Visual-CoG: Stage-Aware Reinforcement Learning with Chain of Guidance for Text-to-Image Generation

Bo Zheng; Haoxiang Shi; Jun Song; Mingyang Han; Peng Chen; Pi Bu; Runzhou Zhao; Xuan Zhang; Yang Yao; Yaqi Li

REVIEW 3 major objections 5 minor 8 references

Visual-CoG: Stage-Aware Reinforcement Learning with Chain of Guidance for Text-to-Image Generation

T0 review · 3 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 A stage-aware reward chain—semantic reasoning, process refining, and outcome evaluation—improves text-to-image alignment on multi-attribute and ambiguous prompts, raising GenEval overall from 68.29 to 83.86.

desk verdict Stage-aware reward decomposition is a good idea, but the two central reward equations are sign-inverted as written and the teacher model is unspecified; major revision before the results can be trusted. read the letter →

arxiv 2508.18032 v2 pith:4PHA7WA2 submitted 2025-08-25 cs.CV

Yaqi Li , Peng Chen , Mingyang Han , Pi Bu , Haoxiang Shi , Runzhou Zhao , Yang Yao , Xuan Zhang

show 2 more authors

Jun Song Bo Zheng

This is my paper · ORCID

classification cs.CV

keywords text-to-imagegenerationreinforcementlearningstage-awarerewardschainofguidancesemanticreasoningautoregressiveunifiedmodelsmasktokenpredictionvisualcognitionbenchmark

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

Visual-CoG is a training strategy for autoregressive text-to-image models that replaces a single final reward with three stage-aware rewards: semantic reasoning, process refining, and outcome evaluation. The paper argues that final-only reward signals, as used by prior CoT and RL text-to-image works, make it hard to tell which stage of generation helped and therefore lead to suboptimal policies. By rewarding each stage immediately, Visual-CoG lets the model allocate credit to early interpretation and intermediate refinement, not just the finished picture. On the GenEval benchmark, the method lifts the Show-o baseline from 68.29 to 83.86 overall, and on the authors' new VisCog-Bench it raises overall alignment from 58.07 to 77.50. If the result holds, stage-aware rewards are a practical way to make text-to-image models follow multi-attribute and ambiguous prompts.

What carries the argument

The machinery is a stage-aware reward chain, $R = R_r + R_p + R_o$, plugged into a GRPO-style clipped policy gradient. $R_r$ is a contrastive semantic-reasoning reward: generate two images from the original prompt and from a reasoned prompt under a fixed seed, then take the difference of their outcome scores. $R_p$ is a process-refining reward defined as an exponential norm between the policy's reconstructed masked patch distribution and a teacher model's preferred distribution $p_\tau$. $R_o$ is a rule-based outcome reward combining counting, color, and spatial checks, via an open-vocabulary detector and a vision-language classifier, with a holistic HPS aesthetic score. This per-stage reward design is what allows credit assignment to early reasoning and intermediate refinement steps rather than only to the final rendered image.

What would settle it

Train Visual-CoG with the teacher model replaced by the policy itself, so $p_\tau = p_\theta$ in Eq. 5: the authors state it would be meaningless for the policy to recover its own masked output, so the process-refining reward should stop contributing and GenEval should drop toward the $R_r+R_o$ ablation's 76.23. If performance stays near 83.86, the process-refining reward is not doing the claimed work. A simpler check: compute Eq. 5 as printed, without a negative sign, and verify whether high-reward samples are closer to or farther from the teacher distribution.

Watch

Extended reading notes

Core claim

The central claim is that decomposing text-to-image synthesis into a chain of guidance and giving an immediate reward at each link improves alignment more than optimizing only the final image. The three stages are: (1) semantic reasoning, where a language model rewrites the prompt and the reward is the outcome-score difference between images generated from the reasoned prompt and the original prompt; (2) process refining, where a masked patch reconstruction step is rewarded for matching a teacher model's preferred distribution; and (3) outcome evaluation, where rule-based spatial, counting, and color checks plus an aesthetic preference score score the final image. The rewards are summed into $R = R_r + R_p + R_o$ and optimized with a clipped policy-gradient objective. In the paper's experiments, this raises Show-o from 68.29 to 83.86 on GenEval, reaches 78.92 on T2I-CompBench Color and 43.71 on Spatial, and lifts VisCog-Bench overall from 58.07 to 77.50.

