Pith. sign in

REVIEW 4 major objections 6 minor 48 references

PanoLlama: Generating Endless and Coherent Panoramas with Next-Token-Prediction LLMs

T0 review · 4 major / 6 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read By reusing the trailing tokens of each generated crop as the starting condition for the next, PanoLlama turns a fixed-size visual autoregressive model into a training-free generator of coherent, arbitrarily long panoramas.

desk verdict Token redirection is a neat training-free idea, but the horizontal expansion relies on an unvalidated off-raster assumption that undermines the headline numbers. read the letter →

arxiv 2411.15867 v3 pith:X7ILXCBN submitted 2024-11-24 cs.CV

Teng Zhou , Xiaoyu Zhang , Yongchuan Tang This is my paper · ORCID
classification cs.CV
keywords panoramicimagegenerationnext-tokenpredictionautoregressivemodelstraining-freeextrapolationtokenredirectiontext-to-imagebenchmarkmulti-scalesynthesis
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

PanoLlama argues that panoramic image generation is best modeled as a next-token prediction problem: a panorama is a sequence of crops, and each crop should be generated in the context of everything before it. Rather than train a new model, it takes a pre-trained visual autoregressive image generator (which normally produces one fixed square image) and, after each crop, feeds the last tokens of the current crop back as the prefix for the next crop. This 'token redirection' gives spatial continuity in both horizontal and vertical directions without any fine-tuning. The paper reports that on its own 1,000-prompt benchmark of 512 by 5120 panoramas, the approach beats diffusion-based panorama methods on crop-connection coherence, fidelity, diversity, and aesthetic score, while also enabling mask-free layout control and multi-directional expansion.

What carries the argument

The central object is token redirection, a manipulation of the raster-scan token order produced by the image tokenizer. Because an autoregressive image generator arranges its $p$ image tokens in a fixed grid of $\sqrt{p}$ by $\sqrt{p}$, the trailing boundary tokens of one crop can be fed back as the conditional prefix for the next block, so the model's learned next-token distribution supplies the spatial continuation. The two redirection rules are Eq. 6 for vertical expansion (feed the last $p - r\sqrt{p}$ tokens back, producing $r$ new rows) and Eq. 8 for horizontal expansion (for each row $j$, feed the last $\sqrt{p} - c$ tokens of that row back, producing $c$ new columns per row). Together with the tokenizer decoder from Eq. 10, these rules turn a fixed-square generator into an endless panorama generator; the overlap parameters $r$ and $c$ set the expansion stride and therefore the trade-off between coherence and generation time.

What would settle it

A decisive experiment is to tokenize the left half of a real wide photograph, apply PanoLlama to extend it rightward, and measure how closely the synthesized right half matches the real right half with a perceptual metric such as LPIPS. If the match is good only for uniform textures but fails for scenes with long-range geometry—crowds, facades, repeating patterns—then the redirection strategy would be shown to produce local continuity without the multilevel coherence the paper claims.

Watch

Extended reading notes

Core claim

The central claim is that next-crop prediction can be obtained for free from a fixed-size autoregressive image model by exploiting the raster-scan order of its discrete tokens. For vertical expansion, the next crop starts from the last $p - r\sqrt{p}$ tokens of the previous crop (Eq. 6); for horizontal expansion, each row of the next crop starts from the last $\sqrt{p} - c$ tokens of the corresponding row of the previous crop (Eq. 8). Iterating this process and decoding the concatenated token stream yields a panorama of arbitrary length. The paper's experiments indicate that this simple reuse of the model's own conditional distribution produces smoother crop junctions than heuristic diffusion-based fusion, and the accompanying benchmark of 1,000 prompts across 25 themes is offered as a standardized test bed. The paper states openly that the method approximates global dependencies with a partial token context rather than fully conditioning on the entire panorama.

