Infinite-Precision Autoregressive Modeling for Vector Graphics and Layouts

Bohyung Han; Bongkeun Kim; Insoo Kim; Keonwoo Bae; Yeonsang Shin

REVIEW 4 major objections 5 minor 7 references

Infinite-Precision Autoregressive Modeling for Vector Graphics and Layouts

T0 review · 4 major / 5 minor · reviewed 2026-08-03 · deepseek-v4-flash

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

Pith's one-line read AGDC jointly predicts discrete identifiers and diffuses continuous coordinates inside one autoregressive transformer, giving variable-length hybrid vectors effectively unbounded precision.

desk verdict AGDC has a genuinely new hybrid autoregressive-diffusion recipe and a potentially useful benchmark, but the main tables hide the appendix's own LT-8 results that beat it on FID. read the letter →

arxiv 2601.05680 v2 pith:D3G5OJKC submitted 2026-01-09 cs.LG cs.AIcs.CV

Yeonsang Shin , Insoo Kim , Bongkeun Kim , Keonwoo Bae , Bohyung Han This is my paper

classification cs.LGcs.AIcs.CV

keywords autoregressivegenerationcontinuousdiffusionhybriddiscrete-continuoussequenceslayoutvectorgraphicsprecisionscalabilitysemiconductorlayoutsdesignrulechecks

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

AGDC argues that token-based discretization caps the precision of autoregressive generation, and that the fix is to stop quantizing continuous coordinates. It builds an autoregressive transformer whose discrete branch classifies element types while a small diffusion network, conditioned on the same latent vector, generates each continuous coordinate vector in its natural real-valued form. Two additions make sequence length controllable: an MLP that adjusts the end-of-sequence logit from context, and a differentiable length-regularization loss. The paper claims this beats discretization-based and fixed-schema baselines on semiconductor layouts (its new ContLayNet benchmark), graphic layouts, and text-to-SVG, and that precision errors, not just visual similarity, should be measured. If right, autoregressive models can scale to high-precision engineering domains without vocabulary explosion.

What carries the argument

The load-bearing object is the atomic unit: a discrete identifier plus a real-valued vector. The architecture is a transformer decoder whose latent z_i feeds two parallel heads — an MLP+softmax for the discrete identifier and an AdaLN-conditioned denoising MLP that predicts the noise to remove from c_i (a MAR-style diffusion head). Sequence length is controlled by adding α·MLPEOS(z_i) to the EOS token logit and by Lℓ=(E[length]−L_target)^2, where E[length] is computed from adjusted EOS probabilities. This is what lets one model do categorical prediction and diffusion in the same autoregressive loop without fixing a schema or a token budget.

What would settle it

Re-run Table 3 and Table 2 with LayoutTransformer at every precision from 4 to 18 bits (and DLT with tuned length control), then compare against AGDC on the same metrics; if an 8-bit or 12-bit LT equals or beats AGDC on FID, overlap, and alignment, the claim that discretization fundamentally caps achievable fidelity is falsified for graphic layouts. On ContLayNet the decisive test is whether any discretized precision passes the DRC metrics at the same rate as AGDC's continuous outputs.

Watch

Extended reading notes

Core claim

The central claim is that a hybrid sequence of discrete identifiers and continuous vectors can be modeled autoregressively without lossy transformation: discrete values remain categorical predictions, and continuous values are sampled by a reverse diffusion process conditioned on the autoregressive latent z_i. The model represents each atomic unit u_i=[d_i,c_i] as e_i=concat(E(onehot(d_i)), W c_i), feeds previous embeddings to a transformer to obtain z_i, then predicts d_{i+1} with softmax and c_{i+1} with a denoising MLP. To terminate at the right length, the EOS logit is adjusted by MLPEOS(z_i), and training includes a squared penalty on expected length. The paper introduces ContLayNet, 33

Load-bearing premise

The main comparison claims rest on evaluating LayoutTransformer at 18-bit precision; the paper's own Table B shows the 8-bit configuration achieves lower FID on PubLayNet and Rico completion (3.00 vs 4.58 and 5.20 vs 9.77), so if the right baseline is each method's best-performing precision, the central outperformance claim weakens substantially.