Load-bearing premise

The process-refining reward assumes that the teacher model $\tau$ provides a preferred distribution that is genuinely better to imitate, yet the paper never specifies $\tau$'s architecture, training data, or origin; if $\tau$ is not a better generative distribution, or if the sign in Eq. 5 is wrong as printed, the intermediate guidance is undefined or actively pushes the policy to diverge.

Editorial extensions

If this is right

On GenEval, stage-aware rewards take Show-o from 68.29 to 83.86 overall, with the position subtask jumping from 31.03 to 79.00 and counting from 66.25 to 80.94.
On T2I-CompBench, Visual-CoG sets the paper's best numbers of 78.92 on color and 43.71 on spatial among the compared models.
On VisCog-Bench, overall alignment rises from 58.07 to 77.50, with the unusual-position subtask improving from 28.75 to 77.50.
Ablations attribute distinct roles to each reward: $R_r$ lifts position by 6.99 points, $R_p$ lifts counting by 7.96 points, and $R_o$ lifts color by 9.53 points; removing any stage lowers overall from 83.86.

Reading between the lines

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

An implication not drawn in the paper: the semantic-reasoning reward is a contrast between two images scored by the same outcome evaluator, so if that evaluator has systematic biases, the reasoning stage may learn to exploit the scorer instead of resolving genuine ambiguity.
The paper leaves unresolved how the teacher distribution $p_\tau$ is obtained; without a specified teacher, the process-refining reward is not reproducible from Eq. 5 as written, since the exponential norm with no negative sign rewards disagreement rather than closeness.
Because VisCog-Bench's unusual-position, unusual-composition, and unusual-color subtasks each contain only 20 prompts, the reported 19-point overall gain on that benchmark likely carries nontrivial sampling uncertainty; extending those subtasks would test whether the effect is stable.

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.

Desk Editor's Note

read the letter

Two things to know about this paper. The stage-aware reward decomposition (semantic reasoning, process refining, outcome evaluation) is a sensible idea, and the VisCog-Bench, though small, targets a real evaluation gap. But the central equations as written invert the optimization: Eq. 5 rewards divergence from the teacher model, and Eq. 7 rewards counting error. That is not a detail; it is the mechanism.

On the positive side, the paper correctly identifies a limitation of final-only rewards in RL for T2I, and the three-stage credit assignment is a plausible remedy. The ablations in Table 4 show each reward component helps, which is interesting if the implementation actually uses the correct signs. The benchmark of 100 prompts for unusual position/composition/color and reasoning is a useful idea, though small.

Now the soft spots, in rough order of severity.

First, Eq. 5 defines Rp = exp(||G(pθ)-G(pτ)||_p). Maximizing that pushes the policy away from the teacher distribution, the opposite of "process refining." Eq. 7 defines Rn as the average of exp(|count - target|/τ), which increases with counting error. The text says both should penalize deviation. Unless the exponential signs are typos (exp(-...)), the policy is being rewarded for doing worse. I think they are likely typos, but as presented the method is not coherent.

Second, the teacher model τ that defines the preferred distribution pτ is never described: no architecture, training data, or source. Eq. 5 cannot be computed or reproduced. That is a load-bearing gap.

Third, the semantic reasoning reward Rr = Ro(P') - Ro(P) uses the same outcome evaluator that appears in the final reward. That makes the "reasoning" signal dependent on the same metric the policy is trained to exploit; it is not an independent measure of reasoning quality. This is a legitimate concern, though less damning than the sign errors.

Fourth, the empirical section has no error bars, no seeds, and omits T2I-R1, a closely related CoT+RL baseline cited in the paper. Also, the reproduction checklist at the end asserts "yes" to a code appendix and statistical tests that do not exist in the manuscript. That is not a great look.

Bottom line: the idea is worth a second look, but the paper as written cannot be relied on. If the sign errors are just typos and the teacher model is specified, the method could be salvaged with substantial revision. As is, I would tell the authors to fix the equations, define τ, add the missing baseline, and release code.

