Pith. sign in

REVIEW 3 major objections 8 minor 1 cited by

RDPM: Solve Diffusion Probabilistic Models via Recurrent Token Prediction

T0 review · 3 major / 8 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read The paper claims that diffusion probabilistic models can be solved by recurrent token prediction, converting the denoising process into a sequence of GPT-style classification steps over a vector-quantized codebook.

desk verdict A genuinely novel discrete-token generation recipe with competitive ImageNet FID, but the 'diffusion' framing outstrips the method; the learnability worry is real but missing evidence, not a demonstrated flaw. read the letter →

arxiv 2412.18390 v2 pith:KPAMUD2L submitted 2024-12-24 cs.CV cs.AIcs.LGcs.MM

Xiaoping Wu , Jie Hu , Xiaoming Wei This is my paper · ORCID
classification cs.CVcs.AIcs.LGcs.MM
keywords recurrentdiffusionprobabilisticmodeldiscretetokenpredictionvectorquantizationimagegenerationGPT-stylelossNet
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

RDPM claims a new way to run diffusion: instead of denoising continuous latents, it converts an image into a stack of vector-quantized codes, one per timestep, and trains a transformer to predict the next timestep's code from the noise, class label, timestep, and previously accumulated codes. Because the loss is cross-entropy over codebook entries, the optimization exactly matches GPT-style next-token prediction, which the authors argue opens the door to unified multimodal models. On ImageNet 256x256, RDPM reaches FID 2.56 in only 10 sampling steps with a 602M-parameter transformer, outperforming prior discrete-token methods and approaching continuous diffusion models. The paper frames its contribution as a new paradigm—'discrete diffusion'—rather than a marginal accuracy gain.

What carries the argument

The central object is the diffusion-based image tokenizer of Algorithm 1 combined with the recurrent token predictor of Algorithm 2. The tokenizer mixes Gaussian noise into the latent over T steps, each time quantizing the residual against a shared codebook, so an image becomes T code maps. The generator is a transformer f that maps (epsilon_t, y, t, z'_{t-1}) to the next code map, with the accumulated quantized vectors z'_{t-1} carrying information forward; the cross-entropy loss in Eq. (7) is identical in form to language-model next-token prediction, which is what the paper says unifies discrete and continuous signal generation.

What would settle it

Fix an image and a step t, then draw many noise samples epsilon_t while keeping z'_{t-1} fixed, and record the empirical distribution of the ground-truth code C_t. If the RDPM transformer's predicted distribution over the codebook at that (z'_{t-1}, epsilon_t, y, t) diverges sharply from this empirical distribution, the recurrent chain is not learning the intended denoising transition, and the method's claim of solving diffusion via token prediction would be falsified.

Watch

Extended reading notes

Core claim

The central discovery is that a diffusion process can be 'solved' by recurrent token prediction: the forward process adds Gaussian noise to the VAE latent and quantizes the residual each step, producing T discrete codes; the reverse process is a transformer that, at each step, predicts the code for that step given the noise, class condition, timestep, and the codes accumulated so far. This turns denoising into a sequence of classification problems with a GPT-style cross-entropy loss, and the paper demonstrates that ten such recurrent steps are enough to reach FID 2.56 on ImageNet 256x256 at 602M parameters.

Load-bearing premise

The method assumes that a model can learn to predict the next quantized code from the noise, the class, the timestep, and the codes already produced, even though the original image encoding that determined that code is never shown to it.

Editorial extensions

If this is right

  • Only 10 recurrent steps are needed for high-quality generation, far fewer than the hundreds of steps typical of continuous diffusion, while preserving quality.
  • The loss function matches GPT-style next-token prediction, so the same optimization framework used for text can be applied to image, video, and audio generation.
  • The diffusion-based image tokenizer reduces discretization information loss compared to single-step VQ, as shown by reconstruction FID dropping to 1.52 with the weight and bias tricks.
  • RDPM outperforms prior methods that use discrete visual tokenizers on ImageNet 256x256, including MaskGIT and VAR at comparable scales.
  • The approach is not limited to images; the tokenization applies to any continuous signal domain.