Editorial extensions

If this is right

High-precision domains like semiconductor layout no longer need exponentially growing vocabularies; the continuous branch's diffusion output is effectively unbounded in resolution (15-bit or beyond) on a fixed model.
Variable-length generation becomes differentiable and controllable: the EOS adjustment and length loss reduce length error from a biased mean (~39) to near zero (~0.15) in the paper's ContLayNet measurement.
Autoregressive methods can be extended to any atomic-unit-style hybrid domain (typed boxes or path commands), not just layouts and SVG, by reusing the same embedding-branch structure.
Functional evaluation becomes feasible: DRC-based metrics catch failures that FID misses, so generative models for chips can be scored by whether the circuit would work.
In text-to-SVG, the continuous model keeps FID/CLIP quality comparable to the best 8-bit baseline while remaining functional at 10+ bits, where the discretized baseline fails entirely.

Reading between the lines

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

Editorial: the main tables evaluate LayoutTransformer at 18 bits, but the supplement's Table B shows 8-bit LT achieves lower FID than AGDC on PubLayNet completion (3.00 vs 4.58) and Rico completion (5.20 vs 9.77); a head-to-head against each baseline's best-performing precision would be the more conservative test of the 'discretization fundamentally limits fidelity' claim.
Editorial: because the continuous branch samples with 10–100 diffusion steps per coordinate vector, inference is much slower per step than a discrete token; the reported 50-step Improved DDPM configuration trades quality for speed, and faster samplers or distillation would likely be needed for interactive design tools.
Editorial: the atomic-unit abstraction suggests immediate testable extensions — PCB routing, floor plans, and CAD sketches — and a natural stress test is adding curved or nested paths, where a fixed 8-dimensional coordinate vector may be insufficient.
Editorial: one could test the length-control mechanism in isolation across datasets by measuring calibration of P(EOS) under distribution shift; the paper only reports Gaussian-fitted length-error statistics on ContLayNet.

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.

Desk Editor's Note

read the letter

You should know before reading: the architecture is genuinely interesting, but the central empirical claim does not survive contact with the paper's own table in the appendix. The main text says AGDC outperforms discretization baselines across all tasks, but Table B shows LayoutTransformer at 8 bits getting FID 3.00 on PubLayNet completion versus AGDC's 4.58, and 5.20 versus 9.77 on Rico. The main tables only run LT at 18 bits, where it does poorly. That is a selective comparison, and the abstract's unqualified "higher-fidelity than discretization-based baselines" is not supported.

What is actually new: combining a softmax discrete head with MAR-style diffusion in a variable-length autoregressive loop, plus the EOS logit adjustment and length regularization. The ablations show both components help, and the precision-scaling analysis in Section 3.1 is correct, if elementary. ContLayNet with DRC metrics is a reasonable attempt to measure functional correctness in a precision-sensitive domain, and AGDC's DRC numbers are much better than the baselines. That part holds up.

Soft spots beyond the cherry-pick: no error bars anywhere, so we cannot tell if AGDC's margins over DLT are meaningful. The dataset and code are not released despite being central to the evaluation. The SVG experiment is also weak — AGDC's FID is 48.73 versus IconShop's 37.35 at 8 bits, so "comparable" is doing a lot of work. The ContLayNet DRC metrics are only evaluated on the authors' own data, with no external check on whether those rules capture realistic design constraints.

I don't think the central method is broken. The idea is coherent, and the ContLayNet results are promising. But the paper as written overclaims, and the fix is straightforward: report LT at its best precision in the main tables, or explicitly restrict the claim to high-precision settings while acknowledging the trade-off. If they do that, the contribution stands on the architecture and the benchmark.

This is worth a serious referee. A good reviewer could push for honest comparisons, error bars, and dataset release. I would bring it to a reading group and I'd cite it for the hybrid AR-diffusion formulation and ContLayNet, but I would not take the empirical superiority claim at face value.

Referee Report

4 major / 5 minor

Summary. The paper proposes AGDC, an autoregressive framework that jointly models discrete and continuous values in variable-length sequences by combining categorical prediction for discrete identifiers with diffusion-based modeling for continuous vectors. Two technical components are introduced: an MLP-based EOS logit adjustment and a length regularization loss. The authors also introduce ContLayNet, a 334K-sample semiconductor layout benchmark with Design Rule Check (DRC) metrics. Experiments cover ContLayNet, PubLayNet/Rico layouts, and FIGR-8-SVG, comparing against LayoutTransformer, DLT, and IconShop. The central claim is that AGDC achieves higher-fidelity hybrid vector representations than discretization-based and fixed-schema baselines, especially at high precision.

