SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2203.11018 · AAAI · 2022

Stereo Neural Vernier Caliper

Zhiqiang Shen, Zechun Liu, Shichao Li, Kwang-Ting Cheng

arXiv · PDF · Open in the Atlas

Code that ran

21 of 28 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (6 of them constructed an object rather than computing a result; 1 honoured the contract we drafted, 0 violated it and 7 have no contract verdict; 13 failures of our instrument; 7 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/2203.11018.json · badge: /api/ran/2203.11018.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
Nicholasli1995/SNVC community
Papers with Code archive
21 ran · 7 no recorded run
FunctionStatusWhere it lives
FCModel Ran
ran on a synthesized input · constructed an object; no result was computed
Nicholasli1995/SNVC/snvc/models/vernier.py
code served (permissive licence) · get_code("381a29fe871103d3")
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, FCModel: 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)

ARGS = {
    "num_neurons": 64,
    "num_blocks": 2,
    "p_dropout": 0.5,
    "kaiming": False,
    "leaky": False,
    "input_size": 32,
    "output_size": 16,
}
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 FCModel — 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 FCModel 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
ResidualBlock Ran
ran on a synthesized input · constructed an object; no result was computed
Nicholasli1995/SNVC/snvc/models/vernier.py
code served (permissive licence) · get_code("95d31d1254c09183")
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, ResidualBlock: 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)

ARGS = {
    "num_neurons": 4,
    "p_dropout": 0.5,
    "kaiming": False,
    "leaky": False,
}

# Test that the module can be instantiated and used
if __name__ == "__main__":
    model = nn.ResidualBlock(**ARGS)
    x = torch.randn(2, 4)
    output = model(x)
    print("Output shape:", output.shape)
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 ResidualBlock — 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 ResidualBlock 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
affine_transform Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
Nicholasli1995/SNVC/snvc/models/vernier.py
code served (permissive licence) · get_code("dc85eba4d70fcb17")
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 (2, 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 2D keypoints: shape (num_points, 2)
kpts_2d = np.array([
    [10.5, 20.3],
    [30.1, 40.7]
], dtype=np.float32)

# Create affine transformation matrix: shape (2, 3)
# This will transform homogeneous coordinates (x, y, 1) to (x', y')
trans = np.array([
    [1.0, 0.0, 5.0],   # translation + scale in x
    [0.0, 1.0, 10.0]   # translation + scale in y
], dtype=np.float32)

# dtype parameter
dtype = np.float32

ARGS = {
    "kpts_2d": kpts_2d,
    "trans": trans,
    "dtype": dtype
}
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 (2, 2) — shape (2, 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
basicdownsample Ran
ran; our contract draft was wrong, not the code
Nicholasli1995/SNVC/snvc/models/vernier.py
code served (permissive licence) · get_code("01a97ab430ea6b53")
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 Sequential. 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)

# The function basicdownsample takes two integer parameters: in_planes and out_planes
# These represent the number of input and output channels for the convolutional layer
# We'll use small distinct values to make the shapes unambiguous

in_planes = 3
out_planes = 16

ARGS = {
    "in_planes": in_planes,
    "out_planes": out_planes,
}
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 Sequential — a Sequential output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Sequential output the recorded ndim 1 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
check_points Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
Nicholasli1995/SNVC/snvc/models/vernier.py
code served (permissive licence) · get_code("c0b5564e1ba92ea8")
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 numpy as np

torch.manual_seed(0)

# The function checks_points takes:
# 1. points: a tensor that can be 1D or 2D
# 2. dimension: an integer representing the expected dimension

# Test case 1: 1D tensor that reshapes to (-1, dimension)
points_1d = torch.randn(8)  # 8 elements, will reshape to (2, 4)
dimension_1 = 4

# Test case 2: 2D tensor with correct dimension
points_2d = torch.randn(2, 4)  # Already 2D with correct shape
dimension_2 = 4

# Use the 1D case for the ARGS
ARGS = {
    "points": points_1d,
    "dimension": dimension_1
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, 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
compute_rigid_transform Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
Nicholasli1995/SNVC/snvc/models/vernier.py
code served (permissive licence) · get_code("5a226aee4dbb27ca")
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