Reading between the lines

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

  • If the learnability assumption holds at scale, RDPM suggests a path toward a single model that handles text and continuous modalities with one cross-entropy objective, potentially simplifying multimodal training recipes.
  • The T-step tokenization increases the number of tokens per image by a factor of T over a single VQ code map; the paper's success at T=10 implies the transformer can exploit temporal structure across timesteps, which might transfer to video generation where the timestep axis and the frame axis could be merged.
  • A direct comparison with a version using a standard VQ-VAE tokenizer (single code map) under the same recurrent predictor would isolate how much of the gain comes from the diffusion-based tokenization versus the recurrent prediction itself; the paper does not report that ablation.
  • Because inference injects fresh Gaussian noise at every step, the model's diversity may be tunable by the noise schedule and Gumbel temperature; this suggests a testable knob for controlling the precision-diversity trade-off in discrete diffusion.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 8 minor

Summary. The paper proposes RDPM, a framework for class-conditional image generation on ImageNet 256×256. A VAE is trained with a diffusion-based quantization scheme (Algorithm 1) that encodes the latent into T=10 discrete code maps by progressively mixing Gaussian noise with residual vectors and quantizing. A transformer is then trained (Eq. 6) to predict the next code map from the current noise, class label, timestep, and accumulated quantized vectors, using a cross-entropy loss (Eq. 7). At inference, Algorithm 2 recurrently predicts code maps starting from zero accumulation and decodes the final sum. The method reports FID 2.56 at 602M parameters with 10 sampling steps, comparable to recent VAR models and better than several prior discrete-tokenizer baselines.

Significance. The empirical results are noteworthy: they demonstrate that a discrete-token generative model trained with a GPT-style cross-entropy loss can reach FID levels close to continuous diffusion models on ImageNet 256×256 while using only 10 steps. The paper also promises code and weight release, which would support reproducibility. However, the theoretical framing as a “diffusion” process is not established, and the central learnability assumption of the recurrent predictor is not examined. If the authors can provide diagnostic evidence for the conditional predictability of the codes and temper the diffusion claims, this would be a useful empirical contribution to discrete-token generation.