Load-bearing premise

The whole method rests on the assumption that when the token sequence of a generated image is cut short and its tail is fed back as a prompt, the model genuinely continues drawing the adjacent part of the scene rather than merely producing plausible-looking but structurally disconnected pixels.

Editorial extensions

If this is right

  • Any future fixed-size autoregressive image generator becomes a panoramic generator with no retraining, so the redirection trick is a drop-in capability for the whole class of models.
  • Because the redirection is row-wise and column-wise, the same loop supports horizontal, vertical, and two-dimensional multi-scale panoramas from a single pipeline.
  • The overlap parameters $r$ and $c$ give a direct knob for the quality/efficiency tradeoff: smaller strides mean more overlap and higher local coherence but longer generation.
  • The main results are measured on a new 1,000-prompt benchmark spanning 25 themes, so future panorama methods can be compared on a standardized test bed rather than small hand-picked prompt sets.
  • The method's token-prefix conditioning also enables mask-free layout control and multi-guidance generation, letting different regions be steered by different prompts or images within one panorama.

Reading between the lines

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

  • Not tested in the paper, but a natural next stress test is very long extrapolation (well beyond 10x): with each step reusing only a bounded context, any slow drift in style or content would accumulate differently from what a single 10x benchmark shows.
  • The benchmark's coherence metrics compare adjacent non-overlapping crops; an inference from that design is that local junction quality is measured directly, while global long-range consistency is only indirectly reflected, so future work could add explicit long-range layout agreement tests.
  • The redirection idea is stated for one raster-scan generator family; whether it transfers to hierarchical or masked next-scale image generators is open, because those models do not expose a single per-row token ordering.
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

4 major / 6 minor

Summary. PanoLlama re-frames panoramic image generation as next-crop prediction using pre-trained autoregressive image models. The method is training-free: it first generates a fixed-size block with LlamaGen, then redirects generation by feeding the trailing tokens of the previous block as context to generate the next block vertically or horizontally, concatenates all token sequences, and decodes them into an arbitrarily larger panorama. The authors report state-of-the-art coherence, fidelity, and aesthetics on a newly constructed 1,000-prompt benchmark, and demonstrate additional capabilities such as multi-scale expansion, mask-free layout control, and multi-guidance synthesis. The paper is clearly written and the idea is simple and attractive, but the quantitative evidence does not yet support the central horizontal-expansion claim.

Significance. The autoregressive next-token-prediction paradigm for panoramic image generation is a potentially valuable departure from the joint-diffusion pipeline, and the training-free token-redirection idea is elegant if it works. The paper's strengths include a concrete algorithm (Algorithm 1), a code release, a new benchmark, and several applications. However, the central mechanism for horizontal expansion relies on the model predicting off-raster tokens, and the evaluation mixes absolute and relative comparisons across different base generators. If these issues are resolved, the contribution could be significant; in the current form, the reported state-of-the-art claims are not yet established.