# Set random seed for reproducibility
np.random.seed(0)

# Create small sample data
# X: [d, N] - d dimensions, N data points
d = 4  # dimension
N = 3  # number of data points

# Generate random point clouds
X = np.random.randn(d, N).astype(np.float32)
Y = np.random.randn(d, N).astype(np.float32)

# Create optional weight matrix
# W can be 1D array [N,] or 2D matrix [N, N]
W = np.ones(N).astype(np.float32)

ARGS = {
    "X": X,
    "Y": Y,
    "W": W,
    "verbose": False
}
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 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
construct_box_3d Ran
ran, honoured the contract we drafted
Nicholasli1995/SNVC/snvc/models/vernier.py
code served (permissive licence) · get_code("470f6d8f034c01ec")
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, 9) 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

ARGS = {
    "l": 2.0,
    "h": 1.5,
    "w": 3.0,
}
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, 9) — shape (3, 9) 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
convbn Ran
ran; our contract draft was wrong, not the code
Nicholasli1995/SNVC/snvc/models/vernier.py
code served (permissive licence) · get_code("621990dde69e9fb2")
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 Sequential. 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)

ARGS = {
    "in_planes": 3,
    "out_planes": 64,
    "kernel_size": 3,
    "stride": 1,
    "pad": 1,
    "dilation": 1,
    "gn": False,
    "groups": 32
}
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 Sequential — a Sequential output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Sequential output the recorded ndim 1 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
convbn_3d Ran
ran; our contract draft was wrong, not the code
Nicholasli1995/SNVC/snvc/models/vernier.py
code served (permissive licence) · get_code("7a0e35eeadc577b2")
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 Sequential. 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)

ARGS = {
    "in_planes": 3,
    "out_planes": 16,
    "kernel_size": 3,
    "stride": 1,
    "pad": 1,
    "dilation": 1,
    "gn": False,
    "groups": 32
}
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 Sequential — a Sequential output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Sequential output the recorded ndim 1 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
get_cam_cord Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
Nicholasli1995/SNVC/snvc/models/vernier.py
code served (permissive licence) · get_code("4e5b2125ab45ed44")
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, 9) 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

# Construct valid arguments for get_cam_cord
# The function expects box_3d which is a numpy array with 7 elements:
# [h, w, l, x, y, z, ry]
# where h, w, l are dimensions and x, y, z are coordinates and ry is rotation angle

box_3d = np.array([
    1.5,      # h (height)
    2.0,      # w (width)
    3.0,      # l (length)
    10.0,     # x (position)
    5.0,      # y (position)
    15.0,     # z (position)
    0.5       # ry (rotation angle in radians)
], dtype=np.float32)

ARGS = {
    'box_3d': box_3d
}
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, 9) — shape (3, 9) 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_fc_model Ran
ran on a synthesized input
Nicholasli1995/SNVC/snvc/models/vernier.py
code served (permissive licence) · get_code("1a1214ec82bbb3bf")
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 FCModel. 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)

ARGS = {}
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 FCModel — a FCModel output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a FCModel 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
get_hg_down_sample Ran
ran; our contract draft was wrong, not the code
Nicholasli1995/SNVC/snvc/models/vernier.py
code served (permissive licence) · get_code("9f266410e35fa4da")
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 Sequential. 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)

ARGS = {
    "channel_in": 16,
    "channel_out": 32,
    "gn": False,
    "downsample": True,
}
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 Sequential — a Sequential 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
get_hg_down_sample_2d Ran
ran; our contract draft was wrong, not the code
Nicholasli1995/SNVC/snvc/models/vernier.py
code served (permissive licence) · get_code("8f3e2407776c028d")
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 Sequential. 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)

ARGS = {
    "channel_in": 3,
    "channel_out": 64,
    "gn": False,
    "downsample": True,
}
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 Sequential — a Sequential 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
get_hg_up_sample Ran
ran; our contract draft was wrong, not the code
Nicholasli1995/SNVC/snvc/models/vernier.py
code served (permissive licence) · get_code("6e03fb539321a405")
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 Sequential. 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)

ARGS = {
    "channel_in": 64,
    "channel_out": 32,
    "gn": 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 Sequential — a Sequential output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Sequential output the recorded ndim 1 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
get_hg_up_sample_2d Ran
ran; our contract draft was wrong, not the code
Nicholasli1995/SNVC/snvc/models/vernier.py
code served (permissive licence) · get_code("b49abc0a36bfd171")
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 Sequential. 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)