Significance. If the claims are substantiated, the work addresses a real limitation of tokenized autoregressive models in continuous-valued domains and proposes a novel architecture that integrates diffusion into autoregressive generation for variable-length hybrid sequences. The ContLayNet benchmark and DRC metrics are potentially valuable resources for the semiconductor layout community. However, the central outperformance claim is not supported by the paper's own data: the appendix reports that an 8-bit discretized baseline (LT) achieves better FID on both layout datasets, and the SVG experiment shows IconShop at 8 bits has lower FID than AGDC. The contribution is therefore more modest than advertised, resting primarily on ContLayNet under DRC metrics. The method is coherent and the ablations suggest the two proposed components help, but the evidence does not justify the unqualified 'higher-fidelity' claim.

major comments (4)

[Section 6.2, Table 3 vs Appendix B.1, Table B] The claim in Section 6.2 that 'AGDC outperforms both LT (18 bits) and DLT across all tasks' omits the 8-bit LT results reported in the paper's own Table B. On PubLayNet completion, LT at 8 bits has FID 3.00 and Overlap 2.78 versus AGDC's 4.58 and 4.59; on Rico completion, LT at 8 bits has FID 5.20 and Alignment 0.21 versus AGDC's 9.77 and 0.25. Since Table B is part of the same manuscript, this is an internal inconsistency. The main comparison should include the best-performing precision of the baseline, or the claim should be restricted to the high-precision setting.
[Section 6.3, Table 4] Table 4 shows IconShop at 8 bits achieves FID 37.35 while AGDC achieves 48.73. The text states AGDC 'maintain quality comparable to IconShop's best outputs,' but the abstract and introduction claim 'higher-fidelity hybrid vector representations than discretization-based ... baselines.' This is not supported. The claim should be revised to state that AGDC is superior only at high precision and comparable or worse at the precision where the discretized baseline performs best.
[Section 6.1, Table 2 and Section 6.2, Table 3] No error bars, confidence intervals, or significance tests are reported for any of the headline numbers. Since several comparisons involve small margins (e.g., PubLayNet ablation rows) and the central claim depends on specific baseline configurations, the absence of variance estimates makes it difficult to assess whether AGDC's improvements are meaningful. I request at least three seeds with standard deviations, or a paired test for the key comparisons.
[Section 6.1, 'Baselines' paragraph] The authors justify evaluating LT at 18-bit precision as 'the maximum precision achievable with a batch size of 2 on a single NVIDIA A6000 GPU.' Precision is a property of the data representation, not the hardware; presumably the constraint is the vocabulary size and memory footprint. This needs clarification, and more importantly, it should be acknowledged that Table B shows 8-bit is the best-performing LT configuration for the layout tasks. The choice of 18-bit as the main baseline appears to select a configuration that is known to perform poorly, rather than the one that best represents the discretization approach.

minor comments (5)

[Section 5.1] The paragraph contains a duplicated statement: 'with substantial structural complexity, each sample averaging 323 layers' appears effectively twice. Please remove the repetition.
[Section 6.1, Implementation details] The sentence 'MLPCONT consists of three blocks with 1024 channels each' is unclear: is this the continuous branch of the transformer, the diffusion conditioning network, or the denoising network? Please define all components consistently.
[Equation (12)] The expected-length formula is written without derivation, and the adjusted EOS probability is not explicitly defined. Please specify that p_i^EOS is computed from softmax over the adjusted logits in Equation (9).
[Figure 7] The caption lists 'LT (8 bits) LT (18 bits) DLT ( bits) AGDC ( bits)' with blank precision values for DLT and AGDC. In several figures (e.g., Figure 5, C, D) the precision annotations are also blank. Fill these in or remove the placeholder text.
[Reproducibility] The ContLayNet dataset is stated to be released 'upon acceptance,' but no code repository or data availability link is provided. For a benchmark paper, releasing the data and evaluation code is essential for the contribution to be verifiable.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: AGDC's derivation and evaluation are self-contained, with external baselines and no load-bearing self-citation.

