Pith. sign in

REVIEW 4 major objections 5 minor 44 references

InfGen: A Resolution-Agnostic Paradigm for Scalable Image Synthesis

T0 review · 4 major / 5 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 InfGen claims that arbitrary high-resolution images can be generated from a fixed-size diffusion latent by replacing only the VAE decoder with a one-step resolution-agnostic generator.

desk verdict A genuinely new decoder-side idea for arbitrary-resolution LDM generation, backed by broad but patch-level-only validation; the 4K latency headline is unmeasured and architecture details are thin. read the letter →

arxiv 2509.10441 v1 pith:6K5RVYDQ submitted 2025-09-12 cs.CV

Tao Han , Wanghan Xu , Junchao Gong , Xiaoyu Yue , Song Guo , Luping Zhou , Lei Bai This is my paper · ORCID
classification cs.CV
keywords arbitraryresolutionimagegenerationlatentdiffusionmodelsone-stepdecoderresolution-agnosticVAEreplacementimplicitneuralpositionalembeddingtraining-freeextrapolationplug-and-playupgrade
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

InfGen claims that once a diffusion model has produced its compact latent, that latent already contains the image's content, so a single trained decoder can turn it into an image at any requested resolution, without retraining the diffusion model and without high-resolution denoising. The paper replaces the VAE decoder of latent diffusion models with a transformer-based generator that uses the fixed latent as keys and values and a resolution-shaped mask token as query, trained on real high-resolution images with reconstruction, perceptual, and adversarial losses. If the claim holds, any diffusion model sharing the same latent space can be upgraded to arbitrary high-resolution generation by swapping in InfGen, with 4K images generated in about 7.4 seconds instead of over 100 seconds. The paper reports that InfGen improves FID-based quality metrics at high resolutions on all tested backbones and is faster than the previous fastest arbitrary-resolution method.

What carries the argument

The central object is InfGen's arbitrary-resolution decoder, a vision transformer that treats the fixed latent $z$ as keys and values and a mask token of shape $(\lceil h/8\rceil, \lceil w/8\rceil)$ as query, with cross-attention in transformer blocks followed by upsampling to the target size. Positional information across dynamic sizes comes from an Implicit Neural Positional Embedding (INPE) that normalizes coordinates, maps them to a unit sphere, applies Fourier features, and feeds them through a small implicit neural network to produce positional encodings for queries and keys. Training uses $ℓ_1$ reconstruction loss, LPIPS perceptual loss, and a PatchGAN adversarial loss on high-resolution crops. For resolutions beyond training, a training-free iterative extrapolation scheme decodes, re-encodes, and decodes again with bounded scale factors, reaching 4K and beyond.

What would settle it

Take a latent from a diffusion model fine-tuned on a domain far from the training data, such as medical or satellite imagery, decode it with InfGen at 2048x2048, and compare semantic fidelity against the same model's native 512x512 output; if the high-resolution output introduces structures absent from or inconsistent with the low-resolution content, the 'fixed latent is complete content' assumption fails. A quantitative version is to encode the 3072x3072 output back into the latent space and measure its divergence from the original fixed latent, where large divergence indicates the decoder changed content rather than only adding detail.

Watch

Extended reading notes

Core claim

The central discovery is that the second stage of latent diffusion, the mapping from latent to pixels, can absorb all the cost of resolution while the first stage only needs to generate a fixed, compact content latent. InfGen trains a one-step decoder, conditioned on the fixed-size latent through cross-attention, to output an image of arbitrary height and width; the decoder is trained on crops of high-resolution images at varying target resolutions, forcing it to invent plausible texture and detail beyond what the latent explicitly specifies. Because the latent is fixed and small, inference cost no longer grows with output resolution. The paper demonstrates this on several generative backbones sharing the VAE latent space and shows that replacing the VAE decoder improves quality at high resolutions compared with upsampling the original outputs, while also beating specialized arbitrary-resolution generators in speed.

