Hierarchical SVG Tokenization: Learning Compact Visual Programs for Scalable Vector Graphics Modeling

Ximing Xing; Ziteng Xue; Zhenxi Li; Weicong Liang; Linqing Wang; Zhantao Yang; Tiankai Hang; Zijin Yin

cs.LGApr 6, 2026

Hierarchical SVG Tokenization: Learning Compact Visual Programs for Scalable Vector Graphics Modeling

Authors: Ximing Xing, Ziteng Xue, Zhenxi Li, Weicong Liang, Linqing Wang, Zhantao Yang, Tiankai Hang, Zijin Yin,

Organizations: Tencent · Visual Computing Group, Beijing

Abstract

Recent large language models have shifted SVG generation from differentiable rendering optimization to autoregressive program synthesis. However, existing approaches still rely on generic byte-level tokenization inherited from natural language processing, which poorly reflects the geometric structure of vector graphics. Numerical coordinates are fragmented into discrete symbols, destroying spatial relationships and introducing severe token redundancy, often leading to coordinate hallucination and inefficient long-sequence generation. To address these challenges, we propose HiVG, a hierarchical SVG tokenization framework tailored for autoregressive vector graphics generation. HiVG decomposes raw SVG strings into structured atomic tokens and further compresses executable command--parameter groups into geometry-constrained segment tokens, substantially improving sequence efficiency while preserving syntactic validity. To further mitigate spatial mismatch, we introduce a Hierarchical Mean--Noise (HMN) initialization strategy that injects numerical ordering signals and semantic priors into new token embeddings. Combined with a curriculum training paradigm that progressively increases program complexity, HiVG enables more stable learning of executable SVG programs. Extensive experiments on both text-to-SVG and image-to-SVG tasks demonstrate improved generation fidelity, spatial consistency, and sequence efficiency compared with conventional tokenization schemes. Our code is publicly available at https://github.com/ximinng/HiVG

Figures & tables

Figure 1: Sequence-length compression, token-efficient scaling, and human evaluation. (a) HiVG tokenization compresses SVG sequences by 62.7%–63.8% (2.68 $\times$ –2.76 $\times$ ). (b) HiVG reaches comparable quality with approximately 2.7 $\times$ fewer training tokens. (c) HiVG achieves the best human evaluation results, scoring 4.06 in usability and winning 58.9%–70.8% in pairwise comparisons against baselines.

Figure 2: Comparison of SVG tokenization strategies. (a) A generic LLM tokenizer [ 42 , 2 ] treats SVG as plain text and splits it into subword tokens, producing long token sequences. (b) An SVG-aware tokenizer [ 37 , 32 ] improves structural awareness by tokenizing SVG elements and attributes, but geometric primitives remain fragmented into many numeric coordinate tokens. (c) Our HiVG tokenizer introduces a hierarchical representation that groups drawing commands together with their associated coordinates into reusable segment tokens, enabling substantial sequence compression (10 → 7 → 2).

Figure 3: Overview of HiVG. (a) Hierarchical SVG Tokenization ( Sec. 3.1 ). Raw SVG strings are first decomposed into atomic tokens and further compressed into executable segment tokens via structure-aware merging. Only complete command–parameter units are merged to ensure geometric validity. (b) Structure Segment Learning . Segment tokens are learned from a large SVG corpus by discovering renderable command–coordinate groups while discarding merges that violate syntactic or geometric constraints. (c) Model Architecture . Atomic and Segment tokens extend the embedding space of the base LLM, while training progressively increases program depth to stabilize structural abstraction and global composition.

Figure 4: Learned segment tokens as renderable geometric primitives . Segment-level merging preserves syntactic validity and geometric coherence while shortening token sequences and improving efficiency.

Figure 5: HMN initialization for structured SVG tokens . Each new token is initialized by combining a global mean-noise prior with a semantic embedding from its textual description; for coordinate tokens, an additional numeric embedding derived from the normalized value through Gaussian–Polynomial basis encoding is added, while non-numeric tokens omit the numeric branch.

Panel of Figure 6 — Figure 6: Image-to-SVG generation results. For each example, the raster input image is shown on the right and the generated SVG rendering on the left. The examples include icons, logos, typography, UI elements, and emoji-style graphics.

Panel of Figure 8 — Figure 8: Text-to-SVG comparison. Each row corresponds to a text prompt (left). Columns show SVG renderings generated by different methods. Compared with existing models, HiVG produces SVG outputs with more consistent layout structure and better alignment with the prompt description.