major comments (3)
  1. [Section 3.2, Eq. (6) and Algorithm 2] The training target v'_t = Q(alpha_t(v - z'_{t-1}) + beta_t epsilon_t) depends on the original latent v, which is not an input to the predictor f(epsilon_t, y, t, z'_{t-1}). Since z'_{t-1} is a lossy function of v (it is a sum of quantized residuals), multiple source latents can map to the same z'_{t-1}, and for fixed inputs (epsilon_t, y, t, z'_{t-1}) the target code is not uniquely determined. The cross-entropy loss in Eq. (7) is therefore minimizing an irreducible conditional entropy, but the paper provides no measurement of this entropy or of the model's teacher-forced code accuracy. Without such diagnostics, it is unclear whether the free-running accumulation in Algorithm 2 stays on the training manifold. Please report teacher-forced top-1 accuracy, an estimate of the conditional entropy of target codes given the inputs, and a comparison of teacher-forced versus free-running FID or reconstruction quality.
  2. [Abstract, Sections 1, 3.2, 4.5] The repeated claim that RDPM “solves diffusion probabilistic models” or performs a “diffusion process on discrete-value domains” is not supported by the methodology. Algorithm 2 is an autoregressive code predictor: there is no reverse Markov chain, no SDE discretization, no ELBO, and no derivation that iterating Eq. (6) maps standard Gaussian noise to the data distribution. The forward process in Algorithm 1 is not inverted in any formal sense; the model simply learns a conditional distribution over codes. The authors should either provide a formal connection between the recurrent prediction objective and a diffusion process, or revise the terminology to describe an autoregressive model over residual quantized codes with noise-conditioned inputs.
  3. [Section 4.3, Figure 5, Table 4] The noise schedule hyperparameter phi is selected by evaluating FID-50K on the same metric used for the final reported performance. The pow schedule with phi=0.75 is chosen because it yields the best FID-50K, which introduces selection bias into the headline numbers. Please report results with a fixed, untuned schedule (e.g., a standard linear or cosine schedule) and clearly state which hyperparameters were tuned on the evaluation metric, or use a validation split for hyperparameter selection.
minor comments (8)
  1. [Introduction / Related Work] The paper claims to pioneer “Discrete Diffusion” but does not cite prior work on discrete diffusion models such as D3PM (Austin et al., 2021) or CDCD (Hoogeboom et al., 2021). Please add references and position RDPM relative to these lines of work.
  2. [Section 3.1, Eq. (2)] The text says alpha_t increases from small to large and reaches 1 at t=T, while beta_t^2 + alpha_t^2 = 1. This is the opposite of the standard diffusion direction, where noise increases over time. Please clarify the intended direction and justify why this is called a diffusion process.
  3. [Section 3.1, Algorithm 1] The quantization operation Q(·) is described as nearest-neighbor search, but the loss in Eq. (4) requires gradients to flow through the quantization step. Please specify whether a straight-through estimator or other gradient approximation is used, as is standard in VQ-VAE and VQGAN.
  4. [Figure 3] The caption says rows 2 and 3 represent discrete quantized tokens, but the figure itself does not label which row corresponds to which method. Please add explicit labels to the row illustrations.
  5. [Table 3] The row labeled “Raw” is unclear; please define what “raw” means (e.g., no CFG, no Gumbel noise, greedy decoding) in the caption or text.
  6. [Eq. (7)] The displayed cross-entropy formula contains an extra “log” inside the softmax expression; it should be Ct,i,j log( e^{bv'_{t,i,j}} / sum_k e^{bv'_{t,i,k}} ) or equivalent. Please correct the notation.
  7. [Algorithm 2, line 3] The phrase “predict the direction of the next data distribution” is unclear; consider replacing with “predict the next quantized vector v'_t” or “estimate the next code map”.
  8. [References] Reference [38] contains a typo: “Tian et at.” should be “Tian et al.” Please fix.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: RDPM's performance claims are empirical comparisons against external baselines, and the identified identifiability gap is a correctness risk rather than a self-referential derivation.

full rationale

The derivation chain is not circular. The tokenization module (Alg. 1, Eqs. 2-5) defines quantized targets from the image latent and Gaussian noise, and the generation module (Eq. 6, Alg. 2) is trained with cross-entropy against those external codebook indices; no predicted quantity is defined in terms of the final FID or of the paper's own conclusions. The reference list contains no self-citations, so there is no load-bearing self-citation chain. The variance schedule and φ=0.75 are selected on FID-50K (Fig. 5, Table 4), but this is explicit hyperparameter selection on an external benchmark, not a fitted parameter disguised as a prediction, and the headline results in Table 5 are measured against independently published models. The genuine weakness is a learnability assumption: in Eq. (2), v'_t = Q(α_t z_t + β_t ε_t) with z_t = v - z'_{t-1}, while f in Eq. (6) receives only ε_t, y, t, and z'_{t-1}; because v is not observed, the target is not a deterministic function of the model's inputs, and the paper supplies no identifiability, conditional-entropy, or teacher-forced accuracy analysis. That is a correctness risk to raise in review, not a circular reduction.