major comments (4)
  1. [Sec. 3.2, Eq. (8)] In horizontal expansion, the last sqrt(p)-c tokens of row j are used as the context to generate c new tokens that are treated as a horizontal extension of that row. In the raster-scan token order used by LlamaGen and other VQGAN-based autoregressive models, the token that follows the end of a row in training is the first token of the next row, not a token to the right. The model therefore receives an out-of-distribution context, and its output is drawn from a distribution trained to predict vertically adjacent content. The paper supplies no quantitative or qualitative ablation that separates the effect of this off-raster redirection from the base model's own continuity. Since the headline results (w'=5120, 10x horizontal expansion) are based on this mechanism, the central claim is not yet established. Please provide evidence that the predicted tokens are horizontally coherent, for example by measuring agreement with ground-truth right-neighbor tokens, comparing against a raster-order baseline, or using human evaluation.
  2. [Sec. 4.1, Table 1] The coherence scores (LPIPS, DISTS, TV, SSIM) are absolute values computed on outputs from different base generators: PanoLlama uses LlamaGen while all diffusion baselines use SD or SDXL. As the paper itself states, PIG performance is closely tied to the reference model, so a direct comparison of absolute coherence scores can confound the PIG method with the base model. To support the claim of state-of-the-art coherence, the comparison should be controlled: use the same base generator where possible, report the coherence degradation relative to each base model, or at least include coherence scores for the unexpanded base models. Without this, the reported 47.50% TV improvement and the 21.15% LPIPS improvement are not interpretable.
  3. [Sec. 4.1 / Sec. 4.2 / Appendix D] The dataset description is internally inconsistent. The main text says the dataset contains 1,000 prompts and that "using all prompts ... and 25 sets of random seeds" yields 2,000 panoramas; Sec. 4.2 says 80 prompts per theme (2,000 prompts for 25 themes); Appendix D says "400 prompts per theme" (10,000 prompts). These numbers cannot all be correct, and they prevent the experiments from being replicated. Please report the exact number of prompts, seeds, and generated images, and correct the statements.
  4. [Sec. 3.2, Eq. (10)] The method assumes the pre-trained tokenizer decoder fT_d can decode concatenated token sequences longer than the training grid into an arbitrary-width panorama. The paper does not verify or discuss whether LlamaGen's VQGAN decoder preserves quality outside its training resolution, nor whether the concatenation order in Eqs. (7)-(9) matches the decoder's expected spatial arrangement. This assumption is part of the method and should be tested, for example by decoding a horizontally concatenated sequence of two unmodified generated blocks and measuring seam artifacts.
minor comments (6)
  1. [Sec. 3.1, Eqs. (1)-(3)] The loss function and the optimization for theta* are not used by the training-free method; they are only motivational. Please state this explicitly to avoid the appearance that the method optimizes Eq. (2).
  2. [Sec. 4.2] The word "construsting" should be "constructing".
  3. [Eq. (6)] The notation v1,r*sqrt(p) and similar expressions lack clear subscripts; please typeset as v_{1,r*sqrt(p)} and v_{1,p} for readability.
  4. [Appendix D] The description of prompt creation ("400 prompts per theme") should be aligned with the 1,000-prompt total mentioned in the abstract and Sec. 4.1.
  5. [Fig. 4] The axes of the quality-efficiency plot are not labeled; please add labels and units.
  6. [Secs. 1-2] The abbreviations "V AR" (visual AR) and "VAR" (visual autoregressive) are used inconsistently; please unify them.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the training-free redirection method uses an external pre-trained model with no parameter fit to its own metrics; the horizontal off-raster assumption is a validity concern, not a circular reduction.

full rationale

PanoLlama's derivation chain is training-free and relies on the external pre-trained LlamaGen model; no parameter is fitted to the paper's own evaluation data. Eqs. (1)-(3) are a modeling re-description of panorama generation as a product of conditional crop probabilities, and while they do not train a model, they do not smuggle the target result into the input. The redirection operations in Eq. (6) and Eq. (8) are constructive heuristics applied to a fixed external model; the paper does not estimate any parameter from the coherence metrics it later reports, so the SOTA numbers are not forced by construction. The same-group citations ([47], [48]) appear only as related-work baselines or prior connection-strategy references and are not invoked to justify the redirection premise or to exclude alternative designs. The horizontal-expansion step does rest on an unverified assumption that a pre-trained raster-scan model's continuation after a row suffix behaves like a horizontal extension rather than the trained next-row continuation; that is an empirical validity risk, not a circular reduction. There is no self-definitional equation, no fitted input renamed as a prediction, no load-bearing self-citation chain, and no uniqueness theorem imported from the authors. Consequently no circular step is exhibited.