full rationale

I walked the derivation chain from Section 3 (precision analysis and autoregressive formulation) through Section 4 (AGDC methodology) to the experiments in Section 6 and the appendices. The continuous branch is a standard diffusion objective conditioned on the autoregressive latent (Eq. 11), the discrete branch is a categorical cross-entropy (Eq. 10), and the length mechanism is a differentiable regularizer computed from the model's own adjusted EOS probabilities (Eqs. 12-13). None of these reduces to a fitted parameter renamed as a prediction: the expected-length formula is a definitional identity of the model's EOS distribution used as a training loss against ground-truth lengths, not as an evaluation prediction. The central empirical claims are assessed against external baselines (LayoutTransformer, DLT, IconShop) on existing datasets (PubLayNet, Rico, FIGR-8-SVG) and on the newly introduced ContLayNet; creating a new benchmark and metrics is not circular, even though the authors also report state-of-the-art results on it. There is no self-citation chain, no imported uniqueness theorem, and no ansatz smuggled in via citation; the cited MAR and improved-DDPM components are standard prior machinery used as building blocks, not as evidence for the paper's own conclusions. The paper's main-table comparison uses LT at 18 bits while appendix Table B reports stronger LT performance at 8 bits; that is a potential correctness or reporting issue about which configuration best supports the headline claim, but it is not an instance of the derivation being equivalent to its inputs by construction. Therefore no circular step meets the evidentiary bar, and the appropriate score is 0.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

The model's central claim relies on standard diffusion and AR mathematics, plus domain assumptions about the atomic-unit representation and the ContLayNet DRC rules. The hand-set loss weights and DRC thresholds are free parameters. No new physical entities are posited.

free parameters (4)

λ1 (continuous loss weight) = 100 (all domains)
Hand-set loss weight balancing diffusion term; reported in Section 6.1/A.1/B.2/C.1.
λ2 (length regularization weight) = 0.1 ContLayNet; 0.005 PubLayNet/Rico; 0.001 FIGR-8-SVG
Hand-set; ablation shows contribution.
α (EOS logit scaling) = 0.1 ContLayNet; 0.05 layouts; 0.005 SVG
Hand-set scaling for MLP EOS adjustment.
DRC thresholds (ε, W, H) = 240, 1200, 1000
Hand-set; define functional correctness in ContLayNet metrics; not derived from actual design rules.

assumptions (5)

domain assumption Continuous coordinates are effectively infinite-precision because Δx→0 for continuous support
Section 3.1 equates mathematical continuity with computational precision; ignores floating-point finite precision.
domain assumption Atomic-unit representation covers the domains studied
Section 4.1 states 'majority of such hybrid vector representations' can be expressed as atomic units; no proof of coverage.
domain assumption The four DRC constraints (CLC, PDC, HSC, VSC) capture circuit functional correctness
Section 5.2 defines these rules as fundamental; no validation against real design-rule checking or electrical simulation.
standard math EOS expected length formula Eq. (12) is exact for autoregressive stopping
Product-of-survival formula; exact if p_i are conditional stop probabilities.
standard math Score-matching objective Eq. (11) trains a valid continuous conditional density
Uses standard diffusion/denoising assumption.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Infinite-Precision Autoregressive Modeling for Vector Graphics and Layouts." pith.science (2026). https://pith.science/paper/D3G5OJKC

@misc{pith2026260105680,
  author       = {Pith},
  title        = {Pith review of: Infinite-Precision Autoregressive Modeling for Vector Graphics and Layouts},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/D3G5OJKC}},
  note         = {Machine review of arXiv:2601.05680}
}

read the original abstract

While Transformer-based autoregressive models excel in data generation, their token discretization strategy inherently limits their precision in continuous domains. We analyze the scalability limitations of existing discretization-based approaches for generating hybrid discrete-continuous sequences, particularly in high-precision domains such as logos, layouts, and semiconductor circuit designs, where precision loss potentially leads to visual artifacts, aesthetic degradation, and even functional failure. To address the challenge, we propose a novel unified framework that jointly models discrete and continuous values for variable-length sequences. Our approach employs a hybrid approach that combines categorical prediction for discrete values with diffusion-based modeling for continuous values, incorporating two key technical components: an end-of-sequence (EOS) logit adjustment mechanism that uses an MLP to dynamically adjust EOS token logits based on sequence context, and a length regularization term integrated into the loss function. Additionally, we present ContLayNet, a large-scale benchmark comprising 334K high-precision semiconductor layout samples with specialized evaluation metrics that capture functional correctness, where precision errors significantly impact performance. Experiments on multiple domains show that our approach achieves higher-fidelity hybrid vector representations than discretization-based and fixed-schema baselines, while effectively scaling to high-precision generation.

