Pith. sign in

REVIEW 4 major objections 6 minor 1 cited by

CountDiffusion: Text-to-Image Synthesis with Training-Free Counting-Guidance Diffusion

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 This paper claims that a training-free detection-and-correction loop can attach to any diffusion text-to-image model and sharply improve how often generated images contain the requested number of objects.

desk verdict Clever training-free counting-correction idea, but the headline numbers are circular—the same detector drives the correction and defines the metric—so the claimed gains are unverified. read the letter →

arxiv 2505.04347 v1 pith:SY7WP5CQ submitted 2025-05-07 cs.CV

Yanyu Li , Pencheng Wan , Liang Han , Yaowei Wang , Liqiang Nie , Min Zhang This is my paper · ORCID
classification cs.CV
keywords text-to-imagesynthesisdiffusionmodelsobjectcountingtraining-freeguidanceattentionmapcontroluniversalquantitygroundedsegmentation
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

CountDiffusion is a training-free way to make diffusion-based text-to-image models produce the exact number of objects a prompt asks for. It first generates an early preview of the image with a single denoising step, counts objects in that preview with a grounded segmentation model, and then uses a universal-guidance loss on the model's attention maps to add missing objects or remove extras. The paper claims this two-stage correction can be attached to any diffusion text-to-image model without retraining and that it improves counting accuracy substantially, for instance raising SDXL from 34% to 59% and Pixart-Sigma from 40% to 60% on the CoCoCount dataset. A curious reader would care because getting object counts right is a known weakness of text-to-image systems, and this fix is model-agnostic and requires no user-supplied boxes or masks.

What carries the argument

The load-bearing object is the one-step denoising preview: from a saved latent $z_{t_{\text{mid}}}$ the method predicts $z_{0,\text{mid}} = P(z_{t_{\text{mid}}}, y, t)$, decodes it to an image, and lets Grounded SAM produce counts and masks. The correction step uses a universal-guidance loss on the cross-attention map $A_t$: with a mask $M$ from the detection stage, the loss is $\pm \frac{1}{P} \sum \mathrm{topk}(A_t \odot M, P)$, taking the average of the top $P\%$ attention values inside the region, and the sign determines whether objects are removed or added. The background is preserved by mixing the corrected latent with the saved original latent outside $M$. For multi-class prompts the organizing identity is $z_{\text{mid}} = \frac{1}{n} \sum_{i=1}^{n} z_{\text{mid},i}$, one single-class correction per class averaged at the intermediate step so the class-specific losses do not fight each other.

What would settle it

Take a fixed set of prompts with small, overlapping, or similarly shaped objects, run the detection stage, and compare the Grounded SAM counts and object positions in the one-step preview $x_{0,\text{mid}}$ against those in the final decoded image. If the preview and final image disagree on counts in a large share of cases where the final image is correct, then the documented correction mechanism is not what is fixing the counts; if they agree reliably, the counting signal is as claimed.

Watch

Extended reading notes

Core claim

The central claim is that the quantity signal needed for correction survives in a one-step denoising prediction, even though that prediction breaks the diffusion model's Markov-chain assumption and produces a blurry image. The paper argues that this preview still carries reliable object quantity and position information, so a counting tool can extract per-class counts and segmentation masks from it. The correction stage then rewrites the cross-attention map while denoising runs: it lowers attention inside a segmented region to remove objects, raises attention elsewhere to add objects, and smooths the attention map so the gradient lands on a continuous region rather than scattered pixels. For multi-class prompts, the method corrects each class separately and averages the resulting intermediate latents, avoiding competition between losses. The evidence reported is that this pipeline raises accurate-count generation from 34% to 59% with SDXL and from 40% to 60% with Pixart-Sigma on CoCoCount, with consistent improvements on the two newly constructed evaluation sets.

Load-bearing premise

The load-bearing premise is that the image obtained by one-step denoising from an intermediate latent faithfully represents the object counts and positions of the final image; the paper asserts this from examples but does not measure how often it actually holds.