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

The central method is an inference-time rewrite of an off-the-shelf AR generator. It introduces no trainable parameters but relies on several unstated assumptions: that the raster-scan token geometry of the pretrained model maps token suffixes to spatial continuations as assumed, that the VQGAN decoder accepts arbitrarily shaped token grids, and that the self-built benchmark is a fair basis for SOTA claims. The expansion stride and blending weight are hand-chosen hyperparameters. No new physical or learned entities are postulated.

free parameters (3)
  • expansion stride u = 0.75 default; swept over {1, 0.75, 0.5, 0.25, 0.125}
    Controls the overlap between consecutive crops. Hand-chosen, not learned, and it directly trades quality against generation time.
  • rows and columns per redirection (r and c) = derived from u and the token grid; not specified numerically
    Determines how many tokens of the previous crop are reused as context. Chosen by hand to balance context range and speed.
  • blending factor lambda = 1.0 for single-prompt generation; 0.5 to 0.8 for multi-prompt applications
    Used only in the optional blending of Eq. 13 for mask-free layout control and multi-guidance. Not needed for the main panorama results.
assumptions (4)
  • domain assumption Off-the-shelf components (LlamaGen, VQGAN, FLAN-T5) work as described and are not modified.
    The whole method is an inference-time wrapper around these pretrained models; if any component does not support the assumed conditioning or decoding, the pipeline fails. Invoked throughout Sec. 3.2 and Appendix A.
  • domain assumption Raster-scan token suffixes correspond to spatially contiguous image regions.
    Eqs. 6 and 8 redirect generation using trailing tokens of the previous crop; this assumes the model's positional encoding treats those tokens as the bottom rows or right columns of a continuation. The paper does not test this directly.
  • ad hoc to paper The VQGAN decoder can decode concatenated token sequences of arbitrary length and shape into a valid panorama.
    Eq. 10 applies the decoder to V whose length can be n*r*sqrt(p) or c*sqrt(p) per row; pretrained decoders are typically fixed-shape, and this generalization is unproven.
  • domain assumption The self-built benchmark and the nine chosen metrics faithfully capture panorama quality.
    The 1,000-prompt dataset is AI-generated by the authors and not released; coherence metrics may reward repetitive textures rather than true semantic continuity.

how reviews work

0 comments
Cite this review

Pith. "Pith review of PanoLlama: Generating Endless and Coherent Panoramas with Next-Token-Prediction LLMs." pith.science (2026). https://pith.science/paper/X7ILXCBN

@misc{pith2026241115867,
  author       = {Pith},
  title        = {Pith review of: PanoLlama: Generating Endless and Coherent Panoramas with Next-Token-Prediction LLMs},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/X7ILXCBN}},
  note         = {Machine review of arXiv:2411.15867}
}
read the original abstract

Panoramic Image Generation (PIG) aims to create coherent images of arbitrary lengths. Most existing methods fall in the joint diffusion paradigm, but their complex and heuristic crop connection designs often limit their ability to achieve multilevel coherence. By deconstructing this challenge into its core components, we find it naturally aligns with next-token prediction, leading us to adopt an autoregressive (AR) paradigm for PIG modeling. However, existing visual AR (VAR) models are limited to fixed-size generation, lacking the capability to produce panoramic images. In this paper, we propose PanoLlama, a novel framework that achieves endless and coherent panorama generation with the autoregressive paradigm. Our approach develops a training-free strategy that utilizes token redirection to overcome the size limitations of existing VAR models, enabling next-crop prediction in both horizontal and vertical directions. This refreshes the PIG pipeline while achieving SOTA performance in coherence (47.50%), fidelity(28.16%), and aesthetics (15%). Additionally, PanoLlama supports applications other PIG methods cannot achieve, including mask-free layout control, multi-scale and multi-guidance synthesis. To facilitate standardized evaluation, we also establish a dataset with 1,000 prompts spanning 100+ themes, providing a new testing benchmark for PIG research. The code is available at https://github.com/0606zt/PanoLlama.

Figures

Figures reproduced from arXiv: 2411.15867 by the authors.