Load-bearing premise

The load-bearing premise is that a fixed-size latent generated by a diffusion model contains enough content information that a decoder trained only on real image crops can invent coherent, semantically correct detail at any target resolution, and that latents from different diffusion models sharing the VAE encoder fall in a distribution this decoder can handle.

Editorial extensions

If this is right

  • Any diffusion model trained on the same VAE latent space can be upgraded to arbitrary-resolution generation by swapping in InfGen, with no retraining of the diffusion model.
  • Generating a 4K image drops from over 100 seconds to roughly 7.4 seconds, because the multi-step denoising runs only on a compact fixed-size latent.
  • High-resolution quality improves over naively upsampling the original model outputs, with the paper reporting FID-based improvements of 41 to 44 percent at 3072x3072 on some backbones.
  • InfGen also functions as a competitive image tokenizer, matching or exceeding standard VAE decoders on reconstruction metrics at native resolutions.
  • The training-free iterative extrapolation extends the same decoder to ultra-high resolutions beyond its training range, preserving quality up to recommended scale factors.

Reading between the lines

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

  • Beyond the paper's claims, the plug-and-play behavior implies a testable distributional assumption: the latents of SD, SDXL, DiT, SiT, and FiTv2 must be close enough in latent space for one decoder to handle them all, and measuring that distance would predict where the patch starts to fail.
  • Because the decoder is trained only on real image crops, the detail it invents is bounded by the training distribution; a diffusion model fine-tuned on a very different domain may produce latents whose high-resolution decoding hallucinates plausible but semantically wrong structures.
  • The iterative extrapolation loop re-encodes generated pixels, so errors can compound across cycles; tracking latent drift between the original fixed latent and the re-encoded latent after a 4K decode would show whether content is preserved or only invented.
  • The same split, a fixed content latent plus a one-step arbitrary-size decoder, could transfer to other generative modalities with a learned tokenizer, such as video, where compact latents would make high-resolution decoding computationally cheap.
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 / 5 minor

Summary. The paper proposes InfGen, a learned decoder that replaces the VAE decoder in latent diffusion models. InfGen takes a fixed-size latent produced by an off-the-shelf diffusion model and a target resolution (h, w), and generates an image of that resolution through a transformer decoder using cross-attention between latent tokens and learnable mask tokens. The decoder is trained on pairs (VAE-encoded resized crops, high-resolution targets) with L1, LPIPS, and adversarial losses. The authors report improved FIDp/sFIDp, precision, and recall across DiT, SiT, MDTv2, FiTv2, and SD1.5/SDXL at resolutions from 512 to 3072 pixels, alongside latency advantages over prior arbitrary-resolution methods, and claim that the method can upgrade any model sharing the SD/SDXL VAE latent space to arbitrary-resolution generation without retraining.

Significance. The central idea is practical and timely: if a fixed-size latent from an existing diffusion model can be decoded at arbitrary resolution by a one-step generator, then the large installed base of latent diffusion models can be upgraded to high-resolution synthesis at low computational cost. The paper provides a broad experimental study across five diffusion backbones and multiple resolutions, and the reported speed advantage is potentially substantial. However, the significance of the claim as stated is conditional on two missing pieces of evidence: a direct measurement of content/semantic preservation during decoding, and a concrete measurement of the headline 4K latency. These gaps are load-bearing because the method is explicitly positioned as a 'content representation' decoder and a 'plug-and-play' replacement for the VAE decoder. If the decoder turns out to be a locally realistic texture prior rather than a faithful content decoder, the paradigm's main practical value—seamless upgrade of existing models—would be undermined.

