Pith. sign in

REVIEW 2 major objections 6 minor 45 references

Generative Diffusion Modeling: A Practical Handbook

T0 review · 2 major / 6 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read This handbook claims a shared notation and explicit velocity mappings can close the paper-to-code gap across diffusion, consistency, and rectified-flow models.

desk verdict A genuinely useful notation-unifying handbook, but two real equation errors—the velocity mapping and the DDPO variance—mean the unified perspective overshoots for general schedules. read the letter →

arxiv 2412.17162 v1 pith:AWLNMFHZ submitted 2024-12-22 cs.LG cs.CV

Zihan Ding , Chi Jin This is my paper · ORCID
classification cs.LGcs.CV
keywords diffusionprobabilisticmodelsscore-basedgenerativeconsistencyrectifiedflowmatchingTrigvelocitymappingpaper-to-codegap
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper is a practical handbook whose central claim is that the many flavors of diffusion-based generation—denoising diffusion probabilistic models, score-based models, consistency models, rectified flow, flow matching, and TrigFlow—can be described in one shared notation that is deliberately aligned with actual code implementations. It argues that the notorious "paper-to-code" gap comes mostly from inconsistent notation and missing implementation details rather than from differences in the underlying mathematics, and that a unified presentation closes that gap. The handbook derives common training objectives, inference procedures, prediction parameterizations, distillation methods, and reward-based fine-tuning in that notation, and it supplies explicit transformations between $\epsilon$-, $x$-, $v$-, and $f$-predictions as well as a velocity mapping between diffusion trajectories and rectified-flow trajectories. If the unified perspective is correct, a practitioner who follows the handbook can implement any of these methods, convert one into another, and compare them fairly without re-deriving formulas from each original paper.

What carries the argument

The carrying object is the linear interpolant $x_t = \alpha_t x_0 + \beta_t x_1$ with differentiable coefficients, together with the derived velocity–score identity $$v(x_t,t) = \frac{\dot{\$\alpha$}_t}{\alpha_t} x_t + \beta_t\!\left(\frac{\dot{\$\alpha$}_t \beta_t}{\alpha_t} - \dot{\$\beta$}_t\right) s(x_t,t),$$ where $s$ is the score function and dots are time derivatives. This identity supplies the "unified formulation" that the handbook places at the center of the family. The other load-bearing piece is the velocity mapping between the diffusion variable $x_t = \sqrt{\bar{\alpha}_t}x_0 + \sqrt{1-\bar{\alpha}_t}\varepsilon$ and the rectified-flow variable $y_t = x_t / (\sqrt{\bar{\alpha}_t} + \sqrt{1-\bar{\alpha}_t})$, written as $$v^y_t = \frac{v^x_t b - x_t(\sqrt{1-\bar{\$\alpha$}_t} - \sqrt{\bar{\$\alpha$}_t})}{$b^{2}$},\quad b = \sqrt{\bar{\$\alpha$}_t} + \sqrt{1-\bar{\$\alpha$}_t},$$ which converts a diffusion trajectory into a rectified-flow trajectory and back. Together these identities let every prediction head and every training or inference scheme be expressed in the same variables, and they are what make the claimed paper-to-code bridge concrete.

What would settle it

Evaluate both sides of the claimed velocity mapping for a standard variance-preserving schedule (for example, a linear noise schedule over 1000 steps) at several timesteps on the same pair $(x_0,\varepsilon)$: if the mapped diffusion velocity $v^y_t$ does not equal $(x_0-\varepsilon)/(1+2\sqrt{\bar{\alpha}_t}\sqrt{1-\bar{\alpha}_t})$ for those timesteps, then the mapping's general validity is refuted.