Figure 1
Figure 1. PanoLlama vs. the most effective current PIG method, joint diffusion. Joint diffusion methods (right) denoise each image crop [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. The framework of PanoLlama consists of three parts: (i) [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Qualitative comparisons between six baselines and ours. Our approach minimizes unnatural connections commonly found in [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (4 more)
Figure 5
Figure 5. Figure 5: Ablation studies on different panorama sizes [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]
Figure 6
Figure 6. Figure 6: Exploration on a wide range of prompt themes. Our [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]
Figure 4
Figure 4. Figure 4: The quantitative and qualitative results demonstrate our [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]
Figure 7
Figure 7. Figure 7: PanoLlama enables versatile applications that other PIG methods cannot achieve, including multi-scale expansion, mask-free [PITH_FULL_IMAGE:figures/full_fig_p008_7.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

48 extracted references · 18 canonical work pages

  1. [1]

    Blended latent diffusion

    Omri Avrahami, Ohad Fried, and Dani Lischinski. Blended latent diffusion. ACM Transactions on Graphics (TOG), 42 (4):1–11, 2023. 2, 6

    work page 2023

  2. [2]

    Sequential modeling enables scalable learn- ing for large vision models

    Yutong Bai, Xinyang Geng, Karttikeya Mangalam, Amir Bar, Alan L Yuille, Trevor Darrell, Jitendra Malik, and Alexei A Efros. Sequential modeling enables scalable learn- ing for large vision models. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 22861–22872, 2024. 1

    work page 2024

  3. [3]

    Multidiffusion: Fusing diffusion paths for controlled image generation

    Omer Bar-Tal, Lior Yariv, Yaron Lipman, and Tali Dekel. Multidiffusion: Fusing diffusion paths for controlled image generation. In International Conference on Machine Learn- ing, pages 1737–1752, 2023. 1, 2, 5, 6

    work page 2023

  4. [4]

    Brown, Benjamin Mann, Nick Ryder, Melanie Sub- biah, Jared Kaplan, Prafulla Dhariwal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, et al

    Tom B. Brown, Benjamin Mann, Nick Ryder, Melanie Sub- biah, Jared Kaplan, Prafulla Dhariwal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, et al. Lan- guage models are few-shot learners. In Advances in Neural Information Processing Systems, pages 1877–1901, 2020. 1, 2

    work page 1901

  5. [5]

    Efficient-vqgan: To- wards high-resolution image generation with efficient vision transformers

    Shiyue Cao, Yueqin Yin, Lianghua Huang, Yu Liu, Xin Zhao, Deli Zhao, and Kaigi Huang. Efficient-vqgan: To- wards high-resolution image generation with efficient vision transformers. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 7368–7377, 2023. 2

    2023

  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

    2022

  7. [7]

    A comprehensive survey for generative data augmentation

    Yunhao Chen, Zihui Yan, and Yunjie Zhu. A comprehensive survey for generative data augmentation. Neurocomputing, page 128167, 2024. 1

    work page 2024

  8. [8]

    Text2light: Zero-shot text-driven hdr panorama generation

    Zhaoxi Chen, Guangcong Wang, and Ziwei Liu. Text2light: Zero-shot text-driven hdr panorama generation. ACM Trans- actions on Graphics (TOG), 41(6):1–16, 2022. 2

    work page 2022

Show all 48 references
  1. [9]

    Scaling instruction- finetuned language models

    Hyung Won Chung, Le Hou, Shayne Longpre, Barret Zoph, Yi Tay, William Fedus, Yunxuan Li, Xuezhi Wang, Mostafa Dehghani, Siddhartha Brahma, et al. Scaling instruction- finetuned language models. Journal of Machine Learning Research, 25(70):1–53, 2024. 1

    2024

  2. [10]

    Image quality assessment: Unifying structure and texture similarity

    Keyan Ding, Kede Ma, Shiqi Wang, and Eero P Simoncelli. Image quality assessment: Unifying structure and texture similarity. IEEE transactions on pattern analysis and ma- chine intelligence, 44(5):2567–2581, 2020. 5

