SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2007.11301 · NeurIPS · 2020

DeepSVG: A Hierarchical Generative Network for Vector Graphics Animation

Alexandre Carlier, Martin Danelljan, Alexandre Alahi, Radu Timofte, Ecole Fédérale De Lausanne

arXiv · PDF · Open in the Atlas

Code that ran

5 of 25 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 1 violated it and 3 have no contract verdict; 1 failure of our instrument; 20 no recorded run), measured 2026-11-22; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/2007.11301.json · badge: /api/ran/2007.11301.svg

Each repository's role is a label: a classifier's output recorded on the link, or, where no role is recorded, what the Papers with Code archive marked. The line under it names its basis. Neither says who wrote the code.

RepositoryRoleRan
alexandre01/deepsvg canonical (classifier label)
classifier basis: classifier_2025
5 ran · 20 no recorded run
FunctionStatusWhere it lives
FCN Ran
ran on a synthesized input
alexandre01/deepsvg/deepsvg/model/model.py
code served (permissive licence) · get_code("0f7f57b9f6743705")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, class construct-then-call prompt, model not recorded) and returned tuple. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn

INIT_ARGS = {
    'd_model': 4,
    'n_commands': 3,
    'n_args': 2,
    'args_dim': 5,
}

ARGS = {
    'out': torch.randn(6, 2, 4),
}

class FCN(nn.Module):

    def __init__(self, d_model, n_commands, n_args, args_dim=256):
        super().__init__()
        self.n_args = n_args
        self.args_dim = args_dim
        self.command_fcn = nn.Linear(d_model, n_commands)
        self.args_fcn = nn.Linear(d_model, n_args * args_dim)

    def forward(self, out):
        S, N, _ = out.shape
        command_logits = self.command_fcn(out)
        args_logits = self.args_fcn(out)
        args_logits = args_logits.reshape(S, N, self.n_args, self.args_dim)
        return (command_logits, args_logits)


if __name__ == '__main__':
    model = FCN(**INIT_ARGS)
    result = model(**ARGS)
    print("Command logits shape:", result[0].shape)
    print("Args logits shape:", result[1].shape)
Setup block drafted by
a language model, class construct-then-call prompt, model not recorded
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple — a tuple output has no array shape of its own; none is stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
HierarchFCN Ran
ran on a synthesized input
alexandre01/deepsvg/deepsvg/model/model.py
code served (permissive licence) · get_code("c4a9704d5f61ac91")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, class construct-then-call prompt, model not recorded) and returned tuple. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn

class HierarchFCN(nn.Module):

    def __init__(self, d_model, dim_z):
        super().__init__()
        self.visibility_fcn = nn.Linear(d_model, 2)
        self.z_fcn = nn.Linear(d_model, dim_z)

    def forward(self, out):
        G, N, _ = out.shape
        visibility_logits = self.visibility_fcn(out)
        z = self.z_fcn(out)
        return (visibility_logits.unsqueeze(0), z.unsqueeze(0))


INIT_ARGS = {
    'd_model': 4,
    'dim_z': 3,
}

ARGS = {
    'out': torch.randn(2, 5, 4),
}

if __name__ == '__main__':
    model = HierarchFCN(**INIT_ARGS)
    result = model(**ARGS)
    print("Output shapes:")
    print(f"  visibility_logits: {result[0].shape}")
    print(f"  z: {result[1].shape}")
Setup block drafted by
a language model, class construct-then-call prompt, model not recorded
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple — a tuple output has no array shape of its own; none is stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
ResNet Ran
ran on a synthesized input
alexandre01/deepsvg/deepsvg/model/model.py
code served (permissive licence) · get_code("32df1a7d7da8dbc7")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, class construct-then-call prompt, model not recorded) and returned Tensor, ndim 2, dtype family float, shape (2, 4) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn

class ResNet(nn.Module):

    def __init__(self, d_model):
        super().__init__()
        self.linear1 = nn.Sequential(nn.Linear(d_model, d_model), nn.ReLU())
        self.linear2 = nn.Sequential(nn.Linear(d_model, d_model), nn.ReLU())
        self.linear3 = nn.Sequential(nn.Linear(d_model, d_model), nn.ReLU())
        self.linear4 = nn.Sequential(nn.Linear(d_model, d_model), nn.ReLU())

    def forward(self, z):
        z = z + self.linear1(z)
        z = z + self.linear2(z)
        z = z + self.linear3(z)
        z = z + self.linear4(z)
        return z