Editorial extensions

If this is right

  • Any diffusion-based text-to-image model can acquire counting ability simply by attaching CountDiffusion, so the improvement compounds as the base model improves.
  • Both single-class and multi-class counting accuracy improve, because the per-class latent averaging strategy prevents the correction losses for different object classes from competing.
  • The number of universal-guidance steps is a tunable trade-off: too few steps under-correct, and too many hurt image quality, with the optimum sitting near 30 guidance steps of 40 total in the paper's ablations.
  • Accuracy is bounded by the base model's and the counting model's ability, so very large requested counts remain a failure mode that the paper explicitly leaves to future work.

Reading between the lines

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

  • A testable extension would be to reuse the same preview-count-adjust loop for other measurable properties, such as object color, size, orientation, or spatial arrangement, by swapping the detector and the loss.
  • The real stress point is the one-step preview: on small, overlapping, or heavily occluded objects the preview's count is likely to diverge from the final image's count, which would make corrections miss exactly the hard cases.
  • Because the correction is applied locally through attention, CountDiffusion should compose naturally with layout-guided methods: a prompt-derived box layout could restrict the candidate regions and reduce false-positive detections before counting.
  • The fixed midpoint, $t_{\text{mid}}=30$ of 40 steps, could become adaptive by monitoring the preview's count confidence and applying guidance only when the count signal is reliable.
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.

Referee Report

4 major / 6 minor

Summary. The manuscript proposes CountDiffusion, a training-free framework for improving object-count accuracy in text-to-image diffusion models. It operates in two stages: a detection stage that one-step denoises an intermediate latent to predict the final image, uses Grounded SAM to count objects and obtain masks, and a correction stage that modifies cross-attention maps with a Universal Guidance style loss to add or remove objects until the detected count matches the prompt. A multi-class strategy runs separate single-class corrections and averages the resulting latents. Experiments on CoCoCount and two new GPT-constructed datasets compare SDXL and Pixart-Sigma baselines and report large gains in accuracy and MAE, together with CLIP-score and ImageReward, plus ablations of loss strategies, guidance steps, Gaussian smoothing, and the multi-class strategy.

Significance. If validated, CountDiffusion would be a practically useful contribution: it is training-free, applicable to different diffusion backbones, and requires no human annotation at inference time. The proposed multi-loss universal guidance and the two new evaluation datasets are also potentially useful resources. The paper is clearly written and includes ablations for each design component. However, the quantitative claims are not credible as presented because the same Grounded SAM model that drives the correction loop is also used to compute the headline Acc. and MAE, making the evaluation circular. In addition, the one-step denoising proxy is asserted without quantitative verification, and key hyperparameters appear to be selected on the same CoCoCount test set that is later reported in the main comparison. As a result, the paper's central claim that CountDiffusion improves counting accuracy by a large margin is unverifiable from the evidence provided.