    2020

  3. [11]

    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. 2

    2021

  4. [12]

    Gans trained by a two time-scale update rule converge to a local nash equilib- rium

    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. 5

    2017

  5. [13]

    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. 1

    2020

  6. [14]

    Dynamicid: Zero-shot multi-id image personalization with flexible facial editabil- ity

    Xirui Hu, Jiahao Wang, Hao Chen, Weizhan Zhang, Benqi Wang, Yikun Li, and Haishun Nan. Dynamicid: Zero-shot multi-id image personalization with flexible facial editabil- ity. arXiv preprint arXiv:2503.06505, 2025. 1

    2025 arXiv

  7. [15]

    Diffusion model-based image editing: A survey

    Yi Huang, Jiancheng Huang, Yifan Liu, Mingfu Yan, Jiaxi Lv, Jianzhuang Liu, Wei Xiong, He Zhang, Liangliang Cao, and Shifeng Chen. Diffusion model-based image editing: A survey. IEEE Transactions on Pattern Analysis & Machine Intelligence, pages 1–27, 2025. 1

    2025

  8. [16]

    A survey of multimodal con- trollable diffusion models

    Rui Jiang, Guang-Cong Zheng, Teng Li, Tian-Rui Yang, Jing-Dong Wang, and Xi Li. A survey of multimodal con- trollable diffusion models. Journal of Computer Science and Technology, 39(3):509–541, 2024. 1

    2024

  9. [17]

    Streammultidiffusion: Real-time interac- tive generation with region-based semantic control

    Jaerin Lee, Daniel Sungho Jung, Kanggeon Lee, and Ky- oung Mu Lee. Streammultidiffusion: Real-time interac- tive generation with region-based semantic control. arXiv preprint arXiv:2403.09055, 2024. 2, 6

    2024 arXiv

  10. [18]

    Syncdiffusion: Coherent montage via synchronized joint diffusions

    Yuseung Lee, Kunho Kim, Hyunjin Kim, and Minhyuk Sung. Syncdiffusion: Coherent montage via synchronized joint diffusions. Advances in Neural Information Processing Systems, 36:50648–50660, 2023. 1, 2, 5, 6

    2023

  11. [19]

    Panogen: Text-conditioned panoramic environment generation for vision-and-language navigation

    Jialu Li and Mohit Bansal. Panogen: Text-conditioned panoramic environment generation for vision-and-language navigation. Advances in Neural Information Processing Sys- tems, 36, 2024. 2

    2024

  12. [20]

    Diffusion models for image restoration and enhancement–a compre- hensive survey

    Xin Li, Yulin Ren, Xin Jin, Cuiling Lan, Xingrui Wang, Wenjun Zeng, Xinchao Wang, and Zhibo Chen. Diffusion models for image restoration and enhancement–a compre- hensive survey. arXiv preprint arXiv:2308.09388, 2023. 1

    2023

  13. [21]

    Controlvar: Exploring con- trollable visual autoregressive modeling

    Xiang Li, Kai Qiu, Hao Chen, Jason Kuen, Zhe Lin, Rita Singh, and Bhiksha Raj. Controlvar: Exploring con- trollable visual autoregressive modeling. arXiv preprint arXiv:2406.09750, 2024. 2

    2024 arXiv

  14. [22]

    Wonderfree: Enhancing novel view quality and cross-view consistency for 3d scene exploration

    Chaojun Ni, Jie Li, Haoyun Li, Hengyu Liu, Xiaofeng Wang, Zheng Zhu, Guosheng Zhao, Boyuan Wang, Chenxin Li, Guan Huang, et al. Wonderfree: Enhancing novel view quality and cross-view consistency for 3d scene exploration. arXiv preprint arXiv:2506.20590, 2025. 2

    2025 arXiv

  15. [23]