Watch

Extended reading notes

Core claim

The central claim is that a single formulation $x_t = \alpha_t x_0 + \beta_t x_1$, with time-dependent coefficients $\alpha_t, \beta_t$ and a velocity–score relationship, covers the diffusion-family models treated in the handbook, and that the apparent differences among them are notation and scheduling choices rather than fundamental divisions. Within this formulation, the three common prediction heads ($\epsilon$-prediction, $x$-prediction, $v$-prediction) are linked by "triangular" identities, and a consistency model's $f$-prediction can be reparameterized through a diffusion model's noise prediction. A further derived velocity mapping connects the velocity along a diffusion trajectory to the velocity along a rectified-flow trajectory, which the handbook uses to justify applying rectified-flow training objectives to pre-trained diffusion models and to enable straight-line few-step sampling. The handbook also presents TrigFlow as the formulation that unifies continuous-time consistency models with rectified flow through an angular parameterization. The paper's discovery, on its own terms, is that these methods form one family whose members are interconvertible in code if the notation is fixed first.

Load-bearing premise

The unified picture in Section 2.6.3 rests on the assumption that the velocity mapping between rectified-flow and diffusion trajectories holds for all the noise schedules the handbook uses, even though the derivation relies on a special trigonometric schedule; if the mapping fails for ordinary variance-preserving schedules, the unification is incomplete.

Editorial extensions

If this is right

  • A practitioner can convert a model trained with $\epsilon$-prediction into $x$- or $v$-prediction using the triangular identities, without retraining the network.
  • The velocity mapping lets a pre-trained diffusion model be fine-tuned with a rectified-flow loss and then sampled along a straighter trajectory, reducing the number of sampling steps.
  • A consistency model can be initialized from a pre-trained diffusion model through the $f$-prediction reparameterization, so distillation starts from the teacher's weights rather than from scratch.
  • TrigFlow's angular parameterization gives one formulation in which the same model can be trained by the diffusion objective or by continuous-time consistency training.
  • Post-training techniques—progressive distillation, score distillation, distribution matching, adversarial distillation, consistency-based distillation, and reward-based fine-tuning—are all expressible in the same notation, which makes fair comparisons of their costs and gains possible.

Reading between the lines

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

  • A plausible extension the handbook leaves implicit is that adopting this notation as a shared convention would let downstream benchmarks report the exact schedule, prediction head, and sampler configuration as part of the result, making reproduction failures traceable to parameter choices instead of notation.
  • Because the velocity mapping's derivation uses a trigonometric schedule, a natural test is whether the same mapping holds, with a correction term, for variance-preserving schedules; if it does not, the unification would need to be reformulated as schedule-dependent rather than universal.
  • If the handbook's code-aligned notation is implemented as a single reusable code path, the time cost of switching between DDPM-style sampling, consistency-model sampling, and rectified-flow sampling should drop to changing a few schedule and prediction parameters; this is directly measurable and would confirm or undermine the paper-to-code claim.
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

2 major / 6 minor

Summary. This manuscript is a practical handbook on diffusion models that aims to unify notation and presentation across diffusion probabilistic models, score-based generative models, consistency models, rectified flow, flow matching, and TrigFlow. It reviews the fundamentals of forward and reverse processes, training objectives, inference algorithms, prediction parameterizations, and post-training techniques including distillation and reward-based fine-tuning. The handbook provides explicit algorithms and pseudocode throughout and attempts to connect the formulations of different methods to facilitate implementation and comparison.

Significance. If the technical claims are corrected, this handbook would be a valuable practical reference for practitioners, offering a single notation system, explicit algorithms, and coverage of recent methods such as LCM, DMD, and DDPO. Its strengths include a clear notation table, detailed derivations of standard DDPM/DDIM/score-matching results, and an extensive set of pseudocode algorithms. However, the claimed unified perspective relies on a velocity mapping between diffusion and rectified-flow trajectories that is established only under a trigonometric time schedule; as written, the proof in §2.6.3 is not valid for the general variance-preserving schedules otherwise used in the handbook. This undermines the generality of the unification as presented and requires either a corrected derivation or a restricted claim.