INIT_ARGS = {
    'd_model': 4
}

ARGS = {
    'z': torch.randn(2, 4)
}

model = ResNet(**INIT_ARGS)
output = model(**ARGS)
Setup block drafted by
a language model, class construct-then-call prompt, model not recorded
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 2, dtype family float, shape (2, 4) — shape (2, 4) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
_generate_square_subsequent_mask Ran
ran, violated the contract we drafted
alexandre01/deepsvg/deepsvg/model/model.py
code served (permissive licence) · get_code("1263393b5a254282")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1) and returned Tensor, ndim 2, dtype family float, shape (4, 4) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

ARGS = {
    "sz": 4
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 2, dtype family float, shape (4, 4) — shape (4, 4) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
_threshold_sample Ran
ran; our contract draft was wrong, not the code
alexandre01/deepsvg/deepsvg/model/model.py
code served (permissive licence) · get_code("ea8d8a3182dcdbbc")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 2, dtype family float, shape (2, 4) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn.functional as F

torch.manual_seed(0)

# arg_logits: tensor passed to F.softmax, so needs shape (..., 2) for dim=-1 indexing
# The function indexes [..., 1] and compares with threshold (scalar)
# Temperature and threshold are scalars with defaults
arg_logits = torch.randn(2, 4, 2)

threshold = 0.5

temperature = 1.0

ARGS = {
    "arg_logits": arg_logits,
    "threshold": threshold,
    "temperature": temperature,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 2, dtype family float, shape (2, 4) — shape (2, 4) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
Bottleneck Not yet run
unverified (harvested, no recorded run)
alexandre01/deepsvg/deepsvg/model/model.py
code served (permissive licence) · get_code("317c8344bea83044")
How it ran

no invocation recorded.

ConstEmbedding Not yet run
unverified (harvested, no recorded run)
alexandre01/deepsvg/deepsvg/model/model.py
code served (permissive licence) · get_code("7dcced8878b64383")
How it ran

no invocation recorded.

Decoder Not yet run
unverified (harvested, no recorded run)
alexandre01/deepsvg/deepsvg/model/model.py
code served (permissive licence) · get_code("a7a568c0aec488a9")
How it ran

no invocation recorded.

Encoder Not yet run
unverified (harvested, no recorded run)
alexandre01/deepsvg/deepsvg/model/model.py
code served (permissive licence) · get_code("27acde109433ec82")
How it ran

no invocation recorded.

LabelEmbedding Not yet run
unverified (harvested, no recorded run)
alexandre01/deepsvg/deepsvg/model/model.py
code served (permissive licence) · get_code("5f0886ba37847885")
How it ran

no invocation recorded.

SVGEmbedding Not yet run
unverified (harvested, no recorded run)
alexandre01/deepsvg/deepsvg/model/model.py
code served (permissive licence) · get_code("79a174d33890ab0e")
How it ran

no invocation recorded.

SVGTensor Not yet run
unverified (harvested, no recorded run)
alexandre01/deepsvg/deepsvg/model/model.py
code served (permissive licence) · get_code("7a0ffcbae95f2fd2")
How it ran

no invocation recorded.

SVGTransformer Not yet run
unverified (harvested, no recorded run)
alexandre01/deepsvg/deepsvg/model/model.py
code served (permissive licence) · get_code("c08f5fffe0b2ad35")
How it ran

no invocation recorded.

VAE Not yet run
unverified (harvested, no recorded run)
alexandre01/deepsvg/deepsvg/model/model.py
code served (permissive licence) · get_code("43e61efc110a1a58")
How it ran

no invocation recorded.