Assumptions & free parameters 6 free parameters · 5 assumptions · 0 invented entities

The method introduces no new physical or mathematical entity; 'discrete diffusion' is a rebranding of a recurrent residual code predictor. The main free parameters are either borrowed from VQGAN or tuned on the reported FID, and several implicit assumptions about the learnability of the conditional mapping are unstated.

free parameters (6)
  • T (number of quantization/generation steps) = 10
    Number of diffusion/quantization and generation steps; chosen in Section 4.3 ablation because T<10 degrades reconstruction and T>10 gives marginal gain (Table 2).
  • noise schedule alpha_t = phi^(T-t) = phi=0.75
    Chosen by FID-50K in Figure 5; the pow schedule is reported to give the best FID among sin, linear, and pow, and phi is further tuned.
  • quantization loss weight gamma_t = gamma_t = alpha_t
    Set proportional to alpha_t in Eq. (4); Table 2, Exp. 7 shows rFID drops from 3.05 to 2.04 when the weight is applied.
  • codebook size K and embedding dim d = K=4096, d=32
    Hyperparameters stated in Section 4.2; no ablation on K or d is reported, so their effect on the central claim is unknown.
  • CFG strength lambda and Gumbel temperature tau = not reported exactly
    Table 3 shows CFG, linear CFG, and Gumbel noise improve the final FID from 19.38 to 3.68; the paper does not give the lambda and tau values used.
  • loss weights delta and eta = delta=0.25, eta=0.75
    Copied from VQGAN as stated in Section 3.1; not ablated.
assumptions (5)
  • ad hoc to paper The iterative noise-injection and quantization process in Eq. (2) defines a diffusion process that can be inverted by recurrent code prediction.
    Section 3.1 and Section 3.2 use the term 'diffusion' without defining a forward/reverse Markov chain, an ELBO, or a stochastic differential equation; the equivalence is asserted, not derived.
  • domain assumption The mapping in Eq. (6) is a well-posed learning problem: v'_t can be predicted from epsilon_t, t, y, and z'_{t-1} despite the hidden original latent v.
    The target depends on v, which never appears in the model input; no identifiability or consistency argument is given.
  • domain assumption Cross-entropy loss over codebook indices in Eq. (7) is an appropriate objective for learning the recursive code sequence, analogous to LLM next-token loss.
    The paper aligns with GPT-style loss, but the tokens are not a natural language sequence and the objective's mode coverage and error-accumulation properties are not analyzed.
  • ad hoc to paper Adding layer normalization at the encoder output ensures the latent scale is compatible with injected Gaussian noise.
    Section 3.1 introduces this design choice without experimental or theoretical support, and the ablation table does not include a no-layer-norm baseline.
  • domain assumption ImageNet-1k and FID-50K are sufficient benchmarks for the claims about a general unified generative paradigm.
    The paper evaluates only one dataset and one resolution; the broader claims about video, audio, and unified multimodal generation are not tested.

how reviews work

0 comments
Cite this review

Pith. "Pith review of RDPM: Solve Diffusion Probabilistic Models via Recurrent Token Prediction." pith.science (2026). https://pith.science/paper/KPAMUD2L

@misc{pith2026241218390,
  author       = {Pith},
  title        = {Pith review of: RDPM: Solve Diffusion Probabilistic Models via Recurrent Token Prediction},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/KPAMUD2L}},
  note         = {Machine review of arXiv:2412.18390}
}
read the original abstract

Diffusion Probabilistic Models (DPMs) have emerged as the de facto approach for high-fidelity image synthesis, operating diffusion processes on continuous VAE latent, which significantly differ from the text generation methods employed by Large Language Models (LLMs). In this paper, we introduce a novel generative framework, the Recurrent Diffusion Probabilistic Model (RDPM), which enhances the diffusion process through a recurrent token prediction mechanism, thereby pioneering the field of Discrete Diffusion. By progressively introducing Gaussian noise into the latent representations of images and encoding them into vector-quantized tokens in a recurrent manner, RDPM facilitates a unique diffusion process on discrete-value domains. This process iteratively predicts the token codes for subsequent timesteps, transforming the initial standard Gaussian noise into the source data distribution, aligning with GPT-style models in terms of the loss function. RDPM demonstrates superior performance while benefiting from the speed advantage of requiring only a few inference steps. This model not only leverages the diffusion process to ensure high-quality generation but also converts continuous signals into a series of high-fidelity discrete tokens, thereby maintaining a unified optimization strategy with other discrete tokens, such as text. We anticipate that this work will contribute to the development of a unified model for multimodal generation, specifically by integrating continuous signal domains such as images, videos, and audio with text. We will release the code and model weights to the open-source community.