major comments (4)
  1. [Section 3.2.1 and Tables 2–4] The central assumption that a fixed-size latent is a complete content representation is not validated by any content-preservation metric. FIDp, sFIDp, precision, and recall are computed on 229×229 patches and measure local texture statistics; they cannot distinguish faithful decoding from a generic high-resolution texture prior that ignores part of the latent. The paper should report at least one content-fidelity experiment: for example, re-encoding consistency (encode the InfGen-decoded image back to latent and compare with the original latent), or class-conditional accuracy for DiT/SiT/MDTv2, or CLIPScore for text-to-image SD1.5/SDXL. This is load-bearing because the abstract and Section 3.2 promise 'decoding arbitrary resolution images from a compact generated latent' as a general plugin, and because the plug-and-play claim for any model sharing the latent space requires demonstrating that decoder does not introduce semantic distortion across different latent distributions.
  2. [Abstract and Section 5] The headline latency claim—4K image generation under 10 seconds in the abstract and 7.4 seconds in Section 5—is not supported by any reported measurement. Table 4 reports latencies only for 1024×1024 and 2048×2048, with no definition of '4K' (e.g., 3840×2160 vs. 4096×4096) and no row for 4K. The contribution statement 'improving generation speed by over ten times' is also not substantiated by the Table 4 numbers: at 2048×2048, InfGen+SDXL-B-1 is 5.4+1.9 seconds versus UltraPixel's 20 seconds, which is roughly 2.7×, not 10×. Please add a dedicated 4K latency table with hardware, base model, resolution, and a breakdown of diffusion time and InfGen decoding time, and reconcile the 'over ten times' claim with the data.
  3. [Section 3.3] The iterative extrapolation scheme encodes the previously generated image back to latent space and decodes again with a scaling factor. The paper does not analyze error accumulation across encode-decode cycles (e.g., drift in the re-encoded latent or amplification of artifacts), and Table 3 does not specify which entries use iterative extrapolation versus a single forward pass. Since the method is presented as a 'training-free resolution extrapolation' tool, the 3072×3072 results are ambiguous without this specification. An explicit comparison of iterative versus single-shot decoding at the same resolution, along with a study of the reliability limits claimed in Table 1, is necessary to support the extrapolation claims.
  4. [Sections 3.2.2 and 3.2.3] The architecture description omits parameters that are essential for reproducibility and for assessing the 'lightweight generator' claim: the number of transformer blocks, hidden widths, number of attention heads, patch size, the dimension of the Fourier feature vector in Eq. (7), and the values of μ and σ used to sample the diagonal matrix B. Without these details, a reader cannot implement the method, and the claim that InfGen is significantly cheaper than the diffusion backbone cannot be verified. Please provide a complete architecture table or specify the missing hyperparameters.
minor comments (5)
  1. [Table 3] The header 'extrapolation→image: 3072×3072' appears to have a formatting issue: '3072×3072' is repeated in two adjacent columns, and the legend for the second column is missing. Please clarify the column meanings.
  2. [Section 4.2] The notation 'InfGen+DiT' and 'InfGen+ SD1.5' is inconsistent (missing subscript or formatting); please make the method naming uniform throughout.
  3. [Table 4 and Section 4.3] The state-of-the-art comparison uses different base models for the InfGen rows (SD1.5 and SDXL-B-1) without a discussion of whether the comparison is fair given the base-model difference. The caption should state which base models are used by each competitor and by InfGen.
  4. [Section 4 (Dataset)] The training data description says 10 million images with resolutions exceeding 1024^2 and a further filtering to 5 million with resolutions over 2048^2. Please clarify whether the 5 million subset is contained in the 10 million set, and how the two-stage training uses these partitions.
  5. [Figure 2] The log-scale latency plot is hard to verify without a table; please include the numerical values or a supplementary table corresponding to the points in the figure.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: InfGen is a supervised decoder trained on external reconstruction pairs and evaluated against external baselines.

full rationale

Walked the claimed derivation chain: (1) treating the fixed latent as a content representation is an architectural premise, not a conclusion derived from the premise; (2) InfGen's generator is learned by supervised reconstruction from pairs (z = VAE_encoder(resize(crop(x_high),512)), x_high) with L1/LPIPS/GAN losses, so the mapping is fit to external image data rather than to the diffusion outputs it later decodes; (3) evaluation at inference uses latents sampled from pretrained DiT/SiT/SD/SDXL/FiTv2, which are not the training targets, and quality is benchmarked against external methods (ScaleCrafter, Inf-DiT, UltraPixel) using FIDp/sFIDp/Precision/Recall; (4) the only self-citations (refs 19 and 38) are background or application listings, not load-bearing uniqueness theorems or ansatz justifications. No equation in the paper reduces to its own input: Eq. (4) defines the mapping, Eq. (5) is a standard reconstruction objective, and Eq. (8) is an iterative extrapolation, none of which assumes the conclusion. A possible concern that patch-based FID does not verify semantic preservation is an evaluation-adequacy issue, not a circularity.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The ledger shows that InfGen contributes no new physical or conceptual entity; its load-bearing assumptions are about latent sufficiency, cross-model distribution compatibility, iterative stability, and evaluation validity. The free parameters are architectural and schedule choices, not scientific constants fitted to data, so the circularity burden is low.