If you are the editor, I would send it to peer review rather than desk-reject, because the fixes are identifiable and the benchmark has value; but the reviewers should be instructed to check the equations carefully. I would not cite this version.

Referee Report

3 major / 5 minor

Summary. The paper proposes Visual-CoG, a stage-aware reinforcement learning framework for text-to-image generation. Generation is decomposed into three stages—semantic reasoning, process refining, and outcome evaluation—with a distinct reward for each: Rr in Eq. (4), Rp in Eq. (5), and Ro in Eqs. (6)–(9). The authors train Show-o with this staged reward and report large gains on GenEval, T2I-CompBench, and a newly introduced VisCog-Bench. The paper also introduces VisCog-Bench, a 100-prompt benchmark with unusual position, composition, color, and reasoning subtasks. The central claim is that providing immediate, stage-aware rewards improves alignment on multi-attribute and ambiguous prompts beyond a single final reward.

Significance. If the method worked as described, the claimed improvements are substantial and the staged-reward idea is a plausible and interesting direction for autoregressive text-to-image alignment. The paper also contributes a benchmark aimed at unusual and reasoning-demanding prompts, which could be useful to the community. However, the central reward equations contain two sign inversions that make the optimization objective as written the opposite of the stated goal, and the teacher model in the process-refining stage is never specified. These issues are not cosmetic: they undermine the mechanism that the paper's results are attributed to. The strengths are the clear problem framing and the breadth of evaluation; the weaknesses are the mathematical errors and the underspecified teacher model, which together prevent the reported results from being reproduced or interpreted as supporting the proposed method.

major comments (3)

[§3.2, Eq. (5)] The process-refining reward is defined as Rp = exp(||G(pθ) - G(pτ)||_p). This is monotonically increasing in the discrepancy between the policy reconstruction and the teacher reconstruction. Maximizing Rp therefore encourages the policy to move away from the teacher distribution, which is the opposite of the stated 'process refining' objective. As written, this reward cannot explain the counting improvement attributed to Rp in Table 4. Please clarify whether a negative exponent (or a minimization objective) was intended, and define the teacher model τ and its preferred distribution pτ—the architecture, training data, and source are not given anywhere in the paper, so Eq. (5) is not computable or reproducible as written.
[§3.2, Eq. (7)] The counting reward is defined as Rn = (1/N_n) Σ_i exp(|En(D(I_{e,i})) - y*_{in}|/τ). This function increases with the absolute counting error: an image with the wrong number of objects receives a higher reward than one with the correct count. The accompanying text states that the design 'penalizes larger deviations more severely,' but the formula rewards them. This inversion corrupts the outcome reward Ro in Eq. (9) and, through Eq. (4), also distorts the semantic reasoning reward Rr. If the intended formula is exp(-|...|/τ), the equations must be corrected and all experiments dependent on them must be rerun before the reported counting results can be attributed to this reward.
[§3.2, Eq. (4) and §4.4] The semantic reasoning reward is a difference of the outcome reward Ro for the reasoning prompt and the original prompt. Because Ro is itself the reward being optimized, Rr measures whether the reasoning prompt produces a better final image under the same proxy, not whether the model performed any form of reasoning. The VisCog-Bench reasoning evaluation in Table 5 uses Qwen2.5-VL as an automatic judge, and the benchmark has only 100 prompts with no reported confidence intervals or human-judgment agreement. The claim that the method improves 'reasoning' is therefore not yet independently established; a reasoning-specific reward or a more controlled evaluation with inter-annotator agreement would be needed.

minor comments (5)

[§3.2, Eq. (10)] The sentence introducing Eq. (10) says 'the loss is defined in Eq. 2: R = Rr + Rp + Ro,' but Eq. (2) is the policy-gradient loss L(θ), not a reward. Please separate the definition of the scalar reward R from the policy-gradient objective.
[§3.2, Eq. (5)] In Eq. (5), the symbol p is used both for the norm order and for the policy distributions pθ and pτ, which is confusing. Please use a different letter for the norm, e.g., ||·||_q.
[Table 4] The claimed per-reward improvements (6.99% for position from Rr, 7.96% for counting from Rp, and 9.53% for color from Ro) are not directly readable from the table because each ablation row combines multiple rewards. Please specify the exact row comparisons used to compute these numbers.
[§4.1 and §4.4] VisCog-Bench contains only 100 prompts, and the human evaluation in Table 5 is reported without confidence intervals or statistical significance tests. Please provide variance estimates or significance testing.
[Reproduction Checklist] The checklist states that all code and data will be released, but no code or data appendix is included in the submission. Please clarify the availability of the promised artifacts at submission time.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found: the reported improvements are measured on external benchmarks and the stage rewards, while not all fully independent, do not reduce the central claim to its inputs by construction.

