SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2407.07532 · NeurIPS · 2024

Neural Localizer Fields for Continuous 3D Human Pose and Shape Estimation

Gerard Pons-Moll, István Sárándi

arXiv · PDF · Open in the Atlas

Code that ran

18 of 20 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; 5 honoured the contract we drafted, 4 violated it and 0 have no contract verdict; 9 failures of our instrument; 2 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/2407.07532.json · badge: /api/ran/2407.07532.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
isarandi/nlf canonical (classifier label)
classifier basis: census
10 ran · 1 no recorded run
isarandi/smplfitter application (classifier label)
classifier basis: classifier_2025
8 ran · 1 no recorded run
FunctionStatusWhere it lives
align_unit_vectors Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
isarandi/smplfitter/src/smplfitter/jax/bodyfitter.py
code served (permissive licence) · get_code("482f6cb812d2864a")
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 ArrayImpl, ndim 3, dtype family float, shape (2, 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 jax.numpy as jnp

# The function align_unit_vectors takes two parameters: a and b
# Both should be unit vectors (or at least vectors that can be normalized)
# The function uses cross product, dot product, and norm operations
# Expected shape: (..., 3) where ... is batch dimensions

# Create small test vectors
# Batch size 2, each vector has 3 components (since cross product is for 3D vectors)
a = jnp.array([
    [1.0, 0.0, 0.0],  # unit vector along x-axis
    [0.0, 1.0, 0.0]   # unit vector along y-axis
])

b = jnp.array([
    [0.0, 1.0, 0.0],  # unit vector along y-axis
    [0.0, 0.0, 1.0]   # unit vector along z-axis
])

ARGS = {
    "a": a,
    "b": b
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
jax — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ArrayImpl, ndim 3, dtype family float, shape (2, 3, 3) — shape (2, 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
check_batch_size Ran
ran, honoured the contract we drafted
isarandi/nlf/nlf/mmapped_smpl.py
code served (permissive licence) · get_code("b8ac5a72d27f4f69")
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 numpy as np
import torch

torch.manual_seed(0)

# Create small numpy arrays with batch size 2
# pose_rotvecs: shape (2, 3) - batch of rotation vectors
pose_rotvecs = np.random.randn(2, 3)

# shape_betas: shape (2, 4) - batch of shape parameters
shape_betas = np.random.randn(2, 4)

# trans: shape (2, 3) - batch of translations
trans = np.random.randn(2, 3)

# rel_rotmats: shape (2, 3, 3) - batch of rotation matrices
rel_rotmats = np.random.randn(2, 3, 3)

ARGS = {
    "pose_rotvecs": pose_rotvecs,
    "shape_betas": shape_betas,
    "trans": trans,
    "rel_rotmats": rel_rotmats,
}
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 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
choice Ran
ran; our contract draft was wrong, not the code
isarandi/nlf/nlf/common/util.py
code served (permissive licence) · get_code("9c6e866d219206ab")
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 numpy as np

# Create a numpy random generator
rng = np.random.default_rng(seed=42)

# Create a simple list of items to choose from
items = ["apple", "banana", "cherry", "date", "elderberry"]

# Define arguments for the choice function
ARGS = {
    "items": items,
    "rng": rng
}
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 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
divide_no_nan Ran
ran, honoured the contract we drafted
isarandi/smplfitter/src/smplfitter/jax/bodyfitter.py
code served (permissive licence) · get_code("0ec9b9a6711cb9f9")
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 ArrayImpl, 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 jax.numpy as jnp
import numpy as np

# Create small test arrays for the divide_no_nan function
a = jnp.array([[1.0, 2.0, 3.0, 4.0],
                [5.0, 6.0, 7.0, 8.0]])

b = jnp.array([[2.0, 0.0, 1.0, 4.0],
                [0.0, 3.0, 2.0, 0.0]])

ARGS = {
    "a": a,
    "b": b
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
jax, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ArrayImpl, 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
fit_scale_and_translation Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
isarandi/smplfitter/src/smplfitter/jax/bodyfitter.py
code served (permissive licence) · get_code("1df08aadfba3944e")
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 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 jax.numpy as jnp
import numpy as np

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

# Create small test data
batch_size = 2
num_vertices = 4
num_joints = 3
coord_dim = 3

# Create target and reference vertices
target_vertices = jnp.array(np.random.randn(batch_size, num_vertices, coord_dim).astype(np.float32))
reference_vertices = jnp.array(np.random.randn(batch_size, num_vertices, coord_dim).astype(np.float32))

# Create target and reference joints
target_joints = jnp.array(np.random.randn(batch_size, num_joints, coord_dim).astype(np.float32))
reference_joints = jnp.array(np.random.randn(batch_size, num_joints, coord_dim).astype(np.float32))

# Create vertex weights (batch_size, num_vertices)
vertex_weights = jnp.array(np.random.rand(batch_size, num_vertices).astype(np.float32))

# Create joint weights (batch_size, num_joints)
joint_weights = jnp.array(np.random.rand(batch_size, num_joints).astype(np.float32))