free parameters (3)
  • INPE Fourier sampling N(mu, sigma) for B
    Equation (7) samples a diagonal matrix B from a Gaussian, but mu and sigma are never reported; this hand-selected hyperparameter defines the positional encoding family.
  • Loss weights lambda_P and lambda_G = 0.1 and 0.1
    Section 3.2.1 fixes both weights at 0.1 without ablation; the balance between reconstruction, perceptual, and adversarial terms shapes all results.
  • Iterative extrapolation scale factors k_s_n = 1 to 2
    Section 3.3 and Table 1 prescribe per-iteration scale factors between 1 and 2, and no analysis is given for choosing a particular sequence; this schedule controls output resolution and quality.
assumptions (4)
  • domain assumption A fixed-size VAE latent contains enough information for coherent arbitrary-resolution detail synthesis.
    Section 3.2 frames the diffusion model as generating "fixed-compact content latent z"; the decoder must invent all detail beyond what the latent encodes, which is the paper's central premise.
  • domain assumption Latents from SD, SDXL, DiT, SiT, and FiTv2 are compatible with a decoder trained on latents from resized real images.
    Section 3.2.1 claims plug-and-play use with "any model using the same latent space"; no distributional comparison between generated latents and training latents is provided.
  • domain assumption Iterative encode-decode extrapolation does not accumulate artifacts.
    Section 3.3 repeatedly applies InfGen to an encoded version of its own output; the paper calls this robust based on Table 3, but no drift or error analysis is shown.
  • domain assumption Patch-based FID with 229x229 crops is a valid high-resolution quality measure.
    Section 4 adopts UltraPixel's FIDp metric without independent validation; scores depend on crop sampling and may favor hallucinated texture over global coherence.

how reviews work

0 comments
Cite this review

Pith. "Pith review of InfGen: A Resolution-Agnostic Paradigm for Scalable Image Synthesis." pith.science (2026). https://pith.science/paper/6K5RVYDQ

@misc{pith2026250910441,
  author       = {Pith},
  title        = {Pith review of: InfGen: A Resolution-Agnostic Paradigm for Scalable Image Synthesis},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/6K5RVYDQ}},
  note         = {Machine review of arXiv:2509.10441}
}
read the original abstract

Arbitrary resolution image generation provides a consistent visual experience across devices, having extensive applications for producers and consumers. Current diffusion models increase computational demand quadratically with resolution, causing 4K image generation delays over 100 seconds. To solve this, we explore the second generation upon the latent diffusion models, where the fixed latent generated by diffusion models is regarded as the content representation and we propose to decode arbitrary resolution images with a compact generated latent using a one-step generator. Thus, we present the \textbf{InfGen}, replacing the VAE decoder with the new generator, for generating images at any resolution from a fixed-size latent without retraining the diffusion models, which simplifies the process, reducing computational complexity and can be applied to any model using the same latent space. Experiments show InfGen is capable of improving many models into the arbitrary high-resolution era while cutting 4K image generation time to under 10 seconds.

Figures

Figures reproduced from arXiv: 2509.10441 by the authors.