major comments (2)
  1. [§2.6.3, Eq. (39); used in §2.4.5 and §2.4.4] The proof of the velocity mapping between rectified-flow and diffusion trajectories assumes without stating a schedule-specific derivative. Differentiating b = sqrt(alpha_bar_t) + sqrt(1 - alpha_bar_t) gives b' = alpha_bar_t' (sqrt(1-alpha_bar_t) - sqrt(alpha_bar_t)) / (2 sqrt(alpha_bar_t(1-alpha_bar_t))), which equals sqrt(1-alpha_bar_t) - sqrt(alpha_bar_t) only when alpha_bar_t' = 2 sqrt(alpha_bar_t(1-alpha_bar_t)), i.e., for a trigonometric time parameterization such as TrigFlow. For standard VP schedules with alpha_bar_t = exp(-integral beta), the displayed derivative is incorrect. Since Eq. (39) and the equalities in §2.4.5 rely on this derivative, the claimed general mapping between diffusion velocities and rectified-flow velocities is not established for the VP schedules otherwise used in the handbook. The same issue also affects the definition of vx_t as dot x_t in the proof. Please either restrict the claim to TrigFlow or provide the corrected general mapping; this is load-bearing for Approach 2 in §2.4.4 and for the claimed unification.
  2. [§3.3.1, Eq. (50)] The DDPO variance formula in Eq. (50) is incorrect. The expression sqrt((1-alpha_bar_{t-1})/(1-alpha_bar_t)) * sqrt((1-alpha_bar_t)/alpha_bar_{t-1}) simplifies to sqrt((1-alpha_bar_{t-1})/alpha_bar_{t-1}), which is not the posterior standard deviation derived in Eq. (9) and used in Eqs. (12) and (49). For example, with alpha_bar_t=0.5 and alpha_bar_{t-1}=0.8, Eq. (50) gives sigma_t=0.5, while the posterior std from Eq. (9) is sqrt(0.15) approx 0.387. Since Eq. (51) uses sigma_t to compute log p_theta(x_{t-1}|x_t), the DDPO policy gradient estimate is affected. Please correct Eq. (50) to sqrt((1-alpha_bar_{t-1})/(1-alpha_bar_t)) * sqrt(1 - alpha_bar_t/alpha_bar_{t-1}) or the equivalent standard form.
minor comments (6)
  1. [§2.2.1, heading] The section title 'Foundamentals' should read 'Fundamentals'.
  2. [Introduction] The sentence 'the formulations in many papers often differ from their their corresponding code implementations' contains a duplicated 'their'.
  3. [§2.2.3 and §2.4.3] The phrase 'due do lack of mean and standard deviation' should be 'due to the lack of mean and standard deviation'.
  4. [§2.4.2] In the discrete timestep inference formula, the notation 'with xtN ∼ N(0,I)' is inconsistent with the section's use of yt for the rectified-flow variable; this should likely refer to ytN or be clarified.
  5. [§2.4.3] The word 'upperscript' appears to be a typo for 'superscript'.
  6. [§2.3.1] The word 'acheiving' should be 'achieving'.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the handbook's derivations follow from stated definitions and external sources; the Eq. 39 schedule dependence is a correctness concern, not a circular reduction.

full rationale

The manuscript is a survey/practical handbook rather than a paper making novel predictions with fitted constants. Its unification claim is supported by an explicit general definition (Definition 2.1, xt = a_t x0 + b_t eps) under which rectified flow and flow matching are transparently special schedules, not by renaming the conclusion into the premise. The DDPM/DDIM equivalences, score-matching objective reductions, consistency-model reparameterizations, and the velocity/score relationship in Section 2.7 are derived from the stated forward process via Bayes' rule, chain rule, and algebraic rearrangement; imported results such as Theorem 2.2, the VSD global optimum, and the LCM reparameterization are attributed to external works, with no load-bearing self-citation chain by the authors. The velocity mapping in Eq. (39) does contain an unstated schedule assumption: b' = sqrt(1-alpha_bar) - sqrt(alpha_bar) only holds under a trigonometric time parameterization such as TrigFlow, while for a general variance-preserving schedule b' includes a factor of alpha_bar'. That is a mathematical correctness issue that may invalidate the claimed general diffusion-to-rectified-flow mapping for DDPM-type schedules, but it is not circularity: the mapping is not an input fitted to the paper's own outputs, and it is not justified solely by the authors' prior work. Under the circularity rubric, no derivation step reduces to its own inputs.

