of roman soldiers in front of courtyard roman buildings,technicolor film ,roman soldier in foreground masculine features nose helmet and silver sword ,eyes,clear sky, arches grass steps field panorama,Canaletto,stone floor,vanishing point,ben-hur flags , a d...
-
[81]
cuddly stuffed dinosaur talking to a microphone
-
[82]
blue apple, red banana
-
[83]
Anime cute little girl Fifth row:
-
[84]
a lonely man inside a old bucolic house surrounded by cats by Richard Billingham
-
[85]
full body space suit with boots, futuristic, character design, cinematic lightning, epic fantasy, hyper realistic, detail 8k 26
-
[86]
a young danish traveller standing at an immigration counter in ho chi minh city
-
[87]
Amigurumi figure of a little pig wearing a red sweater, professional photography, close up, vintage, 8k, product photo J.2 Prompts for User Study Table 10: Positive-negative prompt pairs used in the user study comparing NAG and NASA. Positive prompt Negative prompt A photo of ...
Reviewed August 7, 2026 · model on record in the stance chip above.
Discussion (0). Sign in to comment.