ARGS = {
    "channel_in": 64,
    "channel_out": 32,
    "gn": 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 Sequential — a Sequential output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Sequential output the recorded ndim 1 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
hourglass Ran
ran on a synthesized input · constructed an object; no result was computed
Nicholasli1995/SNVC/snvc/models/vernier.py
code served (permissive licence) · get_code("cbc0b5237a72da12")
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, hourglass: 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
import torch.nn.functional as F

torch.manual_seed(0)

# The hourglass class takes inplanes and gn as parameters
# inplanes: int representing input channels
# gn: bool flag for using GroupNorm instead of BatchNorm

ARGS = {
    "inplanes": 8,
    "gn": 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 hourglass — 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 hourglass 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
hourglass2d Ran
ran on a synthesized input · constructed an object; no result was computed
Nicholasli1995/SNVC/snvc/models/vernier.py
code served (permissive licence) · get_code("9102f1d1f0a4a1f0")
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, hourglass2d: 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)

ARGS = {
    'inplanes': 16,
    'gn': 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 hourglass2d — 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 hourglass2d 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
hourglass2d_downsample_16 Ran
ran on a synthesized input · constructed an object; no result was computed
Nicholasli1995/SNVC/snvc/models/vernier.py
code served (permissive licence) · get_code("6b233193efdfa8b4")
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, hourglass2d_downsample_16: 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)

# The hourglass2d_downsample_16 class takes inplanes and gn as parameters
# inplanes: int - number of input planes
# gn: bool - whether to use GroupNorm instead of BatchNorm

ARGS = {
    "inplanes": 16,
    "gn": False,
}

# Verify the setup by instantiating and doing a forward pass
if __name__ == "__main__":
    model = nn.Module()
    model.__class__ = type('hourglass2d_downsample_16', (nn.Module,), {
        '__init__': lambda self, inplanes, gn=False: (
            super(type(self), self).__init__(),
            setattr(self, 'conv1', nn.Sequential(
                nn.Conv2d(inplanes, inplanes * 2, kernel_size=3, stride=2, padding=1, dilation=1, bias=False),
                nn.BatchNorm2d(inplanes * 2)
            )),
        ),
        'forward': lambda self, x: x
    })
    
    # Create actual model instance
    model = nn.Sequential(
        nn.Conv2d(ARGS["inplanes"], ARGS["inplanes"] * 2, kernel_size=3, stride=2, padding=1, bias=False),
        nn.BatchNorm2d(ARGS["inplanes"] * 2)
    )
    
    # Test with small input tensor
    x = torch.randn(2, ARGS["inplanes"], 64, 64)
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 hourglass2d_downsample_16 — 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 hourglass2d_downsample_16 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
hourglass_downsample_16 Ran
ran on a synthesized input · constructed an object; no result was computed
Nicholasli1995/SNVC/snvc/models/vernier.py
code served (permissive licence) · get_code("472621b08b760d21")
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, hourglass_downsample_16: 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)

# Create a small input tensor: (batch=2, channels=4, depth=4, height=4, width=4)
input_tensor = torch.randn(2, 4, 4, 4, 4)

# The hourglass_downsample_16 requires:
# - inplanes: int (number of input planes)
# - gn: bool (whether to use GroupNorm instead of BatchNorm)

ARGS = {
    "inplanes": 4,
    "gn": 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 hourglass_downsample_16 — 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 hourglass_downsample_16 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
plot_lines Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
Nicholasli1995/SNVC/snvc/models/vernier.py
code served (permissive licence) · get_code("fefdb1ce9ce665eb")
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 Line2D. 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 numpy as np
import matplotlib.pyplot as plt
from mpl_toolkits.mplot3d import Axes3D

torch.manual_seed(0)

# Create a simple 2D case
points_2d = torch.tensor([
    [0.0, 0.0],
    [1.0, 1.0],
    [2.0, 0.5],
    [3.0, 1.5]
], dtype=torch.float32)

connections_2d = torch.tensor([
    [0, 1],
    [1, 2],
    [2, 3]
], dtype=torch.long)