Figures

Figures reproduced from arXiv: 2412.18390 by the authors.

Figure 1
Figure 1. Taxonomy of modern image generation. From the con [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Visualizations of the ImageNet class-conditional [PITH_FULL_IMAGE:figures/full_fig_p001_2.png] view at source ↗
Figure 3
Figure 3. Comparison of various visual generation patterns. The images in rows 2 and 3, segmented into blocks, represent discrete [PITH_FULL_IMAGE:figures/full_fig_p002_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Overview of the RDPM framework. The framework [PITH_FULL_IMAGE:figures/full_fig_p004_4.png]
Figure 5
Figure 5. Figure 5: (Left) Variation of αt across different scheduling strate￾gies. (Right) Changes in FID as the hyperparameter φ varies under the pow schedule. The experiment is conducted on the model with depth of 12 and trained for 100 epochs. but compromises diversity. Building on th…

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. Compressed Image Generation with Denoising Diffusion Codebook Models

    eess.IV 2025-02 conditional novelty 8.0 of 10

    Using fixed codebooks of noise vectors in diffusion sampling yields images that carry their own compressed bit-streams and enables a strong perceptual image codec.

Reference graph

Works this paper leans on

46 extracted references · 11 canonical work pages · cited by 1 Pith paper

  1. [1]

    Gpt-4 technical report

    Josh Achiam, Steven Adler, Sandhini Agarwal, Lama Ah- mad, Ilge Akkaya, Florencia Leoni Aleman, Diogo Almeida, Janko Altenschmidt, Sam Altman, Shyamal Anadkat, et al. Gpt-4 technical report. arXiv preprint arXiv:2303.08774 ,

    arXiv

  2. [2]

    Wasserstein generative adversarial networks

    Martin Arjovsky, Soumith Chintala, and L ´eon Bottou. Wasserstein generative adversarial networks. In Interna- tional conference on machine learning , pages 214–223. PMLR, 2017. 3

    2017

  3. [3]

    Hin- ton

    Lei Ba, Jimmy, Jamie Ryan Kiros, and Geoffrey E. Hin- ton. Layer normalization. arXiv preprint arXiv:1607.06450.,

    arXiv

  4. [4]

    Large scale gan training for high fidelity natural image synthesis

    Andrew Brock, Jeff Donahue, and Karen Simonyan. Large scale gan training for high fidelity natural image synthesis. arXiv preprint arXiv:1809.11096, 2018. 3

    arXiv 2018

  5. [5]

    Lan- guage models are few-shot learners

    Tom Brown, Benjamin Mann, Nick Ryder, Melanie Sub- biah, Jared D Kaplan, Prafulla Dhariwal, Arvind Neelakan- tan, Pranav Shyam, Girish Sastry, Amanda Askell, et al. Lan- guage models are few-shot learners. Advances in neural in- formation processing systems, 33:1877–1901, 2020. 3

    1901

  6. [6]

    Maskgit: Masked generative image transformer

    Huiwen Chang, Han Zhang, Lu Jiang, Ce Liu, and William T Freeman. Maskgit: Masked generative image transformer. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 11315–11325, 2022. 2, 3, 5, 7

    work page 2022

  7. [7]

    Muse: Text- to-image generation via masked generative transformers

    Huiwen Chang, Han Zhang, Jarred Barber, AJ Maschinot, Jose Lezama, Lu Jiang, Ming-Hsuan Yang, Kevin Murphy, William T Freeman, Michael Rubinstein, et al. Muse: Text- to-image generation via masked generative transformers. In International Conference on Machine Learning, 2023. 3

    work page 2023

  8. [8]

    Pixart-alpha: Fast training of diffusion transformer for photorealistic text-to-image synthesis