Figure 2
Figure 2. Inference time (seconds per image) for high-resolution image generation methods. The vertical axis is a logarithmic scale. sive models [1, 35]. Due to the high training costs and slow multi-step inference, most generative methods follow a two￾stage paradigm: First, the generative model produces an intermediate representation of the image, and then the de￾coder from a pre-trained tokenizer maps the intermediate repre… view at source ↗
Figure 3
Figure 3. Illustration of the training and inference processes. The generator is trained in the latent space to reconstruct images at any resolution and aspect ratio. During inference, it can be applied to improve various diffusion models, enabling them to generate images of arbitrary resolution across various tasks. and play: InfGen offers strong generalization as it can be applied to any diffusion model trained on the same … view at source ↗
Figure 4
Figure 4. Visualizations of arbitrary image generation. The proposed InfGen improves the generation ability for LDMs [29] across various resolutions. More visual examples are provided in the appendix. integrated with diverse base models. Quantitative comparison. As indicated in Tab. 4, our method InfGen+SD1.5 performs competitively in terms of FID, sFID, Precision, and Recall in two high resolu￾tions. Notably, InfGen has a si… view at source ↗

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

44 extracted references · 15 canonical work pages

  1. [1]

    Maskgit: Masked generative image transformer

    Huiwen Chang, Han Zhang, Lu Jiang, Ce Liu, and William T Freeman. Maskgit: Masked generative image transformer. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 11315–11325, 2022. 1, 2

    work page 2022

  2. [2]

    Latentpaint: Image inpainting in latent space with diffusion models

    Ciprian Corneanu, Raghudeep Gadde, and Aleix M Mar- tinez. Latentpaint: Image inpainting in latent space with diffusion models. InProceedings of the IEEE/CVF Win- ter Conference on Applications of Computer Vision, pages 4334–4343, 2024. 4

    work page 2024

  3. [3]

    Diffusion models beat gans on image synthesis.Advances in neural informa- tion processing systems, 34:8780–8794, 2021

    Prafulla Dhariwal and Alexander Nichol. Diffusion models beat gans on image synthesis.Advances in neural informa- tion processing systems, 34:8780–8794, 2021. 2

    2021

  4. [4]

    Image super-resolution using deep convolutional net- works.IEEE transactions on pattern analysis and machine intelligence, 38(2):295–307, 2015

    Chao Dong, Chen Change Loy, Kaiming He, and Xiaoou Tang. Image super-resolution using deep convolutional net- works.IEEE transactions on pattern analysis and machine intelligence, 38(2):295–307, 2015. 3

    work page 2015

  5. [5]

    An image is worth 16x16 words: Transformers for image recognition at scale.ICLR, 2021

    Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Syl- vain Gelly, Jakob Uszkoreit, and Neil Houlsby. An image is worth 16x16 words: Transformers for image recognition at scale.ICLR, 2021. 5

    work page 2021

  6. [6]

    Taming transformers for high-resolution image synthesis

    Patrick Esser, Robin Rombach, and Bjorn Ommer. Taming transformers for high-resolution image synthesis. InPro- ceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 12873–12883, 2021. 6

    work page 2021

  7. [7]

    Mdtv2: Masked diffusion transformer is a strong image synthesizer.arXiv preprint arXiv:2303.14389,

    Shanghua Gao, Pan Zhou, Ming-Ming Cheng, and Shuicheng Yan. Mdtv2: Masked diffusion transformer is a strong image synthesizer.arXiv preprint arXiv:2303.14389,

    arXiv

  8. [8]

    Scalecrafter: Tuning-free higher- resolution visual generation with diffusion models

    Yingqing He, Shaoshu Yang, Haoxin Chen, Xiaodong Cun, Menghan Xia, Yong Zhang, Xintao Wang, Ran He, Qifeng Chen, and Ying Shan. Scalecrafter: Tuning-free higher- resolution visual generation with diffusion models. InThe Twelfth International Conference on Learning Representa- tions, 2023. 2, 3, 7, 8

    work page 2023

Show all 44 references
  1. [9]

    Gans trained by a two time-scale update rule converge to a local nash equilib- rium.Advances in neural information processing systems, 30, 2017

    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 equilib- rium.Advances in neural information processing systems, 30, 2017. 6

    2017

  2. [10]

    Denoising dif- fusion probabilistic models.Advances in neural information processing systems, 33:6840–6851, 2020