Method	Img2SVG						Text2SVG
Method	SSIM $\uparrow$	LPIPS $\downarrow$	CLIP-S $\uparrow$	PickS $\uparrow$	HPS $\uparrow$	Aes $\uparrow$	CLIP $\uparrow$	PickS $\uparrow$	HPS $\uparrow$	Aes $\uparrow$
DeepSeekv3.2 [ 5 ]	-	-	-	-	-	-	0.272	20.331	0.192	4.594
Qwen3.5 Plus [ 21 ]	0.775	0.228	0.896	22.019	0.175	4.672	0.291	20.972	0.206	4.671
Gemini-2.5-pro [ 9 ]	0.790	0.215	0.904	22.346	0.185	4.732	0.284	20.943	0.210	4.765
GPT-5.2 [ 18 ]	0.780	0.205	0.930	23.977	0.222	4.841	0.291	21.268	0.214	4.806
Claude-Sonnet-4.5 [ 1 ]	0.669	0.292	0.842	22.012	0.164	4.435	0.281	20.562	0.195	4.711

Table 1: Quantitative comparison on Img2SVG and Text2SVG tasks. ‘-’ denotes the method can only accept text input.

Variant	Text-to-SVG					Image-to-SVG
Variant	CLIP $\uparrow$	DINO $\uparrow$	HPS $\uparrow$	PickS $\uparrow$	Aes $\uparrow$	SSIM $\uparrow$	LPIPS $\downarrow$	CLIP-S $\uparrow$	HPS $\uparrow$	PickS $\uparrow$	Aes $\uparrow$
AR baseline †	0.2146	0.1520	0.162	19.628	4.548	0.301	0.396	0.797	0.179	19.793	4.553
Ours	0.2392	0.2795	0.194	20.576	4.632	0.896	0.114	0.957	0.221	21.652	4.681
Improvement	+11.5%	+83.9%	+19.8%	+4.8%	+1.8%	+197.7%	-39.3%	+20.1%	+23.5%	+9.4%	+2.8%

Table 2: Impact of structured SVG modeling.

\uparrow

higher is better,

\downarrow

lower is better. † AR baseline trains on raw SVG string. “Aes” denotes the Aesthetic score.

Scale	Tokenization Stats			Text-to-SVG					Image-to-SVG
Scale	Avg Toks	Raw $\to$ AT	AT $\to$ ST	CLIP $\uparrow$	DINO $\uparrow$	HPS $\uparrow$	PickS $\uparrow$	Aes $\uparrow$	SSIM $\uparrow$	LPIPS $\downarrow$	CLIP-S $\uparrow$	HPS $\uparrow$	PickS $\uparrow$	Aes $\uparrow$
$\mathcal{D}_{50\text{k}}$	317	2.59x	1.03x	0.2158	0.4072	0.157	19.398	4.412	0.696	0.313	0.803	0.174	19.710	4.398
$\Delta_{50\text{k}\to 500\text{k}}$	+301	+0.04 $\times$	+0.01 $\times$	+4.6%	-3.3%	+9.6%	+3.0%	+2.4%	+9.1%	-26.5%	+10.7%	+13.8%	+5.5%	+3.5%
$\mathcal{D}_{500\text{k}}$	618	2.63x	1.04x	0.2257	0.3938	0.172	19.981	4.518	0.759	0.230	0.889	0.198	20.791	4.551
$\Delta_{500\text{k}\to 1.5\text{M}}$	-66	0.00 $\times$	+0.01 $\times$	+1.2%	-2.3%	+3.5%	+0.7%	+1.0%	+2.4%	-9.6%	+2.4%	+3.5%	+1.3%	+0.8%
$\mathcal{D}_{1.5\text{M}}$	552	2.63x	1.05x	0.2283	0.3848	0.178	20.113	4.564	0.777	0.208	0.910	0.205	21.056	4.587

Table 3: Effect of Structure Segment Learning (SSL) corpus scale. Dataset statistics and evaluation metrics are reported on both Text-to-SVG and Image-to-SVG tasks. All models use

M{=}500

merges. AT: Atomic Token, ST: Segment Token. Delta rows show changes from previous scale.