_DefaultConfig Not yet run
unverified (harvested, no recorded run)
alexandre01/deepsvg/deepsvg/model/model.py
code served (permissive licence) · get_code("d88109fe4a168610")
How it ran

no invocation recorded.

_get_group_mask Not yet run
unverified (harvested, no recorded run)
alexandre01/deepsvg/deepsvg/model/model.py
code served (permissive licence) · get_code("2cd604e7a8260f32")
How it ran

no invocation recorded.

_get_key_padding_mask Not yet run
unverified (harvested, no recorded run)
alexandre01/deepsvg/deepsvg/model/model.py
code served (permissive licence) · get_code("3ac1a54394606a15")
How it ran

no invocation recorded.

_get_key_visibility_mask Not yet run
unverified (harvested, no recorded run)
alexandre01/deepsvg/deepsvg/model/model.py
code served (permissive licence) · get_code("ef8136f817499cbc")
How it ran

no invocation recorded.

_get_padding_mask Not yet run
unverified (harvested, no recorded run)
alexandre01/deepsvg/deepsvg/model/model.py
code served (permissive licence) · get_code("16d404266079eb1f")
How it ran

no invocation recorded.

_get_visibility_mask Not yet run
unverified (harvested, no recorded run)
alexandre01/deepsvg/deepsvg/model/model.py
code served (permissive licence) · get_code("a089d43e1ca8c113")
How it ran

no invocation recorded.

_make_batch_first Not yet run
unverified (harvested, no recorded run)
alexandre01/deepsvg/deepsvg/model/model.py
code served (permissive licence) · get_code("c552e5432337eadf")
How it ran

no invocation recorded.

_make_seq_first Not yet run
unverified (harvested, no recorded run)
alexandre01/deepsvg/deepsvg/model/model.py
code served (permissive licence) · get_code("6706fc8261009bed")
How it ran

no invocation recorded.

_pack_group_batch Not yet run
unverified (harvested, no recorded run)
alexandre01/deepsvg/deepsvg/model/model.py
code served (permissive licence) · get_code("c87a2ecdbe4da91b")
How it ran

no invocation recorded.

_sample_categorical Not yet run
unverified (harvested, no recorded run)
alexandre01/deepsvg/deepsvg/model/model.py
code served (permissive licence) · get_code("58f52852746594ee")
How it ran

no invocation recorded.

_unpack_group_batch Not yet run
unverified (harvested, no recorded run)
alexandre01/deepsvg/deepsvg/model/model.py
code served (permissive licence) · get_code("7af00434bdb31122")
How it ran

no invocation recorded.

Also present in other papers' code

Of 25 samples lifted from this paper's linked repositories, 25 carry a function identity and 0 do not. Of the 25 distinct functions with an identity, 0 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 0 later) and 25 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-22.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Scalable Vector Graphics (SVG) are ubiquitous in modern 2D interfaces due to their ability to scale to different resolutions. However, despite the success of deep learning-based models applied to rasterized images, the problem of vector graphics representation learning and generation remains largely unexplored. In this work, we propose a novel hierarchical generative network, called DeepSVG, for complex SVG icons generation and interpolation. Our architecture effectively disentangles high-level shapes from the low-level commands that encode the shape itself. The network directly predicts a set of shapes in a non-autoregressive fashion. We introduce the task of complex SVG icons generation by releasing a new large-scale dataset along with an open-source library for SVG manipulation. We demonstrate that our network learns to accurately reconstruct diverse vector graphics, and can serve as a powerful animation tool by performing interpolations and other latent space operations. Our code is available at https://github.com/alexandre01/ deepsvg. Figure 1: DeepSVG generates vector graphics by predicting draw commands, such as lines and Bézier curves. Our latent space allows meaningful animations between complex vector graphics icons.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2007.11301")
get_code_for_paper("2007.11301")
have("2007.11301")

The run record, dated, one paper per request, free:

curl https://syntology.ai/api/ran/2007.11301.json

A badge for a README (the split and the date, never a ratio):

[![Syntology run record](https://syntology.ai/api/ran/2007.11301.svg)](https://syntology.ai/paper/2007.11301)

Connect an agent — have() is free.