major comments (4)
  1. [§IV-A, §IV-B, §V-B, Eq. (10)] The evaluation is circular. Grounded SAM is used in the detection stage to count objects in the one-step predicted image and to supply the masks that drive the correction loss (Eqs. 12-14), and in Section V-B the same Grounded SAM is used to compute the Acc. and MAE that appear in Table I. The reported improvements (e.g., SDXL 34% to 59% on CoCoCount and Pixart-Sigma 40% to 60%) therefore partly measure the detector's self-consistency rather than counting accuracy by an independent standard. In the limiting case, the method could change images so that Grounded SAM's detections match the prompt while a human or a different detector would still disagree. An independent counting protocol—human counts, a different detector, or at least a report of Grounded SAM's per-class counting accuracy—is required before the central claim can be credited.
  2. [§IV-A, Eqs. (8)-(9), Fig. 3] The entire detection stage depends on the assertion that a one-step denoising prediction x_{0,mid} accurately preserves the object quantity and positions of the final synthesized image. The paper provides only qualitative examples in Fig. 3 and no quantitative comparison between Grounded SAM counts on x_{0,mid} and on the final image. If the proxy is unreliable, the correction masks and the add/remove decisions are computed from an image that does not reflect the final result. Please add a quantitative validation, such as count agreement and mask overlap between x_{0,mid} and the final x_0 over a sample of prompts.
  3. [Tables II and III vs. Table I] The hyperparameters of the method appear to be selected on the same CoCoCount dataset that is later used for the headline comparison. Table II selects P=50 for the top-k loss strategy and Table III selects the number of universal guidance steps and total steps using accuracy on CoCoCount, and Table I then reports CoCoCount accuracy for the final method. No held-out validation split or nested evaluation is described. This test-set fitting inflates the reported gains and should be fixed by reporting validation-selected hyperparameters or by using a separate validation set.
  4. [§IV-C, Eq. (15)] The multi-class correction strategy averages per-class corrected latents z_{mid,i} at the intermediate step. The paper provides no analysis of whether this averaged latent stays on the denoising manifold or whether the average of class-specific corrections introduces inter-class conflicts in the subsequent DDIM steps. The ablation in Fig. 8 is qualitative only. A quantitative ablation, including per-class accuracy, failure-mode examples, and a comparison against a jointly optimized multi-class loss, is needed to support the claim that the averaging strategy avoids competition among class losses.
minor comments (6)
  1. [§III, Eqs. (2) and (3)] The display equations for the denoising update and the one-step prediction are difficult to parse because of missing parentheses; please format them with unambiguous operator grouping.
  2. [§V-B and Table III] The experiments state T=40 and t_mid=30, but Table III varies universal guidance steps up to 40 with total steps 40; please clarify how the guidance-step count relates to t_mid and the interval on which the loss is applied.
  3. [Fig. 5] The caption says the red smiling face indicates a correct count and the green crying face indicates an error; this color convention is the opposite of typical usage and should be checked and clarified.
  4. [§V-B] All reported Acc., MAE, CLIP-score, and ImageReward values lack error bars or significance tests; given the observed differences between some methods are small, standard deviations over multiple runs or seeds should be reported.
  5. [References] Reference [37] contains a malformed URL ('https://https://github.com/...') and should be corrected.
  6. [§V-A] The paper states that two new datasets are built but does not provide a release link or a detailed description of the data collection protocol beyond the prompt counts; please include availability information for reproducibility.

Circularity Check

1 steps flagged · score 7.0 of 10

Reported counting gains are circular: Grounded SAM both drives the correction and defines Acc/MAE, so the headline numbers measure detector self-consistency rather than independent object-count accuracy.

  1. self definitional [Section IV-A Eq. (10) and Section V-B Evaluation Metrics]
    "Grounded SAM (SAM) is adopted to segment and recognize each object in the predefined classes in image x0,mid, with which we can obtain the quantity of each class of objects and their corresponding segmentation masks Mseg, i.e., Mseg =SAM(x0,mid,tags) ... We employ Grounded SAM to obtain the object quantities required for computing Acc. and MAE."

    The correction stage obtains the masks M from Grounded SAM detections and then optimizes the attention loss (Eq. 12) to add or remove objects according to those detections. The headline accuracy and MAE are computed using the same Grounded SAM counts. Thus the evaluation does not measure object quantity independently; it measures whether the generated image is self-consistent with the very detector used to guide generation. Attention changes that make Grounded SAM split or merge detections can inflate Acc/MAE without changing human-perceived counts, so the reported improvements (e.g., SDXL 34 to 59 on CoCoCount) are partly circular and cannot validate the central claim of 'correct object quantity' without an independent counting protocol or human study.

full rationale

The paper's central quantitative support for 'correct object quantity' is Acc. and MAE, and both are computed from Grounded SAM counts. The same Grounded SAM is used in the detection stage to produce the masks that drive the correction loss, making the guidance loop and the evaluation oracle identical. This is a self-definitional circularity: the notion of 'correct count' is operationalized by the same detector that the method is optimized against. Consequently, the large reported gains do not provide independent evidence that the generated images contain the requested number of objects as judged by humans or by an external counter. Other aspects of the paper, such as the multi-loss universal guidance and the dataset construction, are not circular by themselves, and the paper does not rely on self-citations for its core mechanism. However, because the primary quantitative claim reduces to self-consistency with the control detector, the circularity score is high.