    SDXL: Improving latent diffusion models for high-resolution image synthesis

    Dustin Podell, Zion English, Kyle Lacey, Andreas Blattmann, Tim Dockhorn, Jonas M ¨uller, Joe Penna, and Robin Rombach. SDXL: Improving latent diffusion models for high-resolution image synthesis. In The Twelfth Interna- tional Conference on Learning Representations, 2024. 6

    2024

  16. [24]

    Merging and splitting diffusion paths for se- mantically coherent panoramas

    Fabio Quattrini, Vittorio Pippi, Silvia Cascianelli, and Rita Cucchiara. Merging and splitting diffusion paths for se- mantically coherent panoramas. In European Conference on Computer Vision, pages 234–251. Springer, 2025. 1, 2, 6

    2025

  17. [25]

    Learning transferable visual models from natural language supervi- sion

    Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, et al. Learning transferable visual models from natural language supervi- sion. In International conference on machine learning, ...

    2021

  18. [26]

    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, 2, 6

    2022

  19. [27]

    Mm-diffusion: Learning multi-modal diffusion mod- els for joint audio and video generation

    Ludan Ruan, Yiyang Ma, Huan Yang, Huiguo He, Bei Liu, Jianlong Fu, Nicholas Jing Yuan, Qin Jin, and Baining Guo. Mm-diffusion: Learning multi-modal diffusion mod- els for joint audio and video generation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern...

    2023

  20. [28]

    Nonlinear total variation based noise removal algorithms

    Leonid I Rudin, Stanley Osher, and Emad Fatemi. Nonlinear total variation based noise removal algorithms. Physica D: nonlinear phenomena, 60(1-4):259–268, 1992. 5

    1992

  21. [29]

    Improved techniques for training gans

    Tim Salimans, Ian Goodfellow, Wojciech Zaremba, Vicki Cheung, Alec Radford, and Xi Chen. Improved techniques for training gans. Advances in neural information processing systems, 29, 2016. 5

    2016

  22. [30]

    Laion-5b: An open large-scale dataset for training next generation image-text models

    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- f...

    2022

  23. [31]

    Deep unsupervised learning using nonequilibrium thermodynamics

    Jascha Sohl-Dickstein, Eric Weiss, Niru Maheswaranathan, and Surya Ganguli. Deep unsupervised learning using nonequilibrium thermodynamics. In International confer- ence on machine learning, pages 2256–2265, 2015. 1

    2015

  24. [32]

    Denois- ing diffusion implicit models

    Jiaming Song, Chenlin Meng, and Stefano Ermon. Denois- ing diffusion implicit models. In International Conference on Learning Representations, 2021. 1

    2021

  25. [33]

    Autoregressive model beats diffusion: Llama for scalable image generation

    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. 1, 2, 6

    2024 arXiv

  26. [34]

    Gemini: a family of highly capable multimodal models

    Gemini Team, Rohan Anil, Sebastian Borgeaud, Jean- Baptiste Alayrac, Jiahui Yu, Radu Soricut, Johan Schalkwyk, Andrew M Dai, Anja Hauth, Katie Millican, et al. Gemini: a family of highly capable multimodal models. arXiv preprint arXiv:2312.11805, 2023. 2

    2023 arXiv

  27. [35]

    Visual autoregressive modeling: Scalable im- age generation via next-scale prediction

    Keyu Tian, Yi Jiang, Zehuan Yuan, Bingyue Peng, and Li- wei Wang. Visual autoregressive modeling: Scalable im- age generation via next-scale prediction. arXiv preprint arXiv:2404.02905, 2024. 2

    2024 arXiv

  28. [36]

    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. 2, 1

    2023 arXiv

  29. [37]

    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. 2, 1

    2017

  30. [38]

    Attention is all you need

    Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszko- reit, Llion Jones, Aidan N Gomez, Łukasz Kaiser, and Illia Polosukhin. Attention is all you need. Advances in neural information processing systems, 30, 2017. 1

