SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2109.01066 · ICCV · 2021

4D-Net for Learned Multi-Modal Alignment AJ Piergiovanni Google Research

Anelia Angelova, Vincent Casser, AJ Piergiovanni, Michael S. Ryoo

arXiv · PDF · Open in the Atlas

Code that ran

11 of 11 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, 0 violated it and 10 have no contract verdict; 1 failure of our instrument; 0 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/2109.01066.json · badge: /api/ran/2109.01066.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
chanlilong/4D_NET_pytorch community (classifier label)
classifier basis not recorded on the link
11 ran · 0 no recorded run
FunctionStatusWhere it lives
bound_points_jit Ran
ran on a synthesized input
chanlilong/4D_NET_pytorch/utils/point_cloud_ops.py
code served (permissive licence) · get_code("af7c92882e7d6fbc")
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 ndarray, ndim 1, dtype family int, shape (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 numpy as np

# Create small test data
np.random.seed(0)

# points: shape (n, ndim) - array of points to check
# Should have 2 points with 4 dimensions each
points = np.array([
    [0.5, 1.5, 2.5, 3.5],
    [5.0, 6.0, 7.0, 8.0]
], dtype=np.float64)

# upper_bound: shape (ndim,) - upper bounds for each dimension
upper_bound = np.array([1.0, 2.0, 3.0, 4.0], dtype=np.float64)

# lower_bound: shape (ndim,) - lower bounds for each dimension
lower_bound = np.array([0.0, 1.0, 2.0, 3.0], dtype=np.float64)

ARGS = {
    "points": points,
    "upper_bound": upper_bound,
    "lower_bound": lower_bound
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numba, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 1, dtype family int, shape (2,) — shape (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
box_cxcywh_to_xyxy Ran
ran on a synthesized input
chanlilong/4D_NET_pytorch/utils/box_ops.py
code served (permissive licence) · get_code("a8982a32c4283e7b")
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)

# The function box_cxcywh_to_xyxy expects:
# - x: Tensor with shape (..., 4) where the last dimension contains [x_c, y_c, w, h]
# It unbinds the last dimension into 4 components and converts center-based coordinates
# to corner-based coordinates (x1, y1, x2, y2)

# Create a small tensor with batch size 2 and 4 channels (cxcywh format)
x = torch.randn(2, 4)

ARGS = {
    "x": x
}
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
box_xyxy_to_cxcywh Ran
ran; our contract draft was wrong, not the code
chanlilong/4D_NET_pytorch/utils/box_ops.py
pointer only (licence: MIT); the stored copy get_code serves, from shihuahuang95/deim, is pointer only (licence: NOASSERTION) · get_code("db44d81a87192d0f")
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 shihuahuang95/deim, and GitHub could not classify the licence of that repository (NOASSERTION). 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.

choose_requirement Ran
ran on a synthesized input
chanlilong/4D_NET_pytorch/docker/setup_files/setup_mmcv.py
code served (permissive licence) · get_code("ea0600d3c80887da")
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 str. 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 re
from pkg_resources import DistributionNotFound, get_distribution

# The function choose_requirement takes two string parameters representing
# package requirements (e.g., "numpy>=1.0", "scipy<2.0")
# It tries to parse the package name from the primary requirement,
# checks if it's installed, and returns primary if found, else secondary

ARGS = {
    "primary": "numpy>=1.0.0",
    "secondary": "numpy>=0.9.0"
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
pkg_resources — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str 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
draw_projected_box3d Ran
ran on a synthesized input
chanlilong/4D_NET_pytorch/utils/draw_utils.py
code served (permissive licence) · get_code("27f3875b965b8a33")
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 ndarray, ndim 3, dtype family int, shape (480, 640, 3) 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 cv2
import numpy as np

# Create a small image
image = np.ones((480, 640, 3), dtype=np.uint8) * 255

# Create 8 vertices for a 3D bounding box (8x3 array)
# These represent the 8 corners of a 3D box projected onto 2D image plane
qs = np.array([
    [100, 50, 0],      # vertex 0
    [50, 50, 0],       # vertex 1
    [50, 100, 0],      # vertex 2
    [100, 100, 0],     # vertex 3
    [120, 40, 0],      # vertex 4
    [70, 40, 0],       # vertex 5
    [70, 90, 0],       # vertex 6
    [120, 90, 0],      # vertex 7
], dtype=np.float32)

# Color as RGB tuple
color = (0, 255, 0)

# Line thickness
thickness = 2

ARGS = {
    "image": image,
    "qs": qs,
    "color": color,
    "thickness": thickness,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
cv2, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 3, dtype family int, shape (480, 640, 3) — shape (480, 640, 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_same_padding Ran
ran on a synthesized input
chanlilong/4D_NET_pytorch/models/detector_models.py
code served (permissive licence) · get_code("466f16b4752e5041")
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 int, ndim 0, dtype family int, 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 math

def get_same_padding(x: int, k: int, s: int, d: int) -> int:
    return max((math.ceil(x / s) - 1) * s + (k - 1) * d + 1 - x, 0)

ARGS = {
    "x": 32,
    "k": 3,
    "s": 1,
    "d": 1,
}
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 int, ndim 0, dtype family int, 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
inverse_rigid_trans Ran
ran on a synthesized input
chanlilong/4D_NET_pytorch/utils/calibration_util.py
code served (permissive licence) · get_code("603a527d325787b7")
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 ndarray, ndim 2, dtype family float, shape (3, 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 numpy as np