Figures

Figures reproduced from arXiv: 2601.05680 by the authors.

**Figure 1.** Impact of discretization. Discretization fundamentally compromises the scalability of precision. For precision analysis, refer to Section 3.1. Autoregressive models (Vaswani et al., 2017; Brown et al., 2020; Ramesh et al., 2021; Radford et al., 2023) have shown remarkable success in generating sequences across various domains, typically relying on discrete values such as text tokens and quantized image representa… view at source ↗

**Figure 2.** Atomic unit representation. Many real-world domains, including layouts and SVGs, naturally contain both discrete (e.g., types, classes) and continuous values (e.g., coordinates, amounts). The majority of such hybrid vector representations can be expressed as sequences of atomic units, with each unit combining discrete and continuous components ( [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗

**Figure 3.** Overview of AGDC. This latent representation z i conditions two parallel branches for discrete and continuous generations. In the discrete branch, z i is converted into categorical probabilities using a multi-layer perceptron (MLP) followed by a softmax function: p(d i |z i ) = softmax(MLPDISC(z i )) (9) In the continuous branch, a denoising network εθ, conditioned on z i , iteratively removes noise from corrupte… view at source ↗

Figures from the paper (4 more)

**Figure 4.** Figure 4: The ContLayNet dataset. Left: Example visualization of a sample. Right: Comparison showing functional failure when ContLayNet samples get discretized to low precision. Generation process. For discrete identifiers, samples are drawn from p(d|z) after the EOS adjustment…

**Figure 5.** Figure 5: Qualitative results on ContLayNet in completion task given 50 / 100 layers. AGDC illustrate clearly superior performance to LT and DLT. Results [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]

**Figure 6.** Figure 6: Length error comparisons. Ablation studies [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]

**Figure 7.** Figure 7: Top: Layout Generation on PubLayNet. Bottom: Text-to-SVG Generation on FIGR8-SVG. AGDC achieves superior performance in high-precision settings while maintaining quality comparable to low-precision settings across diverse domains. 7 CONCLUSION This paper introduces AG…

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

7 extracted references · 2 linked inside Pith

[6]

We excluded paired data with text longer than 50 words or SVG longer than 256 atomic units

to get the valid text-SVG pairs. We excluded paired data with text longer than 50 words or SVG longer than 256 atomic units. For the training validation and testing of the models, 72K samples each are extracted from the data set as validation set and test set. Metrics.We evaluate SVG generation quality using two commonly adopted metrics:Fréchet inception ...

2017
[7]

During training, each latent vector z is processed through the denoising MLP, sampling the timestep t 4 times each

For the denoising process, we employ an MLP comprising three blocks, each with 256 channels, following the diffusion approach of (Nichol & Dhariwal, 2021). During training, each latent vector z is processed through the denoising MLP, sampling the timestep t 4 times each. The discrete branch for SVG generation consists of a two-layer MLP with a hidden laye...

2021
[1024]

During training, each latent vector z is processed through the denoising MLP, sampling the timestep t 30 times each

For the denoising process, we employ an MLP comprising three blocks, each with 1024 channels, following the diffusion approach of (Nichol & Dhariwal, 2021). During training, each latent vector z is processed through the denoising MLP, sampling the timestep t 30 times each. MLPCONT consists of three blocks with 1024 channels each, while MLPDISC and MLPEOS ...

2021
[2017]

We focused on the top 13 component types, following the approach of (Lee et al., 2020)

comprises 91K mobile UI layouts with 27 component types. We focused on the top 13 component types, following the approach of (Lee et al., 2020). For both datasets, we limited our analysis to layouts containing 9 or fewer components, consistent with methods (Kikuchi et al., 2021; Levi et al., 2023). We used the train-test split defined by (Kikuchi et al., ...