full rationale

The paper's central claim is empirical: adding the three stage rewards to Show-o improves GenEval, T2I-CompBench, and VisCog-Bench. None of these evaluations is identical to the training reward. GenEval and T2I-CompBench use their own published protocols, and VisCog-Bench uses GenEval metrics plus a Qwen2.5-VL evaluator for reasoning prompts, while the training reward stack uses GroundingDINO, CLIP, and HPS. The semantic-reasoning reward Rr = Ro(Mt2i(P')) - Ro(Mt2i(P)) is indeed a contrast of the outcome reward, so calling it a separate 'semantic' signal is somewhat misleading, but this is a reward-shaping/baseline design choice rather than a circular derivation: it does not make the benchmark outcomes equal to the training objective, and the reported gains are measured on held-out tasks. The process-refining reward Rp and counting reward Rn contain sign inversions (Eqs. 5 and 7 reward larger discrepancies and larger count deviations), and the teacher model tau is never specified; these are correctness and reproducibility defects, not circularity, because they do not redefine the evaluation metric as the training signal. There is no load-bearing self-citation chain, no imported uniqueness theorem, and no fitted parameter renamed as a prediction. The proposed VisCog-Bench is new, but evaluating one's own method on a new benchmark is not circular unless the benchmark metric is the training objective, which it is not. Overall, the derivation chain is self-contained and the central empirical claims stand independently of any definitional shortcut.

Assumptions & free parameters 2 free parameters · 3 assumptions · 0 invented entities

The central method is built on three reward signals. None of them is an externally benchmarked quantity: Rr is defined through Ro, Rp depends on an unspecified teacher, and Ro uses unvalidated proxy models. The ledger is deliberately sparse because the main burden is not in fitted constants but in undefined components and sign-defective reward definitions; see red flags.

free parameters (2)

τ in counting reward (Eq. 7) = not specified
The exponential scaling factor for counting deviation is central to Rn, but no value is given and the sign is inverted as written.
Masking schedule for process refining = not specified
The iterative masked patch reconstruction process requires a masking probability and refinement steps; neither is specified in Sec. 3.2.

assumptions (3)

ad hoc to paper Teacher model τ provides a preferred distribution pτ for masked image tokens.
Introduced in Sec. 3.2 to define Rp, but never described. The reward is not computable without it.
domain assumption Rule-based outcome evaluator (GroundingDINO, CLIP, HPS) is a valid proxy for prompt-image alignment.
All three rewards rely on detector, CLIP, and HPS scores as ground truth; no calibration or error analysis is provided.
domain assumption Masked-patch reconstruction quality is a valid signal for later image quality.
The paper assumes that rewarding agreement with the teacher on masked reconstructions improves final generation, but supplies no evidence connecting intermediate reconstruction accuracy to final fidelity.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Visual-CoG: Stage-Aware Reinforcement Learning with Chain of Guidance for Text-to-Image Generation." pith.science (2026). https://pith.science/paper/4PHA7WA2

@misc{pith2026250818032,
  author       = {Pith},
  title        = {Pith review of: Visual-CoG: Stage-Aware Reinforcement Learning with Chain of Guidance for Text-to-Image Generation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/4PHA7WA2}},
  note         = {Machine review of arXiv:2508.18032}
}

read the original abstract

Despite the promising progress of recent autoregressive models in text-to-image (T2I) generation, their ability to handle multi-attribute and ambiguous prompts remains limited. To address these limitations, existing works have applied chain-of-thought (CoT) to enable stage-aware visual synthesis and employed reinforcement learning (RL) to improve reasoning capabilities. However, most models provide reward signals only at the end of the generation stage. This monolithic final-only guidance makes it difficult to identify which stages contribute positively to the final outcome and may lead to suboptimal policies. To tackle this issue, we propose a Visual-Chain of Guidance (Visual-CoG) paradigm consisting of three stages: semantic reasoning, process refining, and outcome evaluation, with stage-aware rewards providing immediate guidance throughout the image generation pipeline. We further construct a visual cognition benchmark, VisCog-Bench, which comprises four subtasks to evaluate the effectiveness of semantic reasoning. Comprehensive evaluations on GenEval, T2I-CompBench, and the proposed VisCog-Bench show improvements of 15%, 5%, and 19%, respectively, demonstrating the superior performance of the proposed Visual-CoG. We will release all the resources soon.