    Junsong Chen, Jincheng Yu, Chongjian Ge, Lewei Yao, Enze Xie, Yue Wu, Zhongdao Wang, James Kwok, Ping Luo, Huchuan Lu, et al. Pixart-alpha: Fast training of diffusion transformer for photorealistic text-to-image synthesis. arXiv preprint arXiv:2310.00426, 2023. 3

    arXiv

Show all 46 references
  1. [9]

    Flashattention-2: Faster attention with bet- ter parallelism and work partitioning

    Tri Dao. Flashattention-2: Faster attention with bet- ter parallelism and work partitioning. arXiv preprint arXiv:2307.08691, 2023. 6

    2023 arXiv

  2. [10]

    Imagenet: A large-scale hierarchical image database

    Jia Deng, Wei Dong, Richard Socher, Li-Jia Li, Kai Li, and Fei-Fei Li. Imagenet: A large-scale hierarchical image database. In IEEE conference on computer vision and pat- tern recognition, 2009. 5

    2009

  3. [11]

    Diffusion models beat gans on image synthesis

    Prafulla Dhariwal and Alexander Nichol. Diffusion models beat gans on image synthesis. Advances in neural informa- tion processing systems, 34:8780–8794, 2021. 3

    2021

  4. [12]

    Taming transformers for high-resolution image synthesis

    Patrick Esser, Robin Rombach, and Bjorn Ommer. Taming transformers for high-resolution image synthesis. In Pro- ceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 12873–12883, 2021. 1, 3, 6, 7

    2021

  5. [13]

    Scaling recti- fied flow transformers for high-resolution image synthesis

    Patrick Esser, Sumith Kulal, Andreas Blattmann, Rahim Entezari, Jonas M ¨uller, Harry Saini, Yam Levi, Dominik Lorenz, Axel Sauer, Frederic Boesel, et al. Scaling recti- fied flow transformers for high-resolution image synthesis. In International Conference on Machine Learning...

    2024

  6. [14]

    Generative adversarial networks

    Ian Goodfellow, Jean Pouget-Abadie, Mehdi Mirza, Bing Xu, David Warde-Farley, Sherjil Ozair, Aaron Courville, and Yoshua Bengio. Generative adversarial networks. Commu- nications of the ACM, 63(11):139–144, 2020. 3

    2020

  7. [15]

    Classifier-free diffusion guidance

    Jonathan Ho and Tim Salimans. Classifier-free diffusion guidance. arXiv preprint arXiv:2207.12598, 2022. 1, 3, 5, 6

    2022 arXiv

  8. [16]

    Denoising dif- fusion probabilistic models

    Jonathan Ho, Ajay Jain, and Pieter Abbeel. Denoising dif- fusion probabilistic models. Advances in neural information processing systems, 33:6840–6851, 2020. 3

    2020

  9. [17]

    Cascaded diffu- sion models for high fidelity image generation

    Jonathan Ho, Chitwan Saharia, William Chan, David J Fleet, Mohammad Norouzi, and Tim Salimans. Cascaded diffu- sion models for high fidelity image generation. Journal of Machine Learning Research, 23(47):1–33, 2022. 1

    2022

  10. [18]

    Lora: Low-rank adaptation of large language models

    Edward J Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen- Zhu, Yuanzhi Li, Shean Wang, Lu Wang, and Weizhu Chen. Lora: Low-rank adaptation of large language models. arXiv preprint arXiv:2106.09685, 2021. 3

    2021 arXiv

  11. [19]

    Categorical reparameterization with gumbel-softmax

    Eric Jang, Shixiang Gu, and Ben Poole. Categorical reparameterization with gumbel-softmax. arXiv preprint arXiv:1611.01144, 2016. 5

    2016 arXiv

  12. [20]

    Mistral 7b

    Albert Q Jiang, Alexandre Sablayrolles, Arthur Mensch, Chris Bamford, Devendra Singh Chaplot, Diego de las Casas, Florian Bressand, Gianna Lengyel, Guillaume Lam- ple, Lucile Saulnier, et al. Mistral 7b. arXiv preprint arXiv:2310.06825, 2023. 1

    arXiv

  13. [21]