# Create a valid rigid body transformation matrix (3x4 format [R|t])
# R is a 3x3 rotation matrix (orthogonal), t is a 3x1 translation vector

# Create a simple rotation matrix (identity for simplicity, but valid)
R = np.array([
    [1.0, 0.0, 0.0],
    [0.0, 1.0, 0.0],
    [0.0, 0.0, 1.0]
], dtype=np.float64)

# Create a translation vector
t = np.array([1.0, 2.0, 3.0], dtype=np.float64)

# Construct the 3x4 transformation matrix [R|t]
Tr = np.hstack([R, t.reshape(3, 1)])

ARGS = {
    "Tr": Tr
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 2, dtype family float, shape (3, 4) — shape (3, 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
parse_requirements Ran
ran on a synthesized input
chanlilong/4D_NET_pytorch/docker/setup_files/setup_mmcv.py
code served (permissive licence) · get_code("d44560204b456049")
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 list, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim 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 pathlib
import tempfile
import os

# Create a temporary requirements file for testing
temp_dir = tempfile.mkdtemp()
requirements_file = os.path.join(temp_dir, 'requirements.txt')

# Write some sample requirements to the file
with open(requirements_file, 'w') as f:
    f.write("numpy>=1.19.0\n")
    f.write("scipy==1.5.4\n")
    f.write("torch>1.7.0\n")
    f.write("# This is a comment\n")
    f.write("pandas>=1.0.0;python_version>='3.6'\n")

ARGS = {
    'fname': requirements_file,
    'with_version': True,
}
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 list, ndim 1, dtype family float — a list output has no array shape of its own; none is stated
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
points_to_voxel Ran
ran on a synthesized input
chanlilong/4D_NET_pytorch/utils/point_cloud_ops.py
code served (permissive licence) · get_code("0ce0640739715030")
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 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 numpy as np

# Create small sample point cloud data
# points: [N, ndim] where ndim >= 3 (xyz + possibly reflectivity)
points = np.array([
    [0.5, 0.5, 0.5, 0.9],
    [1.5, 1.5, 1.5, 0.8],
    [2.5, 2.5, 2.5, 0.7],
    [0.2, 0.3, 0.4, 0.6],
], dtype=np.float32)

# voxel_size: [3] for xyz dimensions
voxel_size = np.array([1.0, 1.0, 1.0], dtype=np.float32)

# coors_range: [6] format xyzxyz (minmax)
# xmin, ymin, zmin, xmax, ymax, zmax
coors_range = np.array([0.0, 0.0, 0.0, 4.0, 4.0, 4.0], dtype=np.float32)

# max_points: maximum points per voxel
max_points = 35

# max_voxels: maximum number of voxels to create
max_voxels = 20000

# reverse_index: whether to reverse coordinates (zyx instead of xyz)
reverse_index = True

ARGS = {
    "points": points,
    "voxel_size": voxel_size,
    "coors_range": coors_range,
    "max_points": max_points,
    "max_voxels": max_voxels,
    "reverse_index": reverse_index,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numba, numpy — 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: none by the harness; any seed is in the setup block
rotx Ran
ran on a synthesized input
chanlilong/4D_NET_pytorch/utils/calibration_util.py
code served (permissive licence) · get_code("af3a0c0fae8b7c89")
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 ndarray, ndim 2, dtype family float, shape (3, 3) 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 numpy as np

# Create a small numpy array representing an angle in radians
t = np.array(0.5)

ARGS = {
    "t": t
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 2, dtype family float, shape (3, 3) — shape (3, 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
roty Ran
ran on a synthesized input
chanlilong/4D_NET_pytorch/utils/calibration_util.py
code served (permissive licence) · get_code("e5ce5f132be82d00")
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 ndarray, ndim 2, dtype family float, shape (3, 3) 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 numpy as np

# The function roty takes a rotation angle t (in radians) as a numpy array
# and returns a 3x3 rotation matrix about the y-axis.
# The parameter t should be a scalar or array of angles.

t_angle = np.array(0.5)  # A single rotation angle in radians

ARGS = {
    "t": t_angle
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 2, dtype family float, shape (3, 3) — shape (3, 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

Also present in other papers' code

Of 11 samples lifted from this paper's linked repositories, 11 carry a function identity and 0 do not. Of the 11 distinct functions with an identity, 1 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 1 later) and 10 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
box_xyxy_to_cxcywh 1 other paper
1 later
2412.04234 (2024-12)
DEIM: DETR with Improved Matching for Fast Convergence

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

We present 4D-Net, a 3D object detection approach, which utilizes 3D Point Cloud and RGB sensing information, both in time. We are able to incorporate the 4D information by performing a novel dynamic connection learning across various feature representations and levels of abstraction, as well as by observing geometric constraints. Our approach outperforms the state-of-the-art and strong baselines on the Waymo Open Dataset. 4D-Net is better able to use motion cues and dense image information to detect distant objects more successfully. We will open source the code.

For agents

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

get_harvested_code_for_paper("2109.01066")
get_code_for_paper("2109.01066")
have("2109.01066")

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

curl https://syntology.ai/api/ran/2109.01066.json

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

[![Syntology run record](https://syntology.ai/api/ran/2109.01066.svg)](https://syntology.ai/paper/2109.01066)

Connect an agent — have() is free.