    Jonathan Ho, Ajay Jain, and Pieter Abbeel. Denoising dif- fusion probabilistic models.Advances in neural information processing systems, 33:6840–6851, 2020. 2

    2020

  3. [11]

    Fouriscale: A frequency perspective on training-free high-resolution image synthesis.arXiv preprint arXiv:2403.12963, 2024

    Linjiang Huang, Rongyao Fang, Aiping Zhang, Guanglu Song, Si Liu, Yu Liu, and Hongsheng Li. Fouriscale: A frequency perspective on training-free high-resolution image synthesis.arXiv preprint arXiv:2403.12963, 2024. 2, 3

    2024 arXiv

  4. [12]

    Image-to-image translation with conditional adver- sarial networks

    Phillip Isola, Jun-Yan Zhu, Tinghui Zhou, and Alexei A Efros. Image-to-image translation with conditional adver- sarial networks. InProceedings of the IEEE conference on computer vision and pattern recognition, pages 1125–1134,

  5. [13]

    Training- free diffusion model adaptation for variable-sized text-to- image synthesis.Advances in Neural Information Processing Systems, 36:70847–70860, 2023

    Zhiyu Jin, Xuli Shen, Bin Li, and Xiangyang Xue. Training- free diffusion model adaptation for variable-sized text-to- image synthesis.Advances in Neural Information Processing Systems, 36:70847–70860, 2023. 3

    2023

  6. [14]

    Transformers in vision: A survey.ACM computing surveys (CSUR), 54(10s):1–41, 2022

    Salman Khan, Muzammal Naseer, Munawar Hayat, Syed Waqas Zamir, Fahad Shahbaz Khan, and Mubarak Shah. Transformers in vision: A survey.ACM computing surveys (CSUR), 54(10s):1–41, 2022. 5

    2022

  7. [15]

    Auto-encoding varia- tional bayes.In International Conference on Learning Rep- resentations, 2014

    Diederik P Kingma and Max Welling. Auto-encoding varia- tional bayes.In International Conference on Learning Rep- resentations, 2014. 2, 3

    2014

  8. [16]

    Improved precision and recall met- ric for assessing generative models.Advances in neural in- formation processing systems, 32, 2019

    Tuomas Kynk ¨a¨anniemi, Tero Karras, Samuli Laine, Jaakko Lehtinen, and Timo Aila. Improved precision and recall met- ric for assessing generative models.Advances in neural in- formation processing systems, 32, 2019. 6

    2019

  9. [17]

    Swinir: Image restoration us- ing swin transformer

    Jingyun Liang, Jiezhang Cao, Guolei Sun, Kai Zhang, Luc Van Gool, and Radu Timofte. Swinir: Image restoration us- ing swin transformer. InProceedings of the IEEE/CVF inter- national conference on computer vision, pages 1833–1844,

  10. [18]

    Decoupled weight decay regularization.arXiv preprint arXiv:1711.05101, 2017

    I Loshchilov. Decoupled weight decay regularization.arXiv preprint arXiv:1711.05101, 2017. 6

    2017 arXiv

  11. [19]

    Fit: Flexible vision trans- former for diffusion model.ICML, 2024

    Zeyu Lu, Zidong Wang, Di Huang, Chengyue Wu, Xihui Liu, Wanli Ouyang, and Lei Bai. Fit: Flexible vision trans- former for diffusion model.ICML, 2024. 3, 4, 6, 7

    2024

  12. [20]

    Sit: Explor- ing flow and diffusion-based generative models with scalable interpolant transformers.arXiv preprint arXiv:2401.08740,

    Nanye Ma, Mark Goldstein, Michael S Albergo, Nicholas M Boffi, Eric Vanden-Eijnden, and Saining Xie. Sit: Explor- ing flow and diffusion-based generative models with scalable interpolant transformers.arXiv preprint arXiv:2401.08740,

    arXiv

  13. [21]

    Generating images with sparse representations

    Charlie Nash, Jacob Menick, Sander Dieleman, and Peter W Battaglia. Generating images with sparse representations. arXiv preprint arXiv:2103.03841, 2021. 6