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

No free parameters are introduced by this paper; constants such as sigma = 0.5 and epsilon = 0.002 are inherited from cited works. No invented entities are proposed. The central compilation relies on standard diffusion literature results, one unstated schedule-specific assumption in the velocity-mapping proof, and the untested assumption that the presented pseudocode matches real implementations.

assumptions (4)
  • ad hoc to paper The velocity mapping between diffusion and rectified flow trajectories relies on d/dt[sqrt(alpha_bar_t) + sqrt(1 - alpha_bar_t)] = sqrt(1 - alpha_bar_t) - sqrt(alpha_bar_t).
    Stated in Sec 2.6.3 Eq. (39) without justification; holds for the trigonometric (TrigFlow) time parametrization but not for the general VP schedules used elsewhere in the handbook.
  • standard math Reverse-time SDE and ODE formulations (Anderson 1982, Song et al. 2021) are accepted as background.
    Invoked in Sec 2.1 Eqs. (2)-(4) without proof; standard in the diffusion literature.
  • standard math Score matching and Tweedie's formula are taken as given for relating noise, data, and score predictions.
    Used throughout Sec 2.2 and 2.6; standard results.
  • domain assumption The handbook assumes its pseudocode and equations align with actual code implementations of the covered methods.
    This is the paper's stated purpose (abstract, Sec 1) but no code or empirical validation is provided; the assumption is load-bearing for the practical-guide claim.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Generative Diffusion Modeling: A Practical Handbook." pith.science (2026). https://pith.science/paper/AWLNMFHZ

@misc{pith2026241217162,
  author       = {Pith},
  title        = {Pith review of: Generative Diffusion Modeling: A Practical Handbook},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/AWLNMFHZ}},
  note         = {Machine review of arXiv:2412.17162}
}
read the original abstract

This handbook offers a unified perspective on diffusion models, encompassing diffusion probabilistic models, score-based generative models, consistency models, rectified flow, and related methods. By standardizing notations and aligning them with code implementations, it aims to bridge the "paper-to-code" gap and facilitate robust implementations and fair comparisons. The content encompasses the fundamentals of diffusion models, the pre-training process, and various post-training methods. Post-training techniques include model distillation and reward-based fine-tuning. Designed as a practical guide, it emphasizes clarity and usability over theoretical depth, focusing on widely adopted approaches in generative modeling with diffusion models.

Figures

Figures reproduced from arXiv: 2412.17162 by the authors.

Figure 1
Figure 1. Relationship of v-prediction in diffusion and InstaFlow-prediction in rectified flow The InstaFlow-prediction [Liu et al., 2023b] follows the previously introduced rectified flow (RF) method. In InstaFlow-prediction, the method works on a conjugate space (yt , v y θ ) T t=0 ∈ 19 [PITH_FULL_IMAGE:figures/full_fig_p020_1.png] view at source ↗
Figure 2
Figure 2. for a vector representation of the variables [PITH_FULL_IMAGE:figures/full_fig_p024_2.png] view at source ↗
Figure 3
Figure 3. The “triangular”-formula in diffusion model for three parameterization: ϵ-prediction, x-prediction and v-prediction [PITH_FULL_IMAGE:figures/full_fig_p026_3.png] view at source ↗
Figures from the paper (9 more)
Figure 4
Figure 4. Figure 4: The “triangular”-formula in rectified flow model for three parameterization: ϵ￾prediction, x-prediction and v-prediction. 25 [PITH_FULL_IMAGE:figures/full_fig_p026_4.png]
Figure 5
Figure 5. Figure 5: ϵ-prediction loss along the diffusion trajectory [PITH_FULL_IMAGE:figures/full_fig_p027_5.png]
Figure 6
Figure 6. Figure 6: x-prediction loss along the diffusion trajectory. 26 [PITH_FULL_IMAGE:figures/full_fig_p027_6.png]
Figure 7
Figure 7. Figure 7: Visualization of v-prediction (left) and reverse ¯v-prediction (right) in diffusion pro￾cess. 2.6.2 v-prediction in Diffusion Model Apart from ϵ-prediction, there is also a commonly used v-prediction. Eq. (6) can be rewritten in angular coordinate as: xt = x0 sin ϕt + …
Figure 8
Figure 8. Figure 8: v-prediction loss along the RF trajectory. as v y t . We also call it InstaFlow-prediction. The diagram of v-prediction loss for training the RF model is shown as [PITH_FULL_IMAGE:figures/full_fig_p029_8.png]
Figure 9
Figure 9. Figure 9: Consistency loss along the diffusion trajectory. Based on the “triangular”-formula relationship, we show a diagram of consistency loss in [PITH_FULL_IMAGE:figures/full_fig_p030_9.png]
Figure 10
Figure 10. Figure 10: Overview of progressive distillation. Figure is adapted from [Salimans and Ho, 2022]. As illustrated in [PITH_FULL_IMAGE:figures/full_fig_p034_10.png]
Figure 11
Figure 11. Figure 11: Overview of distribution matching distillation. Figure is adapted from [Yin et al., 2024a]. Variational Score Distillation Loss. Similar as in [Wang et al., 2024c], the variational score distillation (VSD) loss is designed to minimize the Kullback-Leibler (KL) of real…
Figure 12
Figure 12. Figure 12: Overview of adversarial diffusion distillation. Figure is adapted from [Sauer et al., 2023]. As displayed in [PITH_FULL_IMAGE:figures/full_fig_p039_12.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