2020
[2023]

Dp-tbart: A transformer- based autoregressive model for differentially private tabular data generation.arXiv preprint arXiv:2307.10430,

Rodrigo Castellon, Achintya Gopal, Brian Bloniarz, and David Rosenberg. Dp-tbart: A transformer- based autoregressive model for differentially private tabular data generation.arXiv preprint arXiv:2307.10430,

arXiv
[2024]

Figr: Few-shot image generation with reptile

Louis Clouâtre and Marc Demers. Figr: Few-shot image generation with reptile. InarXiv preprint arXiv:1901.02199,

arXiv 1901
[4090]

B.3 ADDITIONALRESULTS PubLayNet.Figure C and Figure D present additional qualitative results for the completion task and unconditioned generation task on PubLayNet, respectively

We setλ 1 = 100,λ 2 = 0.005, andα= 0.05. B.3 ADDITIONALRESULTS PubLayNet.Figure C and Figure D present additional qualitative results for the completion task and unconditioned generation task on PubLayNet, respectively. AGDC achieves comparable structural 15 Preprint coherence to low-precision LT while maintaining higher precision capabilities. Each recta...

2023

Pith tools

Reviewed August 3, 2026 · model on record in the stance chip above.

[1] [6]

We excluded paired data with text longer than 50 words or SVG longer than 256 atomic units

to get the valid text-SVG pairs. We excluded paired data with text longer than 50 words or SVG longer than 256 atomic units. For the training validation and testing of the models, 72K samples each are extracted from the data set as validation set and test set. Metrics.We evaluate SVG generation quality using two commonly adopted metrics:Fréchet inception ...

2017

[2] [7]

During training, each latent vector z is processed through the denoising MLP, sampling the timestep t 4 times each

For the denoising process, we employ an MLP comprising three blocks, each with 256 channels, following the diffusion approach of (Nichol & Dhariwal, 2021). During training, each latent vector z is processed through the denoising MLP, sampling the timestep t 4 times each. The discrete branch for SVG generation consists of a two-layer MLP with a hidden laye...

2021

[3] [1024]

During training, each latent vector z is processed through the denoising MLP, sampling the timestep t 30 times each

For the denoising process, we employ an MLP comprising three blocks, each with 1024 channels, following the diffusion approach of (Nichol & Dhariwal, 2021). During training, each latent vector z is processed through the denoising MLP, sampling the timestep t 30 times each. MLPCONT consists of three blocks with 1024 channels each, while MLPDISC and MLPEOS ...

2021

[4] [2017]

We focused on the top 13 component types, following the approach of (Lee et al., 2020)

comprises 91K mobile UI layouts with 27 component types. We focused on the top 13 component types, following the approach of (Lee et al., 2020). For both datasets, we limited our analysis to layouts containing 9 or fewer components, consistent with methods (Kikuchi et al., 2021; Levi et al., 2023). We used the train-test split defined by (Kikuchi et al., ...

2020

[5] [2023]

Dp-tbart: A transformer- based autoregressive model for differentially private tabular data generation.arXiv preprint arXiv:2307.10430,

Rodrigo Castellon, Achintya Gopal, Brian Bloniarz, and David Rosenberg. Dp-tbart: A transformer- based autoregressive model for differentially private tabular data generation.arXiv preprint arXiv:2307.10430,

arXiv

[6] [2024]

Figr: Few-shot image generation with reptile

Louis Clouâtre and Marc Demers. Figr: Few-shot image generation with reptile. InarXiv preprint arXiv:1901.02199,

arXiv 1901

[7] [4090]

B.3 ADDITIONALRESULTS PubLayNet.Figure C and Figure D present additional qualitative results for the completion task and unconditioned generation task on PubLayNet, respectively

We setλ 1 = 100,λ 2 = 0.005, andα= 0.05. B.3 ADDITIONALRESULTS PubLayNet.Figure C and Figure D present additional qualitative results for the completion task and unconditioned generation task on PubLayNet, respectively. AGDC achieves comparable structural 15 Preprint coherence to low-precision LT while maintaining higher precision capabilities. Each recta...

2023