SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2104.12229 · ICCV · 2021

Vector Neurons: A General Framework for SO(3)-Equivariant Networks

Leonidas Guibas, Andrea Tagliasacchi, Or Litany, Adrien Poulenard, Yueqi Duan, Congyue Deng

arXiv · PDF · Open in the Atlas

Code that ran

8 of 17 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (1 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 1 violated it and 4 have no contract verdict; 3 failures of our instrument; 9 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/2104.12229.json · badge: /api/ran/2104.12229.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
FlyingGiraffe/vnn-neural-implicits community (classifier label)
classifier basis not recorded on the link
6 ran · 0 no recorded run
sidhikabalachandar/oavnn community
Papers with Code archive
1 ran · 9 no recorded run
FlyingGiraffe/vnn marked official in Papers with Code
Papers with Code archive
1 ran · 0 no recorded run
FunctionStatusWhere it lives
VNLinear Ran
ran on a synthesized input · constructed an object; no result was computed
FlyingGiraffe/vnn/models/vn_layers.py
code served (permissive licence) · get_code("154a3564fcf7fc91")
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 an instance of its own class, VNLinear: the run constructed the object and did not call it, so no result was computed. 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

torch.manual_seed(0)

# VNLinear takes in_channels and out_channels as parameters
in_channels = 4
out_channels = 3

ARGS = {
    "in_channels": in_channels,
    "out_channels": out_channels,
}
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 VNLinear — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a VNLinear output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not 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: none by the harness; any seed is in the setup block
VNReflectionLinearLeakyReLU Ran
ran on a synthesized input
sidhikabalachandar/oavnn/model.py
pointer only (licence: NONE) · get_code("1080ada53cb2d4c1")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, sidhikabalachandar/oavnn when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

get_graph_feature Ran
ran on a synthesized input
FlyingGiraffe/vnn-neural-implicits/im2mesh/layers_equi.py
code served (permissive licence) · get_code("e09e2de9ccd4a1ee")
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 5, dtype family float, shape (2, 8, 3, 5, 3) as a later run recorded it. The image it ran in was not recorded. The code mentions cuda, and the harness aliases cuda to the CPU for every sample; whether this run needed that is 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)

# The function get_graph_feature expects:
# - x: a tensor that will be viewed as (batch_size, -1, num_points)
#   After view, it should have shape (batch_size, num_dims, num_points)
#   where num_dims must be divisible by 3
# - k: an integer for number of neighbors
# - idx: optional pre-computed indices (None here to test knn path)
# - x_coord: optional coordinate tensor for knn (we'll provide x instead)

batch_size = 2
num_points = 5
num_dims = 12  # Must be divisible by 3 for the num_dims // 3 operation

# Create x with shape that when viewed becomes (batch_size, num_dims, num_points)
# Input shape should be (batch_size, num_dims, 1, num_points) based on the view operation
x = torch.randn(batch_size, num_dims, 1, num_points)

k = 3

# idx is None, so knn will be called
idx = None

# x_coord is None, so knn will use x instead
x_coord = None

ARGS = {
    "x": x,
    "k": k,
    "idx": idx,
    "x_coord": x_coord
}
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 5, dtype family float, shape (2, 8, 3, 5, 3) — shape (2, 8, 3, 5, 3) 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
get_graph_feature_cross Ran
ran; our contract draft was wrong, not the code
FlyingGiraffe/vnn-neural-implicits/im2mesh/layers_equi.py
pointer only (licence: MIT); the stored copy get_code serves, from crtie/leveraging-se-3-equivariance-for-learning-3d-geometric-shape-assembly, is pointer only (licence: NONE) · get_code("d5dbd8d8b3a240d1")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from crtie/leveraging-se-3-equivariance-for-learning-3d-geometric-shape-assembly, and GitHub detected no licence in that repository when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

is_url Ran
ran, violated the contract we drafted
FlyingGiraffe/vnn-neural-implicits/im2mesh/checkpoints.py
code served (permissive licence) · get_code("e39d881796e8d67e")
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 bool, ndim 0, dtype family float, shape () 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 urllib

def is_url(url):
    scheme = urllib.parse.urlparse(url).scheme
    return scheme in ('http', 'https')