    2021 arXiv

  14. [22]

    Glide: Towards photorealistic image generation and editing with text-guided diffusion models.arXiv preprint arXiv:2112.10741, 2021

    Alex Nichol, Prafulla Dhariwal, Aditya Ramesh, Pranav Shyam, Pamela Mishkin, Bob McGrew, Ilya Sutskever, and Mark Chen. Glide: Towards photorealistic image generation and editing with text-guided diffusion models.arXiv preprint arXiv:2112.10741, 2021. 2

    2021 arXiv

  15. [23]

    Scalable diffusion models with transformers

    William Peebles and Saining Xie. Scalable diffusion models with transformers. InProceedings of the IEEE/CVF Inter- national Conference on Computer Vision, pages 4195–4205,

  16. [24]

    Sdxl: Improving latent diffusion mod- els for high-resolution image synthesis.arXiv preprint arXiv:2307.01952, 2023

    Dustin Podell, Zion English, Kyle Lacey, Andreas Blattmann, Tim Dockhorn, Jonas M ¨uller, Joe Penna, and Robin Rombach. Sdxl: Improving latent diffusion mod- els for high-resolution image synthesis.arXiv preprint arXiv:2307.01952, 2023. 1, 2, 3, 4, 6, 7

    2023 arXiv

  17. [25]

    Hierarchical text-conditional image gener- ation with clip latents.arXiv preprint arXiv:2204.06125, 1 (2):3, 2022

    Aditya Ramesh, Prafulla Dhariwal, Alex Nichol, Casey Chu, and Mark Chen. Hierarchical text-conditional image gener- ation with clip latents.arXiv preprint arXiv:2204.06125, 1 (2):3, 2022. 3

    2022 arXiv

  18. [26]

    Gener- ating diverse high-fidelity images with vq-vae-2.Advances in neural information processing systems, 32, 2019

    Ali Razavi, Aaron Van den Oord, and Oriol Vinyals. Gener- ating diverse high-fidelity images with vq-vae-2.Advances in neural information processing systems, 32, 2019. 3

    2019

  19. [27]

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

    Jingjing Ren, Wenbo Li, Haoyu Chen, Renjing Pei, Bin Shao, Yong Guo, Long Peng, Fenglong Song, and Lei Zhu. Ultrapixel: Advancing ultra-high-resolution image synthesis to new peaks. InNeurIPs, 2024. 2, 3, 6, 7

    2024

  20. [28]

    Stochastic backpropagation and approximate inference in deep generative models

    Danilo Jimenez Rezende, Shakir Mohamed, and Daan Wier- stra. Stochastic backpropagation and approximate inference in deep generative models. InInternational conference on machine learning, pages 1278–1286. PMLR, 2014. 2, 3

    2014

  21. [29]

    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. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 10684–10695, 2022. 1, 2, 3, 4, 6, 7, 8

    2022

  22. [30]

    Photorealistic text-to-image diffusion models with deep language understanding.Advances in neural information processing systems, 35:36479–36494, 2022

    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 ...

    2022

  23. [31]

    Laion-5b: An open large-scale dataset for training next generation image-text models.Advances in Neural In- formation Processing Systems, 35:25278–25294, 2022

    Christoph Schuhmann, Romain Beaumont, Richard Vencu, Cade Gordon, Ross Wightman, Mehdi Cherti, Theo Coombes, Aarush Katta, Clayton Mullis, Mitchell Worts- man, et al. Laion-5b: An open large-scale dataset for training next generation image-text models.Advances in Neural In- fo...