    2017

  31. [39]

    360-degree panorama generation from few unregis- tered nfov images

    Jionghao Wang, Ziyu Chen, Jun Ling, Rong Xie, and Li Song. 360-degree panorama generation from few unregis- tered nfov images. In Proceedings of the 31st ACM Interna- tional Conference on Multimedia , pages 6811–6821, 2023. 2

    2023

  32. [40]

    Emu3: Next-token prediction is all you need

    Xinlong Wang, Xiaosong Zhang, Zhengxiong Luo, Quan Sun, Yufeng Cui, Jinsheng Wang, Fan Zhang, Yueze Wang, Zhen Li, Qiying Yu, et al. Emu3: Next-token prediction is all you need. arXiv preprint arXiv:2409.18869, 2024. 2

    2024 arXiv

  33. [41]

    Mul- tiscale structural similarity for image quality assessment

    Zhou Wang, Eero P Simoncelli, and Alan C Bovik. Mul- tiscale structural similarity for image quality assessment. In The Thrity-Seventh Asilomar Conference on Signals, Systems & Computers, 2003, pages 1398–1402. Ieee, 2003. 5

    2003

  34. [42]

    Panod- iffusion: 360-degree panorama outpainting via diffusion

    Tianhao Wu, Chuanxia Zheng, and Tat-Jen Cham. Panod- iffusion: 360-degree panorama outpainting via diffusion. In The Twelfth International Conference on Learning Represen- tations, 2024. 2

    2024

  35. [43]

    Autoregressive models in vision: A sur- vey

    Jing Xiong, Gongye Liu, Lun Huang, Chengyue Wu, Taiqiang Wu, Yao Mu, Yuan Yao, Hui Shen, Zhongwei Wan, Jinfa Huang, et al. Autoregressive models in vision: A sur- vey. arXiv preprint arXiv:2411.05902, 2024. 2

    2024 arXiv

  36. [44]

    Diffusion models: A comprehensive survey of methods and applications

    Ling Yang, Zhilong Zhang, Yang Song, Shenda Hong, Run- sheng Xu, Yue Zhao, Wentao Zhang, Bin Cui, and Ming- Hsuan Yang. Diffusion models: A comprehensive survey of methods and applications. ACM Computing Surveys, 56(4): 1–39, 2023. 1

    2023

  37. [45]

    Language model beats diffusion - tokenizer is key to visual generation

    Lijun Yu, Jose Lezama, Nitesh Bharadwaj Gundavarapu, Luca Versari, Kihyuk Sohn, David Minnen, Yong Cheng, Agrim Gupta, Xiuye Gu, Alexander G Hauptmann, Boqing Gong, Ming-Hsuan Yang, Irfan Essa, David A Ross, and Lu Jiang. Language model beats diffusion - tokenizer is key to vi...

    2024

  38. [46]

    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. In Proceedings of the IEEE conference on computer vision and pattern recogni- tion, pages 586–595, 2018. 5

    2018

  39. [47]

    Multi-scale diffusion: Enhancing spatial layout in high-resolution panoramic image generation

    Xiaoyu Zhang, Teng Zhou, Xinlong Zhang, Jia Wei, and Yongchuan Tang. Multi-scale diffusion: Enhancing spatial layout in high-resolution panoramic image generation. arXiv preprint arXiv:2410.18830, 2024. 2

    2024 arXiv

  40. [48]

    Twindiffusion: Enhanc- ing coherence and efficiency in panoramic image generation with diffusion models

    Teng Zhou and Yongchuan Tang. Twindiffusion: Enhanc- ing coherence and efficiency in panoramic image generation with diffusion models. In ECAI 2024, pages 386–393. IOS Press, 2024. 1, 2, 5, 6 A. PanoLlama Generation Process We provide detailed pseudocode in Alg. 1 to facilitat...

    2024

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.