# Create figure with 2D subplot
fig, ax = plt.subplots()

ARGS = {
    'ax': ax,
    'points': points_2d,
    'connections': connections_2d,
    'dimension': 2,
    'lw': 4,
    'c': 'k',
    'linestyle': '-',
    'alpha': 0.8,
    'add_index': False
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
matplotlib, mpl_toolkits, numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Line2D — a Line2D output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Line2D 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
unnormalize Ran
ran; our contract draft was wrong, not the code
Nicholasli1995/SNVC/snvc/models/vernier.py
code served (permissive licence) · get_code("27710015939338e1")
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 float, shape (4, 2, 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 torch
import numpy as np

torch.manual_seed(0)

# Create a small stand-in config class with img_mean and img_std attributes
class Config:
    def __init__(self):
        self.img_mean = [0.485, 0.456, 0.406]
        self.img_std = [0.229, 0.224, 0.225]

# Create a tensor with shape (3, 4, 2) so that after permute(1, 2, 0) it becomes (4, 2, 3)
# This matches the expected output shape for an image with 3 channels
tensor = torch.randn(3, 4, 2)

cfg = Config()

ARGS = {
    "tensor": tensor,
    "cfg": cfg
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — 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 float, shape (4, 2, 3) — shape (4, 2, 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
VernierScale Not yet run
unverified (harvested, no recorded run)
Nicholasli1995/SNVC/snvc/models/vernier.py
code served (permissive licence) · get_code("274db02832658cfc")
How it ran

no invocation recorded.

get_feat_extraction Not yet run
unverified (harvested, no recorded run)
Nicholasli1995/SNVC/snvc/models/vernier.py
code served (permissive licence) · get_code("fd9a001b6adf12fc")
How it ran

no invocation recorded.

get_model Not yet run
unverified (harvested, no recorded run)
Nicholasli1995/SNVC/snvc/models/vernier.py
code served (permissive licence) · get_code("9374704e49ec804f")
How it ran

no invocation recorded.

plot_3d_bbox Not yet run
unverified (harvested, no recorded run)
Nicholasli1995/SNVC/snvc/models/vernier.py
code served (permissive licence) · get_code("39daff41d1789e4d")
How it ran

no invocation recorded.

plot_3d_points Not yet run
unverified (harvested, no recorded run)
Nicholasli1995/SNVC/snvc/models/vernier.py
code served (permissive licence) · get_code("ed47bc08c084e3b1")
How it ran

no invocation recorded.

set_3d_axe_limits Not yet run
unverified (harvested, no recorded run)
Nicholasli1995/SNVC/snvc/models/vernier.py
code served (permissive licence) · get_code("e58d3fd7b0658d63")
How it ran

no invocation recorded.

test_visualization Not yet run
unverified (harvested, no recorded run)
Nicholasli1995/SNVC/snvc/models/vernier.py
code served (permissive licence) · get_code("dbc5584e69ef257d")
How it ran

no invocation recorded.

Also present in other papers' code

Of 28 samples lifted from this paper's linked repositories, 28 carry a function identity and 0 do not. Of the 28 distinct functions with an identity, 1 are also present in other papers whose code we hold (1 with at least one paper dated earlier, 0 the same month, 0 later) and 27 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
basicdownsample
brick of this paper's assembly: helper
1 other paper
1 dated earlier
2007.09548 (2020-07)
Kinematic 3D Object Detection in Monocular Video

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 propose a new object-centric framework for learningbased stereo 3D object detection. Previous studies build scene-centric representations that do not consider the significant variation among outdoor instances and thus lack the flexibility and functionalities that an instance-level model can offer. We build such an instance-level model by formulating and tackling a local update problem, i.e., how to predict a refined update given an initial 3D cuboid guess. We demonstrate how solving this problem can complement scene-centric approaches in (i) building a coarse-to-fine multi-resolution system, (ii) performing model-agnostic object location refinement and, (iii) conducting stereo 3D tracking-by-detection. Extensive experiments demonstrate the effectiveness of our approach, which achieves state-of-the-art performance on the KITTI benchmark. Code and pre-trained models are available at https://github.com/Nicholasli1995/SNVC.

For agents

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

get_harvested_code_for_paper("2203.11018")
get_code_for_paper("2203.11018")
have("2203.11018")

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

curl https://syntology.ai/api/ran/2203.11018.json

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

[![Syntology run record](https://syntology.ai/api/ran/2203.11018.svg)](https://syntology.ai/paper/2203.11018)

Connect an agent — have() is free.