    2022

  24. [32]

    Implicit neural representa- tions with periodic activation functions.Advances in neural information processing systems, 33:7462–7473, 2020

    Vincent Sitzmann, Julien Martel, Alexander Bergman, David Lindell, and Gordon Wetzstein. Implicit neural representa- tions with periodic activation functions.Advances in neural information processing systems, 33:7462–7473, 2020. 3

    2020

  25. [33]

    Deep unsupervised learning using nonequilibrium thermodynamics

    Jascha Sohl-Dickstein, Eric Weiss, Niru Maheswaranathan, and Surya Ganguli. Deep unsupervised learning using nonequilibrium thermodynamics. InInternational confer- ence on machine learning, pages 2256–2265. PMLR, 2015. 2

    2015

  26. [34]

    Denoising diffusion implicit models.arXiv preprint arXiv:2010.02502, 2020

    Jiaming Song, Chenlin Meng, and Stefano Ermon. Denoising diffusion implicit models.arXiv preprint arXiv:2010.02502, 2020. 2

    2010 arXiv

  27. [35]

    Autoregressive model beats diffusion: Llama for scalable image generation.arXiv preprint arXiv:2406.06525, 2024

    Peize Sun, Yi Jiang, Shoufa Chen, Shilong Zhang, Bingyue Peng, Ping Luo, and Zehuan Yuan. Autoregressive model beats diffusion: Llama for scalable image generation.arXiv preprint arXiv:2406.06525, 2024. 2

    2024 arXiv

  28. [36]

    Neural discrete representation learning.Advances in neural information pro- cessing systems, 30, 2017

    Aaron Van Den Oord, Oriol Vinyals, et al. Neural discrete representation learning.Advances in neural information pro- cessing systems, 30, 2017. 3

    2017

  29. [37]

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

    Xintao Wang, Liangbin Xie, Chao Dong, and Ying Shan. Real-esrgan: Training real-world blind super-resolution with pure synthetic data. InProceedings of the IEEE/CVF inter- national conference on computer vision, pages 1905–1914,

    1905

  30. [38]

    Exploring representation-aligned latent space for better generation.arXiv preprint arXiv:2502.00359, 2025

    Wanghan Xu, Xiaoyu Yue, Zidong Wang, Yao Teng, Wen- long Zhang, Xihui Liu, Luping Zhou, Wanli Ouyang, and Lei Bai. Exploring representation-aligned latent space for better generation.arXiv preprint arXiv:2502.00359, 2025. 3

    2025 arXiv

  31. [39]

    Inf-dit: Upsampling any-resolution image with memory-efficient diffusion transformer.arXiv preprint arXiv:2405.04312, 2024

    Zhuoyi Yang, Heyang Jiang, Wenyi Hong, Jiayan Teng, Wendi Zheng, Yuxiao Dong, Ming Ding, and Jie Tang. Inf-dit: Upsampling any-resolution image with memory-efficient diffusion transformer.arXiv preprint arXiv:2405.04312, 2024. 2, 3, 7, 8

    2024 arXiv

  32. [40]

    Vector-quantized image modeling with improved vqgan.arXiv preprint arXiv:2110.04627, 2021

    Jiahui Yu, Xin Li, Jing Yu Koh, Han Zhang, Ruoming Pang, James Qin, Alexander Ku, Yuanzhong Xu, Jason Baldridge, and Yonghui Wu. Vector-quantized image modeling with improved vqgan.arXiv preprint arXiv:2110.04627, 2021. 2

    2021 arXiv

  33. [41]

    Designing a practical degradation model for deep blind image super-resolution

    Kai Zhang, Jingyun Liang, Luc Van Gool, and Radu Timo- fte. Designing a practical degradation model for deep blind image super-resolution. InProceedings of the IEEE/CVF International Conference on Computer Vision, pages 4791– 4800, 2021. 3

    2021

  34. [42]

    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, pages 3836–3847, 2023. 4

    2023

  35. [43]

    The unreasonable effectiveness of deep features as a perceptual metric

    Richard Zhang, Phillip Isola, Alexei A Efros, Eli Shecht- man, and Oliver Wang. The unreasonable effectiveness of deep features as a perceptual metric. InProceedings of the IEEE conference on computer vision and pattern recogni- tion, pages 586–595, 2018. 4

    2018

  36. [44]

    Fast training of diffusion models with masked transformers

    Hongkai Zheng, Weili Nie, Arash Vahdat, and Anima Anandkumar. Fast training of diffusion models with masked transformers. InTransactions on Machine Learning Re- search (TMLR), 2024. 3, 4, 6

    2024

Pith tools

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