ARGS = {
    'url': 'https://example.com'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type bool, ndim 0, dtype family float, shape () — shape () 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
knn Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
FlyingGiraffe/vnn-neural-implicits/im2mesh/layers_equi.py
code served (permissive licence) · get_code("cdd0141594039dcb")
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 3, dtype family int, shape (2, 3, 2) 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)

# x is used with:
# - transpose(2, 1): expects 3D tensor
# - matmul with transposed version: matrix multiplication
# - sum(x ** 2, dim=1, keepdim=True): sum over dimension 1
# - topk(k=k, dim=-1): get top k indices along last dimension
# 
# Shape should be (batch, features, num_points) based on the transpose operations
# Let's use batch=2, features=4, num_points=3

x = torch.randn(2, 4, 3)

# k is the number of nearest neighbors to find
# Should be less than num_points (3), so k=2
k = 2

ARGS = {
    "x": x,
    "k": k,
}
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 3, dtype family int, shape (2, 3, 2) — shape (2, 3, 2) 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
maxpool Ran
ran; our contract draft was wrong, not the code
FlyingGiraffe/vnn-neural-implicits/im2mesh/encoder/vnn_tnet.py
pointer only (licence: MIT); the stored copy get_code serves, from leolyliu/core4d-instructions, is pointer only (licence: NONE) · get_code("94c6fe25807ef01e")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from leolyliu/core4d-instructions, and GitHub detected no licence in that repository when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

meanpool Ran
ran on a synthesized input
FlyingGiraffe/vnn-neural-implicits/im2mesh/encoder/vnn_tnet.py
code served (permissive licence) · get_code("86ab3883f4591f31")
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

torch.manual_seed(0)

# Create a small tensor to pass to meanpool
x = torch.randn(2, 4, 3)

# Define arguments dict with all parameters
ARGS = {
    'x': x,
    'dim': -1,
    'keepdim': False
}
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
VNEmbedding Not yet run
unverified (harvested, no recorded run)
sidhikabalachandar/oavnn/model.py
pointer only (licence: NONE) · get_code("44997ad3144d060a")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, sidhikabalachandar/oavnn when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. No invocation recorded.

VNLinear Not yet run
unverified (harvested, no recorded run)
sidhikabalachandar/oavnn/model.py
pointer only (licence: NONE) · get_code("913a5ed3a4619465")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, sidhikabalachandar/oavnn when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. No invocation recorded.

VNLinearAndLeakyReLU Not yet run
unverified (harvested, no recorded run)
sidhikabalachandar/oavnn/model.py
pointer only (licence: NONE) · get_code("f8f8559195990e90")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, sidhikabalachandar/oavnn when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. No invocation recorded.

VNLinearLeakyReLU Not yet run
unverified (harvested, no recorded run)
sidhikabalachandar/oavnn/model.py
pointer only (licence: NONE) · get_code("d62d926017129a0a")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, sidhikabalachandar/oavnn when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. No invocation recorded.

VNN Not yet run
unverified (harvested, no recorded run)
sidhikabalachandar/oavnn/model.py
pointer only (licence: NONE) · get_code("9e4f7f50d9b26e83")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, sidhikabalachandar/oavnn when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. No invocation recorded.

VNReflectionLinear Not yet run
unverified (harvested, no recorded run)
sidhikabalachandar/oavnn/model.py
pointer only (licence: NONE) · get_code("5926f396f1921e83")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, sidhikabalachandar/oavnn when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. No invocation recorded.

VNReflectionLinearAndLeakyReLU Not yet run
unverified (harvested, no recorded run)
sidhikabalachandar/oavnn/model.py
pointer only (licence: NONE) · get_code("417f212371ebed0d")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, sidhikabalachandar/oavnn when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. No invocation recorded.