#	Method	Components		Img2SVG			Text2SVG
#	Method	Semantic	Numeric	LPIPS $\downarrow$	SSIM $\uparrow$	CLIP-S $\uparrow$	CLIP $\uparrow$	PickScore $\uparrow$	HPS $\uparrow$	Aes $\uparrow$
1	Noise	✗	✗	0.226	0.440	0.795	0.207	19.831	0.144	4.250
2	Mean	✗	✗	0.242	0.244	0.755	0.195	20.110	0.142	4.830
3	Mean+Noise	✗	✗	0.237	0.523	0.821	0.205	19.645	0.137	4.785
4	Semantic	✓	✗	0.236	0.477	0.811	0.205	19.535	0.132	4.675
5	Semantic+Noise	✓	✗	0.233	0.550	0.830	0.206	19.585	0.135	4.715

Table 4: Ablation on token embedding initialization strategies. We report image-to-SVG reconstruction metrics and text-to-SVG generation quality after 1 epoch of training. Bold : best result; underline : second best. All methods use the same

{\sim}3{,}000

SVG tokens and identical training hyperparameters. † Lerp: linear interpolation between text embeddings of “0” and “784”, which does not preserve numeric semantics due to character-level BPE tokenization of digit strings.

	Validity / Efficiency				Visual Similarity			Preference / Aesthetic
Method	Render $\uparrow$	TokCnt	PathCnt	CmdCnt	SSIM $\uparrow$	LPIPS $\downarrow$	CLIP-S $\uparrow$	ImgR $\uparrow$	HPS $\uparrow$	PickS $\uparrow$	Aes $\uparrow$
Stage1-L1	95.20%	273	3.5	54.8	$0.8052\pm 0.13$	$0.1716\pm 0.09$	$0.9360\pm 0.07$	-0.0764	0.2110	21.348	4.6159
Stage1-L2	94.20%	489	5.9	107.2	$0.7223\pm 0.13$	$0.2392\pm 0.09$	$0.8884\pm 0.10$	-0.1836	0.2026	20.945	4.6175
Stage1-L3	90.09%	656	6.7	144.2	$0.7015\pm 0.13$	$0.2405\pm 0.10$	$0.8750\pm 0.11$	-0.2500	0.2050	21.015	4.6285
$\Delta$ S1 $\rightarrow$ S2	+0.2%	+27.0%	+15.3%	+19.5%	+4.1%	-19.7%	+4.9%	+151.7%	+5.9%	+21.0%	+1.3%
Stage2-L1	93.50%	335	4.8	70.4	$0.8152\pm 0.12$	$0.1611\pm 0.08$	$0.9540\pm 0.06$	0.0294	0.2150	21.522	4.6226

Table 5: Effect of three-stage curriculum training on Image-to-SVG with stage-to-stage changes.

\uparrow

higher is better,

\downarrow

lower is better. Delta rows show changes from previous stage (positive values indicate improvement for

\uparrow

metrics).

	Validity / Efficiency				Semantic	Diversity	Preference / Aesthetic
Method	Render $\uparrow$	TokCnt $\downarrow$	PathCnt $\downarrow$	CmdCnt $\downarrow$	CLIP $\uparrow$	DINO $\uparrow$	ImgR $\uparrow$	HPS $\uparrow$	PickS $\uparrow$	Aes $\uparrow$
Stage1-L1	95.60%	288	3.1	60.9	0.2346	0.2949	-0.4789	0.1909	20.552	4.5891
Stage1-L2	95.37%	403	4.3	89.7	0.2322	0.3535	-0.7787	0.1751	20.046	4.5663
Stage1-L3	94.08%	446	5.3	104.9	0.2305	0.4015	-0.8520	0.1730	19.895	4.5580
$\Delta$ S1 $\rightarrow$ S2	-0.8%	+58.1%	+37.2%	+52.4%	+0.6%	-1.3%	+19.3%	+5.1%	+1.2%	+0.6%
Stage2-L1	95.45%	428	4.3	87.6	0.2356	0.2931	-0.3768	0.1953	20.643	4.6279

Table 6: Effect of three-stage curriculum training on Text-to-SVG with stage-to-stage changes.

\uparrow

indicates higher is better,

\downarrow

indicates lower is better. Delta rows show changes from previous stage (positive values indicate improvement for

\uparrow

metrics).

Figure 10: Analysis of structural noise and learned segment properties. (a) Cleaning statistics: Commands removed per sample across data scales; noise is primarily concentrated in l , c , and h types. (b) Redundancy patterns: Frequency mass of strictly degenerate patterns, where <d_0><d_0> pairs represent the most prevalent redundant structure. (c) Command distribution: Relative share of command types within segments across frequency buckets; complex commands like cubic Béziers ( c ) and arcs ( a ) are well-captured. (d) Token length: Atomic token length distribution shows a stable median of $\sim$ 9 tokens, indicating consistent segment complexity regardless of frequency.

Appendix figures & tables8 assets

Supplementary material from the paper’s appendix.