Assumptions & free parameters 5 free parameters · 4 assumptions · 0 invented entities

The method relies on four explicit assumptions, none machine-checked or independently verified, plus five hyperparameters tuned on the evaluation set. No new physical entities are introduced.

free parameters (5)
  • topk P percentage = 50
    Selected as the best loss strategy on CoCoCount in Table II; P is a parameter of Eq. 12 and is tuned on the evaluation set.
  • intermediate denoising step t_mid = 30 (with total steps T=40)
    Section V-B sets t_mid=30; Table III explores the number of guidance steps, so this choice is data-dependent on the evaluation set.
  • universal guidance steps = varies; best around 30 to 35 in Table III
    Table III shows accuracy peaks for intermediate guidance-step counts, so the final count is selected based on CoCoCount performance.
  • control scale sigma = not reported
    Introduced in Eq. 13 as the control intensity for guidance; no value or schedule is given for the main experiments.
  • Gaussian smoothing kernel for attention map = not reported
    Section IV-B states that Gaussian smoothing greatly improves quality and correction accuracy, but the kernel size and standard deviation are not specified.
assumptions (4)
  • domain assumption A one-step denoising prediction from z_{t_mid} retains enough object quantity and position information to guide the final image.
    Section IV-A and Figure 3 assert this despite the acknowledged violation of the Markov assumption; the detection stage depends entirely on this transferability.
  • domain assumption Grounded SAM object counts are sufficiently accurate for both guidance and evaluation.
    Grounded SAM is used to build masks (Eq. 10) and to compute Acc and MAE (Section V-B); any systematic counting error biases both the correction and the reported accuracy.
  • domain assumption Increasing or decreasing top-K cross-attention values inside a mask causes objects to appear or disappear.
    The correction loss (Eq. 12) relies on this causal link between attention values and rendered objects, inherited from universal guidance [31] and BoxDiff [9].
  • ad hoc to paper Averaging per-class corrected latents at step t_mid avoids competition among class losses and yields valid multi-class guidance.
    Proposed in Eq. 15 with no derivation; the only support is the ablation shown in Figure 8, which does not explain why averaging removes loss competition.

how reviews work

0 comments
Cite this review

Pith. "Pith review of CountDiffusion: Text-to-Image Synthesis with Training-Free Counting-Guidance Diffusion." pith.science (2026). https://pith.science/paper/SY7WP5CQ

@misc{pith2026250504347,
  author       = {Pith},
  title        = {Pith review of: CountDiffusion: Text-to-Image Synthesis with Training-Free Counting-Guidance Diffusion},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/SY7WP5CQ}},
  note         = {Machine review of arXiv:2505.04347}
}
read the original abstract

Stable Diffusion has advanced text-to-image synthesis, but training models to generate images with accurate object quantity is still difficult due to the high computational cost and the challenge of teaching models the abstract concept of quantity. In this paper, we propose CountDiffusion, a training-free framework aiming at generating images with correct object quantity from textual descriptions. CountDiffusion consists of two stages. In the first stage, an intermediate denoising result is generated by the diffusion model to predict the final synthesized image with one-step denoising, and a counting model is used to count the number of objects in this image. In the second stage, a correction module is used to correct the object quantity by changing the attention map of the object with universal guidance. The proposed CountDiffusion can be plugged into any diffusion-based text-to-image (T2I) generation models without further training. Experiment results demonstrate the superiority of our proposed CountDiffusion, which improves the accurate object quantity generation ability of T2I models by a large margin.

Figures

Figures reproduced from arXiv: 2505.04347 by the authors.