    Scal- ing up gans for text-to-image synthesis

    Minguk Kang, Jun-Yan Zhu, Richard Zhang, Jaesik Park, Eli Shechtman, Sylvain Paris, and Taesung Park. Scal- ing up gans for text-to-image synthesis. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 10124–10134, 2023. 3

    2023

  14. [22]

    Auto-encoding varia- tional bayes

    Diederik P Kingma and Max Welling. Auto-encoding varia- tional bayes. arXiv preprint arXiv:1312.6114, 2013. 3

    2013 arXiv

  15. [23]

    Autoregressive image generation using residual quantization

    Doyup Lee, Chiheon Kim, Saehoon Kim, Minsu Cho, and Wook-Shin Han. Autoregressive image generation using residual quantization. In Proceedings of the IEEE/CVF Con- ference on Computer Vision and Pattern Recognition, pages 11523–11532, 2022. 3, 4, 7

    2022

  16. [24]

    Mage: Masked generative encoder to unify representation learning and image synthe- sis

    Tianhong Li, Huiwen Chang, Shlok Mishra, Han Zhang, Dina Katabi, and Dilip Krishnan. Mage: Masked generative encoder to unify representation learning and image synthe- sis. In Proceedings of the IEEE/CVF Conference on Com- puter Vision and Pattern Recognition , pages 2142–2152,

  17. [25]

    Autoregressive image generation without vec- tor quantization

    Tianhong Li, Yonglong Tian, He Li, Mingyang Deng, and Kaiming He. Autoregressive image generation without vec- tor quantization. arXiv preprint arXiv:2406.11838, 2024. 2, 8

    2024 arXiv

  18. [26]

    Dpm-solver: A fast ode solver for diffusion probabilistic model sampling in around 10 steps

    Cheng Lu, Yuhao Zhou, Fan Bao, Jianfei Chen, Chongxuan Li, and Jun Zhu. Dpm-solver: A fast ode solver for diffusion probabilistic model sampling in around 10 steps. Advances in Neural Information Processing Systems , 35:5775–5787,

  19. [27]

    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. 3

    2023 arXiv

  20. [28]

    Improved denoising diffusion probabilistic models

    Alexander Quinn Nichol and Prafulla Dhariwal. Improved denoising diffusion probabilistic models. In International conference on machine learning, pages 8162–8171. PMLR,

  21. [29]

    Scalable diffusion models with transformers

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

  22. [30]

    Gener- ating diverse high-fidelity images with vq-vae-2

    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. 1, 3, 7

    2019

  23. [31]

    High-resolution image synthesis with latent diffusion models

    Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser, and Bj ¨orn Ommer. High-resolution image synthesis with latent diffusion models. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 10684–10695, 2022. 1, 7

    2022

  24. [32]

    U- net: Convolutional networks for biomedical image segmen- tation

    Olaf Ronneberger, Philipp Fischer, and Thomas Brox. U- net: Convolutional networks for biomedical image segmen- tation. In Medical image computing and computer-assisted intervention–MICCAI 2015: 18th international conference, Munich, Germany, October 5-9, 2015, proceedings, pa...

    2015

  25. [33]

    Scaling stylegan to large diverse datasets

    A Sauer, K Schwarz, and A StyleGAN-XL Geiger. Scaling stylegan to large diverse datasets. InProceedings of the SIG- GRAPH Conference. ACM, pages 1–10, 2022. 3, 7

    2022

  26. [34]

    Very deep convo- lutional networks for large-scale image recognition

    Karen Simonyan and Andrew Zisserman. Very deep convo- lutional networks for large-scale image recognition. arXiv preprint arXiv:1409.1556, 2014. 4

    2014 arXiv

  27. [35]

    Denoising diffusion implicit models

    Jiaming Song, Chenlin Meng, and Stefano Ermon. Denoising diffusion implicit models. arXiv preprint arXiv:2010.02502, 2020. 1, 3