45 extracted references · 10 canonical work pages

  1. [1]

    Albergo, M. S. and Vanden-Eijnden, E. (2022). Building normalizing flows with stochastic interpolants. arXiv preprint arXiv:2209.15571

    arXiv 2022

  2. [2]

    Anderson, B. D. (1982). Reverse-time diffusion equation models. Stochastic Processes and their Applications , 12(3):313--326

    1982

  3. [3]

    Black, K., Janner, M., Du, Y., Kostrikov, I., and Levine, S. (2023). Training diffusion models with reinforcement learning. arXiv preprint arXiv:2305.13301

    arXiv 2023

  4. [4]

    Brooks, T., Peebles, B., Holmes, C., DePue, W., Guo, Y., Jing, L., Schnurr, D., Taylor, J., Luhman, T., Luhman, E., et al. (2024). Video generation models as world simulators. 2024. URL https://openai. com/research/video-generation-models-as-world-simulators , 3

    work page 2024

  5. [5]

    T., Klasky, M

    Chung, H., Kim, J., Mccann, M. T., Klasky, M. L., and Ye, J. C. (2022). Diffusion posterior sampling for general noisy inverse problems. arXiv preprint arXiv:2209.14687

    arXiv 2022

  6. [6]

    Clark, K., Vicol, P., Swersky, K., and Fleet, D. J. (2023). Directly fine-tuning diffusion models on differentiable rewards. arXiv preprint arXiv:2309.17400

    arXiv 2023

  7. [7]

    Ho, J., Jain, A., and Abbeel, P. (2020). Denoising diffusion probabilistic models. Advances in neural information processing systems , 33:6840--6851

    2020

  8. [8]

    and Salimans, T

    Ho, J. and Salimans, T. (2022). Classifier-free diffusion guidance. arXiv preprint arXiv:2207.12598

    arXiv 2022

Show all 45 references
  1. [9]

    and Dayan, P

    Hyv \"a rinen, A. and Dayan, P. (2005). Estimation of non-normalized statistical models by score matching. Journal of Machine Learning Research , 6(4)

    2005

  2. [10]

    Karras, T., Aittala, M., Aila, T., and Laine, S. (2022). Elucidating the design space of diffusion-based generative models. Advances in neural information processing systems , 35:26565--26577

    2022

  3. [11]

    Kingma, D. P. (2013). Auto-encoding variational bayes. arXiv preprint arXiv:1312.6114

    2013 arXiv

  4. [12]

    T., Ben-Hamu, H., Nickel, M., and Le, M

    Lipman, Y., Chen, R. T., Ben-Hamu, H., Nickel, M., and Le, M. (2022). Flow matching for generative modeling. arXiv preprint arXiv:2210.02747