VNReflectionLinearLeakyReLUBOnly Not yet run
unverified (harvested, no recorded run)
sidhikabalachandar/oavnn/model.py
pointer only (licence: NONE) · get_code("5d28be1fb471af27")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, sidhikabalachandar/oavnn when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. No invocation recorded.

VNStdFeature Not yet run
unverified (harvested, no recorded run)
sidhikabalachandar/oavnn/model.py
pointer only (licence: NONE) · get_code("a1460ecb6d2c8cd1")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, sidhikabalachandar/oavnn when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. No invocation recorded.

Also present in other papers' code

Of 17 samples lifted from this paper's linked repositories, 17 carry a function identity and 0 do not. Of the 17 distinct functions with an identity, 5 are also present in other papers whose code we hold (2 with at least one paper dated earlier, 1 the same month, 5 later) and 12 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.

FunctionAlso present inEarliest-datedOthers named
knn
brick of this paper's assembly: helper
52 other papers
9 dated earlier, 43 later · 12 only as an assembly brick
1801.07829 (2018-01)
Dynamic Graph CNN for Learning on Point Clouds
1905.06292 (2019-05)
3D Point Cloud Generative Adversarial Network Based on Tree Structured Graph Convolutions
2005.06734 (2020-05)
Dense-Resolution Network for Point Cloud Classification and Segmentation
and 49 more not named here
is_url 12 other papers
3 dated earlier, 1 the same month, 8 later
1912.07372 (2019-12)
Differentiable Volumetric Rendering: Learning Implicit 3D Representations without 3D Supervision
2001.05119 (2020-01)
Learning multiview 3D point cloud registration
2003.04618 (2020-03)
Convolutional Occupancy Networks
and 9 more not named here
maxpool 4 other papers
4 later
2203.16529 (2022-03)
CaDeX: Learning Canonical Deformation Coordinate Space for Dynamic Surface Representation via Neural Homeomorphism
2303.14505 (2023-03)
Unsupervised Inference of Signed Distance Functions from Single Sparse Point Clouds without Learning Priors
2310.03740 (2023-10)
ContactGen: Generative Contact Modeling for Grasp Generation
and 1 more not named here
get_graph_feature_cross 2 other papers
2 later
2211.09786 (2022-11)
SE(3)-Equivariant Relational Rearrangement with Neural Descriptor Fields
2309.06810 (2023-09)
Leveraging SE(3) Equivariance for Learning 3D Geometric Shape Assembly
get_graph_feature 1 other paper
1 later
2211.09786 (2022-11)
SE(3)-Equivariant Relational Rearrangement with Neural Descriptor Fields

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

Invariance and equivariance to the rotation group have been widely discussed in the 3D deep learning community for pointclouds. Yet most proposed methods either use complex mathematical tools that may limit their accessibility, or are tied to specific input data types and network architectures. In this paper, we introduce a general framework built on top of what we call Vector Neuron representations for creating SO(3)-equivariant neural networks for pointcloud processing. Extending neurons from 1D scalars to 3D vectors, our vector neurons enable a simple mapping of SO(3) actions to latent spaces thereby providing a framework for building equivariance in common neural operations -including linear layers, non-linearities, pooling, and normalizations. Due to their simplicity, vector neurons are versatile and, as we demonstrate, can be incorporated into diverse network architecture backbones, allowing them to process geometry inputs in arbitrary poses. Despite its simplicity, our method performs comparably well in accuracy and generalization with other more complex and specialized state-of-the-art methods on classification and segmentation tasks. We also show for the first time a rotation equivariant reconstruction network. Source code is available at https://github.com/FlyingGiraffe/vnn.

For agents

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

get_harvested_code_for_paper("2104.12229")
get_code_for_paper("2104.12229")
have("2104.12229")

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

curl https://syntology.ai/api/ran/2104.12229.json

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

[![Syntology run record](https://syntology.ai/api/ran/2104.12229.svg)](https://syntology.ai/paper/2104.12229)

Connect an agent — have() is free.