    2010 arXiv

  28. [36]

    Score-based generative modeling through stochastic differential equa- tions

    Yang Song, Jascha Sohl-Dickstein, Diederik P Kingma, Ab- hishek Kumar, Stefano Ermon, and Ben Poole. Score-based generative modeling through stochastic differential equa- tions. arXiv preprint arXiv:2011.13456, 2020. 1

    2011 arXiv

  29. [37]

    Consistency models

    Yang Song, Prafulla Dhariwal, Mark Chen, and Ilya Sutskever. Consistency models. arXiv preprint arXiv:2303.01469, 2023. 3

    2023 arXiv

  30. [38]

    Visual autoregressive modeling: Scalable image generation via next-scale prediction

    Keyu Tian, Yi Jiang, Zehuan Yuan, Bingyue Peng, and Li- wei Wang. Visual autoregressive modeling: Scalable image generation via next-scale prediction. Advances in neural in- formation processing systems, 2024. 2, 7, 8

    2024

  31. [39]

    Llama: Open and efficient foundation language models

    Hugo Touvron, Thibaut Lavril, Gautier Izacard, Xavier Martinet, Marie-Anne Lachaux, Timoth´ee Lacroix, Baptiste Rozi`ere, Naman Goyal, Eric Hambro, Faisal Azhar, et al. Llama: Open and efficient foundation language models. arXiv preprint arXiv:2302.13971, 2023. 1, 3, 6

    2023 arXiv

  32. [40]

    Givt: Generative infinite-vocabulary transformers

    Michael Tschannen, Cian Eastwood, and Fabian Mentzer. Givt: Generative infinite-vocabulary transformers. arXiv preprint arXiv:2312.02116, 2023. 1

    2023 arXiv

  33. [41]

    Neural discrete representation learning

    Aaron Van Den Oord, Oriol Vinyals, et al. Neural discrete representation learning. Advances in neural information pro- cessing systems, 30, 2017. 1, 3

    2017

  34. [42]

    Show-o: One single transformer to unify multimodal understanding and generation

    Jinheng Xie, Weijia Mao, Zechen Bai, David Junhao Zhang, Weihao Wang, Kevin Qinghong Lin, Yuchao Gu, Zhijie Chen, Zhenheng Yang, and Mike Zheng Shou. Show-o: One single transformer to unify multimodal understanding and generation. arXiv preprint arXiv:2408.12528, 2024. 8

    2024 arXiv

  35. [43]

    Vector-quantized image modeling with improved vqgan

    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. 1, 3, 4, 7

    2021 arXiv

  36. [44]

    Magvit: Masked generative video transformer

    Lijun Yu, Yong Cheng, Kihyuk Sohn, Jos ´e Lezama, Han Zhang, Huiwen Chang, Alexander G Hauptmann, Ming- Hsuan Yang, Yuan Hao, Irfan Essa, et al. Magvit: Masked generative video transformer. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, p...

    2023

  37. [45]

    Language model beats diffusion–tokenizer is key to visual generation

    Lijun Yu, Jos ´e Lezama, Nitesh B Gundavarapu, Luca Ver- sari, Kihyuk Sohn, David Minnen, Yong Cheng, Agrim Gupta, Xiuye Gu, Alexander G Hauptmann, et al. Language model beats diffusion–tokenizer is key to visual generation. arXiv preprint arXiv:2310.05737, 2023. 1

    arXiv

  38. [46]

    Transfusion: Pre- dict the next token and diffuse images with one multi-modal model

    Chunting Zhou, Lili Yu, Arun Babu, Kushal Tirumala, Michihiro Yasunaga, Leonid Shamis, Jacob Kahn, Xuezhe Ma, Luke Zettlemoyer, and Omer Levy. Transfusion: Pre- dict the next token and diffuse images with one multi-modal model. arXiv preprint arXiv:2408.11039, 2024. 8

    2024 arXiv

Pith tools

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