Figures

Figures reproduced from arXiv: 2508.18032 by the authors.

**Figure 2.** Visualization of the image generation pipeline of Visual-Chain of Guidance [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗

**Figure 3.** An overview of Visual-CoG framework, which comprises three key stages: semantic reasoning, process refining, and [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗

Figures from the paper (2 more)

**Figure 4.** Figure 4: Qualitative results from GenEval for various attributes such as composition, color, position and counting. [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]

**Figure 5.** Figure 5: Qualitative results of image generation for prompts [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

8 extracted references · 5 canonical work pages

[1]

Includes a conceptual outline and/or pseudocode descrip- tion of AI methods introduced (yes/partial/no/NA) yes 1.2

General Paper Structure 1.1. Includes a conceptual outline and/or pseudocode descrip- tion of AI methods introduced (yes/partial/no/NA) yes 1.2. Clearly delineates statements that are opinions, hypoth- esis, and speculation from objective facts and results (yes/no) yes 1.3. Provides well-marked pedagogical references for less- familiar readers to gain bac...

work page
[2]

Does this paper make theoretical contributions? (yes/no) yes If yes, please address the following points: 2.2

Theoretical Contributions 2.1. Does this paper make theoretical contributions? (yes/no) yes If yes, please address the following points: 2.2. All assumptions and restrictions are stated clearly and formally (yes/partial/no) yes 2.3. All novel claims are stated formally (e.g., in theorem statements) (yes/partial/no) yes 2.4. Proofs of all novel claims are ...

work page
[3]

Does this paper rely on one or more datasets? (yes/no) yes If yes, please address the following points: 3.2

Dataset Usage 3.1. Does this paper rely on one or more datasets? (yes/no) yes If yes, please address the following points: 3.2. A motivation is given for why the experiments are con- ducted on the selected datasets (yes/partial/no/NA) yes 3.3. All novel datasets introduced in this paper are included in a data appendix (yes/partial/no/NA) yes 3.4. All nove...

work page
[4]

In European Conference on Computer Vi- sion, 423–439

Compositional visual generation with composable dif- fusion models. In European Conference on Computer Vi- sion, 423–439. Springer. Liu, S.; Zeng, Z.; Ren, T.; Li, F.; Zhang, H.; Yang, J.; Jiang, Q.; Li, C.; Yang, J.; Su, H.; et al. 2024. Grounding dino: Marrying dino with grounded pre-training for open-set ob- ject detection. In European Conference on Co...

arXiv 2024
[8]

Does this paper include computational experiments? (yes/no) yes If yes, please address the following points: 4.2

Computational Experiments 4.1. Does this paper include computational experiments? (yes/no) yes If yes, please address the following points: 4.2. This paper states the number and range of values tried per (hyper-) parameter during development of the pa- per, along with the criterion used for selecting the final parameter setting (yes/partial/no/NA) yes 4.3...

work page
[755]

Liu, N.; Li, S.; Du, Y .; Torralba, A.; and Tenenbaum, J

Springer. Liu, N.; Li, S.; Du, Y .; Torralba, A.; and Tenenbaum, J. B

work page
[2022]

In Proceedings of the IEEE/CVF conference on computer vi- sion and pattern recognition, 11315–11325

Maskgit: Masked generative image transformer. In Proceedings of the IEEE/CVF conference on computer vi- sion and pattern recognition, 11315–11325. Chen, J.; Yu, J.; Ge, C.; Yao, L.; Xie, E.; Wu, Y .; Wang, Z.; Kwok, J.; Luo, P.; Lu, H.; et al. 2023. Pixart-alpha: Fast training of diffusion transformer for photorealistic text-to- image synthesis. arXiv pre...