Appendix

Method	Semantic Layering $\uparrow$	Editability $\uparrow$	Redundancy Control $\uparrow$	Overall Code Usability $\uparrow$
SVGen-7B [ 31 ]	2.88	2.83	2.74	2.82
InternSVG-8B [ 32 ]	3.22	3.18	3.09	3.16
Gemini-2.5-pro [ 4 ]	3.39	3.34	3.23	3.32
GPT-5.2 [ 18 ]	3.56	3.49	3.37	3.47
HiVG-3B	4.11	4.05	3.96	4.06

Appendix

Table S1 : Expert review of SVG code usability in Adobe Illustrator. Eight professional SVG practitioners import the generated SVGs into Adobe Illustrator and score their structural usability on a 1–5 Likert scale. Higher is better for all metrics.

Figure S1 : Text-to-SVG generation results. For each example, we show the text prompt together with the rendered SVG output generated by HiVG.

Figure S2 : Image-to-SVG generation results. For each example, the raster input image is shown on the right and the generated SVG rendering on the left.

Figure S3 : Additional Image-to-SVG comparison.

Figure S4 : Additional Text-to-SVG comparison.

Hyperparameter	HiVG
Architecture / Tokenization
Backbone model	Qwen2.5-VL-3B-Instruct
Canvas size	$784\times 784$
Atomic token range	$30$ – $1000$
Number of curriculum stages	3
Context length scaling	progressive across stages

Appendix

Table S2 : Hyperparameter settings of HiVG.

Panel of Figure 21 — Figure S5 : Training templates for text- and image-conditioned SVG token generation. For text-to-SVG Token generation (T2ST), the model takes a textual description as input and autoregressively predicts hierarchical SVG tokens. For image-to-SVG Token generation (I2ST), an image placeholder ( <image> ) is inserted into the user turn, and the model outputs the corresponding SVG token sequence under the same conversational format. Using unified templates across both tasks simplifies multi-task training and keeps the supervision interface consistent.

Figure S6 : Structural noise patterns in raw SVGs. Redundant command groups and zero-move operations contribute no meaningful geometry to the rendered image, yet they artificially inflate token length and degrade computational efficiency.

Vector Graphics Generation

See all 26 papers

Evaluating Constrained Iterative Refinement for Scalable Vector Graphics Generation with Off-the-Shelf VLMs
Aug 26, 2026Matthew Perlman, James Beetham, Niels Da Vitoria Lobo +2Vector Graphics Generation VLM Adaptation
University of Massachusetts Amherst · University of Central Florida
A Scalable Vector Graphics Latent Space
Aug 22, 2026Leonardo Zini, Elia Frigieri, Lorenzo BaraldiRepresentation Learning Visual Representation Learning
University of Modena and Reggio Emilia, Italy
CANVAS: Consistency-Aware Navigation via Visual Adaptive Sampling for Long-Context Text-to-SVG Generation
Aug 31, 2026Yichen Wu, Haoxuan Qu, Yihang Lou +2T2I Generation Text-to-SVG Generation
University of Nottingham · Lancaster University · Peking University

Aug 26, 2026J/K move · Enter open · S save

Evaluating Constrained Iterative Refinement for Scalable Vector Graphics Generation with Off-the-Shelf VLMs

Matthew Perlman, James Beetham, Niels Da Vitoria Lobo, Amrit Singh Bedi+1

University of Massachusetts Amherst · University of Central Florida

Vector Graphics Generation VLM Adaptation Iterative Refinement VLM Evaluation Constrained Generative Modeling

Open paper AskPDF arXiv

Scalable Vector Graphics (SVGs) power much of the modern visual ecosystem, yet state-of-the-art generative models focus almost entirely on rasterized images. We explore whether inference-time methods can unlock SVG generation capabilities in off-the-shelf vision-language models (VLMs). We systematically evaluate a constrained iterative refinement harness that combines visual feedback, structured editing, and constrained decoding to characterize the capabilities and limitations of current VLMs for SVG generation. Across multiple VLMs and generation settings, we find that constrained decoding improves compilation success rates, while iterative refinement reveals a deficit in visual reasoning and self-correction. Our results highlight both the promise and current limitations of using inference-time methods to adapt general-purpose VLMs for SVG generation.

Abstract

Vector Graphics Generation

Explore similar work

Evaluating Constrained Iterative Refinement for Scalable Vector Graphics Generation with Off-the-Shelf VLMs

A Scalable Vector Graphics Latent Space

CANVAS: Consistency-Aware Navigation via Visual Adaptive Sampling for Long-Context Text-to-SVG Generation