Figure 1
Figure 1. Samples generated by SDXL, Pixart-Σ, Ranni, InstanceDiffusion, and CountDiffusion (Ours). Existing models struggle with generating images with correct objects counts. Even when provided with bounding boxes, the models may still generate target objects in the background. Furthermore, Ranni, which combines a LLM for T2I generation, is limited by the language model and is still highly likely to generate incorrect objec… view at source ↗
Figure 2
Figure 2. Pipeline of the proposed CountDiffusion. CountDiffusion consists of a detection stage and a correction stage. In the detection stage, an intermediate [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. The synthesized images from intermediate denoising results of [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (6 more)
Figure 4
Figure 4. Figure 4: Comparison between the proposed CountDiffusion and state of the [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]
Figure 5
Figure 5. Figure 5: Qualitative comparisons of all models. Our method successfully generates the correct number of objects, while other methods struggle in some or all of the examples. The red smiling face in the bottom right corner of the image indicates that the correct number of object…
Figure 6
Figure 6. Figure 6: Qualitative comparisons based on SDXL. Our method successfully [PITH_FULL_IMAGE:figures/full_fig_p006_6.png]
Figure 7
Figure 7. Figure 7: Qualitative comparisons based on pixart- [PITH_FULL_IMAGE:figures/full_fig_p006_7.png]
Figure 8
Figure 8. Figure 8: Ablation study of the proposed CountDiffusion. [PITH_FULL_IMAGE:figures/full_fig_p007_8.png]
Figure 9
Figure 9. Figure 9: The effect of Gaussian smooth on attention maps. with Gaussian [PITH_FULL_IMAGE:figures/full_fig_p007_9.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

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

  1. Transferability Between Understanding and Generation in Unified Multimodal Models

    cs.CV 2026-07 conditional novelty 6.0 of 10

    Cross-task capability transfer in UMMs is architecture-dependent and can be exploited by training understanding to improve generation while avoiding distribution shift.

Reference graph

Works this paper leans on

38 extracted references · 22 canonical work pages · cited by 1 Pith paper

  1. [1]

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

    Dustin Podell, Zion English, Kyle Lacey, Andreas Blattmann, Tim Dockhorn, Jonas M ¨uller, Joe Penna, and Robin Rombach, “Sdxl: Improving latent diffusion models for high-resolution image synthesis,” arXiv preprint arXiv:2307.01952 , 2023

    arXiv 2023

  2. [2]

    Midjourney,

    Leap Motion, “Midjourney,” Website, 2022, https://www.midjourney. com

    work page 2022

  3. [3]

    Dall-e: Creating images from text,

    Mr D Murahari Reddy, Mr Sk Masthan Basha, Mr M Chinnaiahgari Hari, and Mr N Penchalaiah, “Dall-e: Creating images from text,” UGC Care Group I Journal , vol. 8, no. 14, pp. 71–75, 2021

    work page 2021

  4. [4]

    OpenAI, “Dalle3,” Website, 2023, https://openai.com/index/dall-e-3/

    work page 2023

  5. [5]

    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,” Advances in neural information processing systems , vol. 35, pp. 36479–36494, 2022

    work page 2022

  6. [6]

    Ranni: Taming text-to-image diffusion for accurate instruction following,

    Yutong Feng, Biao Gong, Di Chen, Yujun Shen, Yu Liu, and Jingren Zhou, “Ranni: Taming text-to-image diffusion for accurate instruction following,” 2024

    work page 2024

  7. [7]

    Scenecomposer: Any-level semantic image synthesis,

    Yu Zeng, Zhe Lin, Jianming Zhang, Qing Liu, John Collomosse, Jason Kuen, and Vishal M Patel, “Scenecomposer: Any-level semantic image synthesis,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2023, pp. 22468–22478

    work page 2023

  8. [8]

    Dense text-to-image generation with attention modulation,

    Yunji Kim, Jiyoung Lee, Jin-Hwa Kim, Jung-Woo Ha, and Jun-Yan Zhu, “Dense text-to-image generation with attention modulation,” in Proceedings of the IEEE/CVF International Conference on Computer Vision, 2023, pp. 7701–7711

    2023

Show all 38 references
  1. [9]

    Boxdiff: Text-to-image synthesis with training-free box-constrained diffusion,

    Jinheng Xie, Yuexiang Li, Yawen Huang, Haozhe Liu, Wentian Zhang, Yefeng Zheng, and Mike Zheng Shou, “Boxdiff: Text-to-image synthesis with training-free box-constrained diffusion,” in Proceedings of the IEEE/CVF International Conference on Computer Vision , 2023, pp. 7452–7461

    2023

  2. [10]

    Layoutllm-t2i: Eliciting layout guidance from llm for text-to-image generation,

    Leigang Qu, Shengqiong Wu, Hao Fei, Liqiang Nie, and Tat-Seng Chua, “Layoutllm-t2i: Eliciting layout guidance from llm for text-to-image generation,” in Proceedings of the 31st ACM International Conference on Multimedia , 2023, pp. 643–654