    2022 arXiv

  5. [13]

    Liu, H., Chen, Z., Yuan, Y., Mei, X., Liu, X., Mandic, D., Wang, W., and Plumbley, M. D. (2023a). Audioldm: Text-to-audio generation with latent diffusion models. arXiv preprint arXiv:2301.12503

    2023 arXiv

  6. [14]

    Liu, X., Gong, C., and Liu, Q. (2022). Flow straight and fast: Learning to generate and transfer data with rectified flow. arXiv preprint arXiv:2209.03003

    2022 arXiv

  7. [15]

    Liu, X., Zhang, X., Ma, J., Peng, J., et al. (2023b). Instaflow: One step is enough for high-quality diffusion-based text-to-image generation. In The Twelfth International Conference on Learning Representations

    2023

  8. [16]

    and Song, Y

    Lu, C. and Song, Y. (2024). Simplifying, stabilizing and scaling continuous-time consistency models. arXiv preprint arXiv:2410.11081

    2024 arXiv

  9. [17]

    Luo, S., Tan, Y., Huang, L., Li, J., and Zhao, H. (2023). Latent consistency models: Synthesizing high-resolution images with few-step inference. arXiv preprint arXiv:2310.04378

    2023 arXiv

  10. [18]

    Meng, C., Rombach, R., Gao, R., Kingma, D., Ermon, S., Ho, J., and Salimans, T. (2023). On distillation of guided diffusion models. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 14297--14306

    2023

  11. [19]

    B., Kumar, A., Zhang, G., and Levine, S

    Peng, X. B., Kumar, A., Zhang, G., and Levine, S. (2019). Advantage-weighted regression: Simple and scalable off-policy reinforcement learning. arXiv preprint arXiv:1910.00177

    2019 arXiv

  12. [20]

    and Schaal, S

    Peters, J. and Schaal, S. (2007). Reinforcement learning by reward-weighted regression for operational space control. In Proceedings of the 24th international conference on Machine learning , pages 745--750

    2007

  13. [21]

    Polyak, A., Zohar, A., Brown, A., Tjandra, A., Sinha, A., Lee, A., Vyas, A., Shi, B., Ma, C.-Y., Chuang, C.-Y., et al. (2024). Movie gen: A cast of media foundation models. arXiv preprint arXiv:2410.13720

    2024 arXiv

  14. [22]

    T., and Mildenhall, B

    Poole, B., Jain, A., Barron, J. T., and Mildenhall, B. (2022). Dreamfusion: Text-to-3d using 2d diffusion. arXiv preprint arXiv:2209.14988

    2022 arXiv

  15. [23]

    Psenka, M., Escontrela, A., Abbeel, P., and Ma, Y. (2023). Learning a diffusion model policy from rewards via q-score matching. arXiv preprint arXiv:2312.11752

    2023 arXiv

  16. [24]

    Ramesh, A., Dhariwal, P., Nichol, A., Chu, C., and Chen, M. (2022). Hierarchical text-conditional image generation with clip latents. arXiv preprint arXiv:2204.06125 , 1(2):3

    2022 arXiv

  17. [25]

    Z., Lidard, J., Ankile, L

    Ren, A. Z., Lidard, J., Ankile, L. L., Simeonov, A., Agrawal, P., Majumdar, A., Burchfiel, B., Dai, H., and Simchowitz, M. (2024). Diffusion policy policy optimization. arXiv preprint arXiv:2409.00588

    2024 arXiv

  18. [26]

    Rombach, R., Blattmann, A., Lorenz, D., Esser, P., and Ommer, B. (2022). 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

  19. [27]

    and Ho, J

    Salimans, T. and Ho, J. (2022). Progressive distillation for fast sampling of diffusion models. arXiv preprint arXiv:2202.00512

    2022 arXiv

  20. [28]

    Sauer, A., Lorenz, D., Blattmann, A., and Rombach, R. (2023). Adversarial diffusion distillation. arXiv preprint arXiv:2311.17042

    2023 arXiv

  21. [29]