arXiv 2023
[2024]

In Forty-first international conference on machine learning

Scaling rectified flow transformers for high-resolution image synthesis. In Forty-first international conference on machine learning. Fang, R.; Duan, C.; Wang, K.; Huang, L.; Li, H.; Yan, S.; Tian, H.; Zeng, X.; Zhao, R.; Dai, J.; et al. 2025. Got: Un- leashing reasoning capability of multimodal large language model for visual generation and editing. arXi...

arXiv 2025

Pith tools

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

[1] [1]

Includes a conceptual outline and/or pseudocode descrip- tion of AI methods introduced (yes/partial/no/NA) yes 1.2

General Paper Structure 1.1. Includes a conceptual outline and/or pseudocode descrip- tion of AI methods introduced (yes/partial/no/NA) yes 1.2. Clearly delineates statements that are opinions, hypoth- esis, and speculation from objective facts and results (yes/no) yes 1.3. Provides well-marked pedagogical references for less- familiar readers to gain bac...

work page

[2] [2]

Does this paper make theoretical contributions? (yes/no) yes If yes, please address the following points: 2.2

Theoretical Contributions 2.1. Does this paper make theoretical contributions? (yes/no) yes If yes, please address the following points: 2.2. All assumptions and restrictions are stated clearly and formally (yes/partial/no) yes 2.3. All novel claims are stated formally (e.g., in theorem statements) (yes/partial/no) yes 2.4. Proofs of all novel claims are ...

work page

[3] [3]

Does this paper rely on one or more datasets? (yes/no) yes If yes, please address the following points: 3.2

Dataset Usage 3.1. Does this paper rely on one or more datasets? (yes/no) yes If yes, please address the following points: 3.2. A motivation is given for why the experiments are con- ducted on the selected datasets (yes/partial/no/NA) yes 3.3. All novel datasets introduced in this paper are included in a data appendix (yes/partial/no/NA) yes 3.4. All nove...

work page

[4] [4]

In European Conference on Computer Vi- sion, 423–439

Compositional visual generation with composable dif- fusion models. In European Conference on Computer Vi- sion, 423–439. Springer. Liu, S.; Zeng, Z.; Ren, T.; Li, F.; Zhang, H.; Yang, J.; Jiang, Q.; Li, C.; Yang, J.; Su, H.; et al. 2024. Grounding dino: Marrying dino with grounded pre-training for open-set ob- ject detection. In European Conference on Co...

arXiv 2024

[5] [8]

Does this paper include computational experiments? (yes/no) yes If yes, please address the following points: 4.2

Computational Experiments 4.1. Does this paper include computational experiments? (yes/no) yes If yes, please address the following points: 4.2. This paper states the number and range of values tried per (hyper-) parameter during development of the pa- per, along with the criterion used for selecting the final parameter setting (yes/partial/no/NA) yes 4.3...

work page

[6] [755]

Liu, N.; Li, S.; Du, Y .; Torralba, A.; and Tenenbaum, J

Springer. Liu, N.; Li, S.; Du, Y .; Torralba, A.; and Tenenbaum, J. B

work page

[7] [2022]

In Proceedings of the IEEE/CVF conference on computer vi- sion and pattern recognition, 11315–11325

Maskgit: Masked generative image transformer. In Proceedings of the IEEE/CVF conference on computer vi- sion and pattern recognition, 11315–11325. Chen, J.; Yu, J.; Ge, C.; Yao, L.; Xie, E.; Wu, Y .; Wang, Z.; Kwok, J.; Luo, P.; Lu, H.; et al. 2023. Pixart-alpha: Fast training of diffusion transformer for photorealistic text-to- image synthesis. arXiv pre...

arXiv 2023

[8] [2024]

In Forty-first international conference on machine learning

Scaling rectified flow transformers for high-resolution image synthesis. In Forty-first international conference on machine learning. Fang, R.; Duan, C.; Wang, K.; Huang, L.; Li, H.; Yan, S.; Tian, H.; Zeng, X.; Zhao, R.; Dai, J.; et al. 2025. Got: Un- leashing reasoning capability of multimodal large language model for visual generation and editing. arXi...

arXiv 2025