    2023

  3. [11]

    Instancediffusion: Instance-level control for image generation,

    Xudong Wang, Trevor Darrell, Sai Saketh Rambhatla, Rohit Girdhar, and Ishan Misra, “Instancediffusion: Instance-level control for image generation,” arXiv preprint arXiv:2402.03290 , 2024

    2024 arXiv

  4. [12]

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

    Lvmin Zhang, Anyi Rao, and Maneesh Agrawala, “Adding conditional control to text-to-image diffusion models,” in Proceedings of the IEEE/CVF International Conference on Computer Vision , 2023, pp. 3836–3847

    2023

  5. [13]

    Detector guidance for multi-object text-to-image generation,

    Luping Liu, Zijian Zhang, Yi Ren, Rongjie Huang, Xiang Yin, and Zhou Zhao, “Detector guidance for multi-object text-to-image generation,” arXiv preprint arXiv:2306.02236 , 2023

    2023 arXiv

  6. [14]

    Adver- sarial supervision makes layout-to-image diffusion models thrive,

    Yumeng Li, Margret Keuper, Dan Zhang, and Anna Khoreva, “Adver- sarial supervision makes layout-to-image diffusion models thrive,” 2024

    2024

  7. [15]

    Grounded sam: Assembling open-world models for diverse visual tasks,

    Tianhe Ren, Shilong Liu, Ailing Zeng, Jing Lin, Kunchang Li, He Cao, Jiayu Chen, Xinyu Huang, Yukang Chen, Feng Yan, Zhaoyang Zeng, Hao Zhang, Feng Li, Jie Yang, Hongyang Li, Qing Jiang, and Lei Zhang, “Grounded sam: Assembling open-world models for diverse visual tasks,” 2024

    2024

  8. [16]

    Denoising diffusion prob- abilistic models,

    Jonathan Ho, Ajay Jain, and Pieter Abbeel, “Denoising diffusion prob- abilistic models,” Advances in neural information processing systems , vol. 33, pp. 6840–6851, 2020

    2020

  9. [17]

    High-resolution image synthesis with latent diffu- sion models,

    Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser, and Bj ¨orn Ommer, “High-resolution image synthesis with latent diffu- sion models,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , 2022, pp. 10684–10695

    2022

  10. [18]

    Ilvr: Conditioning method for denoising diffusion probabilistic models,

    Jooyoung Choi, Sungwon Kim, Yonghyun Jeong, Youngjune Gwon, and Sungroh Yoon, “Ilvr: Conditioning method for denoising diffusion probabilistic models,” arXiv preprint arXiv:2108.02938 , 2021

    2021 arXiv

  11. [19]

    Diffusion models beat gans on image synthesis,

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

    2021

  12. [20]

    Sdedit: Image synthesis and editing with stochastic differential equations,

    Chenlin Meng, Yang Song, Jiaming Song, Jiajun Wu, Jun-Yan Zhu, and Stefano Ermon, “Sdedit: Image synthesis and editing with stochastic differential equations,” arXiv preprint arXiv:2108.01073 , 2021

    2021 arXiv

  13. [21]

    Deep unsupervised learning using nonequilibrium thermody- namics,

    Jascha Sohl-Dickstein, Eric Weiss, Niru Maheswaranathan, and Surya Ganguli, “Deep unsupervised learning using nonequilibrium thermody- namics,” in International conference on machine learning . PMLR, 2015, pp. 2256–2265

    2015

  14. [22]

    Modelscope text-to-video technical report,

    Jiuniu Wang, Hangjie Yuan, Dayou Chen, Yingya Zhang, Xiang Wang, and Shiwei Zhang, “Modelscope text-to-video technical report,” 2023

    2023

  15. [23]

    Magictime: Time-lapse video generation models as metamorphic simulators,

    Shenghai Yuan, Jinfa Huang, Yujun Shi, Yongqi Xu, Ruijie Zhu, Bin Lin, Xinhua Cheng, Li Yuan, and Jiebo Luo, “Magictime: Time-lapse video generation models as metamorphic simulators,” 2024