    Sohl-Dickstein, J., Weiss, E., Maheswaranathan, N., and Ganguli, S. (2015). Deep unsupervised learning using nonequilibrium thermodynamics. In International conference on machine learning , pages 2256--2265. PMLR

    2015

  22. [30]

    Song, J., Meng, C., and Ermon, S. (2020a). Denoising diffusion implicit models. arXiv preprint arXiv:2010.02502

    2020 arXiv

  23. [31]

    and Dhariwal, P

    Song, Y. and Dhariwal, P. (2023). Improved techniques for training consistency models. arXiv preprint arXiv:2310.14189

    2023 arXiv

  24. [32]

    Song, Y., Dhariwal, P., Chen, M., and Sutskever, I. (2023). Consistency models. arXiv preprint arXiv:2303.01469

    2023 arXiv

  25. [33]

    and Ermon, S

    Song, Y. and Ermon, S. (2019). Generative modeling by estimating gradients of the data distribution. Advances in neural information processing systems , 32

    2019

  26. [34]

    and Ermon, S

    Song, Y. and Ermon, S. (2020). Improved techniques for training score-based generative models. Advances in neural information processing systems , 33:12438--12448

    2020

  27. [35]

    P., Kumar, A., Ermon, S., and Poole, B

    Song, Y., Sohl-Dickstein, J., Kingma, D. P., Kumar, A., Ermon, S., and Poole, B. (2020b). Score-based generative modeling through stochastic differential equations. arXiv preprint arXiv:2011.13456

    2020 arXiv

  28. [36]

    P., Kumar, A., Ermon, S., and Poole, B

    Song, Y., Sohl-Dickstein, J., Kingma, D. P., Kumar, A., Ermon, S., and Poole, B. (2021). Score-based generative modeling through stochastic differential equations. In International Conference on Learning Representations

    2021

  29. [37]

    Vincent, P. (2011). A connection between score matching and denoising autoencoders. Neural computation , 23(7):1661--1674

    2011

  30. [38]

    Wang, F.-Y., Huang, Z., Shi, X., Bian, W., Song, G., Liu, Y., and Li, H. (2024a). Animatelcm: Accelerating the animation of personalized diffusion models and adapters with decoupled consistency learning. arXiv preprint arXiv:2402.00769

    2024 arXiv

  31. [39]

    Wang, Y., Wang, X., Chen, Z., Wang, Z., Sun, F., and Zhu, J. (2024b). Vidu4d: Single generated video to high-fidelity 4d reconstruction with dynamic gaussian surfels. arXiv preprint arXiv:2405.16822

    2024 arXiv

  32. [40]

    Wang, Z., Lu, C., Wang, Y., Bao, F., Li, C., Su, H., and Zhu, J. (2024c). Prolificdreamer: High-fidelity and diverse text-to-3d generation with variational score distillation. Advances in Neural Information Processing Systems , 36

    2024

  33. [41]

    Weng, L. (2021). What are diffusion models? lilianweng.github.io

    2021

  34. [42]

    Xu, J., Liu, X., Wu, Y., Tong, Y., Li, Q., Ding, M., Tang, J., and Dong, Y. (2024). Imagereward: Learning and evaluating human preferences for text-to-image generation. Advances in Neural Information Processing Systems , 36

    2024

  35. [43]

    Yin, T., Gharbi, M., Park, T., Zhang, R., Shechtman, E., Durand, F., and Freeman, W. T. (2024a). Improved distribution matching distillation for fast image synthesis. arXiv preprint arXiv:2405.14867

    2024 arXiv

  36. [44]

    T., and Park, T

    Yin, T., Gharbi, M., Zhang, R., Shechtman, E., Durand, F., Freeman, W. T., and Park, T. (2024b). One-step diffusion with distribution matching distillation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 6613--6623

    2024

  37. [45]

    A., Shechtman, E., and Wang, O

    Zhang, R., Isola, P., Efros, A. A., Shechtman, E., and Wang, O. (2018). The unreasonable effectiveness of deep features as a perceptual metric. In Proceedings of the IEEE conference on computer vision and pattern recognition , pages 586--595

    2018

Pith tools

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