    2024

  16. [24]

    Champ: Controllable and consistent human image animation with 3d parametric guidance,

    Shenhao Zhu, Junming Leo Chen, Zuozhuo Dai, Yinghui Xu, Xun Cao, Yao Yao, Hao Zhu, and Siyu Zhu, “Champ: Controllable and consistent human image animation with 3d parametric guidance,” 2024

    2024

  17. [25]

    Renderdiffusion: Image diffusion for 3d reconstruction, inpainting and generation,

    Titas Anciukevi ˇcius, Zexiang Xu, Matthew Fisher, Paul Henderson, Hakan Bilen, Niloy J Mitra, and Paul Guerrero, “Renderdiffusion: Image diffusion for 3d reconstruction, inpainting and generation,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recog...

    2023

  18. [26]

    Magic123: One image to high-quality 3d object generation using both 2d and 3d diffusion priors,

    Guocheng Qian, Jinjie Mai, Abdullah Hamdi, Jian Ren, Aliaksandr Siarohin, Bing Li, Hsin-Ying Lee, Ivan Skorokhodov, Peter Wonka, Sergey Tulyakov, et al., “Magic123: One image to high-quality 3d object generation using both 2d and 3d diffusion priors,” arXiv preprint arXiv:2306...

    2023 arXiv

  19. [27]

    Generative adversarial text to image synthesis,

    Scott Reed, Zeynep Akata, Xinchen Yan, Lajanugen Logeswaran, Bernt Schiele, and Honglak Lee, “Generative adversarial text to image synthesis,” in International conference on machine learning . PMLR, 2016, pp. 1060–1069

    2016

  20. [28]

    Learning what and where to draw,

    Scott E Reed, Zeynep Akata, Santosh Mohan, Samuel Tenka, Bernt Schiele, and Honglak Lee, “Learning what and where to draw,” Advances in neural information processing systems , vol. 29, 2016

    2016

  21. [29]

    More control for free! image synthesis with semantic diffusion guidance,

    Xihui Liu, Dong Huk Park, Samaneh Azadi, Gong Zhang, Arman Chopikyan, Yuxiao Hu, Humphrey Shi, Anna Rohrbach, and Trevor Darrell, “More control for free! image synthesis with semantic diffusion guidance,” in Proceedings of the IEEE/CVF Winter Conference on Applications of Comp...

    2023

  22. [30]

    Classifier-free diffusion guidance,

    Jonathan Ho and Tim Salimans, “Classifier-free diffusion guidance,” arXiv preprint arXiv:2207.12598 , 2022

    2022 arXiv

  23. [31]

    Universal guidance for diffusion models,

    Arpit Bansal, Hong-Min Chu, Avi Schwarzschild, Soumyadip Sengupta, Micah Goldblum, Jonas Geiping, and Tom Goldstein, “Universal guidance for diffusion models,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2023, pp. 843– 852

    2023

  24. [32]

    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 Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , 2023, pp. 10146–10156

    2023

  25. [33]

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

    Amir Hertz, Ron Mokady, Jay Tenenbaum, Kfir Aberman, Yael Pritch, and Daniel Cohen-Or, “Prompt-to-prompt image editing with cross attention control,” arXiv preprint arXiv:2208.01626 , 2022

    2022 arXiv

  26. [34]

    Make it count: Text-to-image generation with an accurate number of objects,

    Lital Binyamin, Yoad Tewel, Hilit Segev, Eran Hirsch, Royi Rassin, and Gal Chechik, “Make it count: Text-to-image generation with an accurate number of objects,” arXiv preprint arXiv:2406.10210 , 2024

    2024 arXiv

  27. [35]

    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

  28. [36]

    Attention is all you need,

    Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, Łukasz Kaiser, and Illia Polosukhin, “Attention is all you need,” Advances in neural information processing systems , vol. 30, 2017

    2017

  29. [37]

    clip-vit-large-patch14,

    Suraj Patil, “clip-vit-large-patch14,” Website, 2021, https://https://github. com/a736875071/clip-vit-large-patch14

    2021

  30. [38]

    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,” 2023

    2023

Pith tools

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