ARGS = {
    "target_vertices": target_vertices,
    "reference_vertices": reference_vertices,
    "target_joints": target_joints,
    "reference_joints": reference_joints,
    "vertex_weights": vertex_weights,
    "joint_weights": joint_weights,
    "scale": True,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
jax, 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
fixed_padding_layer Ran
ran; our contract draft was wrong, not the code
isarandi/nlf/nlf/pt/backbones/efficientnet.py
code served (permissive licence) · get_code("2f6a6eccfb577d4a")
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 ZeroPad2d. 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

torch.manual_seed(0)

ARGS = {
    "kernel_size": 3,
    "rate": 1,
    "shifts": (0, 0)
}
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 ZeroPad2d — a ZeroPad2d output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a ZeroPad2d 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_structuring_element Ran
ran; our contract draft was wrong, not the code
isarandi/nlf/nlf/common/improc.py
code served (permissive licence) · get_code("da4d2a8a96f9b163")
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 int, shape (5, 5) 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 functools
import cv2
import numpy as np

@functools.lru_cache()
def get_structuring_element(shape, ksize, anchor=None):
    if not isinstance(ksize, tuple):
        ksize = (ksize, ksize)
    return cv2.getStructuringElement(shape, ksize, anchor)

# The function uses cv2.getStructuringElement which expects:
# - shape: an integer constant like cv2.MORPH_RECT, cv2.MORPH_ELLIPSE, cv2.MORPH_CROSS
# - ksize: a tuple of (width, height) for kernel size
# - anchor: optional tuple of (x, y) for anchor point

ARGS = {
    "shape": cv2.MORPH_RECT,
    "ksize": (5, 5),
    "anchor": None
}
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 2, dtype family int, shape (5, 5) — shape (5, 5) 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
image_extents Ran
ran, honoured the contract we drafted
isarandi/nlf/nlf/common/improc.py
code served (permissive licence) · get_code("adae3a87f2d2d5f5")
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 tempfile
import PIL.Image
import numpy as np

# Create a temporary image file
with tempfile.NamedTemporaryFile(suffix='.png', delete=False) as tmp:
    # Create a small test image with dimensions 640x480
    img = PIL.Image.new('RGB', (640, 480), color='red')
    img.save(tmp.name)
    filepath = tmp.name

ARGS = {
    'filepath': filepath
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
PIL, 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
is_hand_joint Ran
ran, violated the contract we drafted
isarandi/nlf/nlf/pt/models/nlf_model.py
code served (permissive licence) · get_code("a6acc64679279c78")
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 torch

torch.manual_seed(0)

ARGS = {
    'name': 'thumb_joint_01'
}
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 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
lstsq Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
isarandi/smplfitter/src/smplfitter/jax/bodyfitter.py
code served (permissive licence) · get_code("90f81d3dcc51571e")
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 ArrayImpl, ndim 3, dtype family float, shape (2, 4, 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 jax.numpy as jnp
import numpy as np

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

# Define dimensions
batch_size = 2
n_points = 5
n_params = 4
n_outputs = 3

# Create matrix: shape (batch, n_points, n_params)
matrix = jnp.array(np.random.randn(batch_size, n_points, n_params), dtype=jnp.float32)

# Create rhs: shape (batch, n_points, n_outputs)
rhs = jnp.array(np.random.randn(batch_size, n_points, n_outputs), dtype=jnp.float32)

# Create weights: shape (batch, n_points)
weights = jnp.array(np.random.rand(batch_size, n_points) + 0.1, dtype=jnp.float32)

# Create l2_regularizer: shape (n_params,)
l2_regularizer = jnp.array(np.random.rand(n_params) * 0.1, dtype=jnp.float32)

# Create l2_regularizer_rhs: shape (batch, n_params, n_outputs)
l2_regularizer_rhs = jnp.array(np.random.randn(batch_size, n_params, n_outputs) * 0.01, dtype=jnp.float32)

# shared flag
shared = False

ARGS = {
    "matrix": matrix,
    "rhs": rhs,
    "weights": weights,
    "l2_regularizer": l2_regularizer,
    "l2_regularizer_rhs": l2_regularizer_rhs,
    "shared": shared,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
jax, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ArrayImpl, ndim 3, dtype family float, shape (2, 4, 3) — shape (2, 4, 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
lstsq_partial_share Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
isarandi/smplfitter/src/smplfitter/jax/bodyfitter.py
code served (permissive licence) · get_code("ec92e05c976f40a8")
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 ArrayImpl, ndim 3, dtype family float, shape (2, 4, 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 jax.numpy as jnp
import numpy as np

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

# Define small dimensions for testing
batch_size = 2
n_points = 5
n_params = 4
n_outputs = 3
n_shared = 2

# Create test data
matrix = jnp.array(np.random.randn(batch_size, n_points, n_params), dtype=jnp.float32)
rhs = jnp.array(np.random.randn(batch_size, n_points, n_outputs), dtype=jnp.float32)
weights = jnp.array(np.random.rand(batch_size, n_points) + 0.1, dtype=jnp.float32)
l2_regularizer = jnp.array(np.random.rand(n_params) * 0.1, dtype=jnp.float32)
l2_regularizer_rhs = jnp.array(np.random.randn(batch_size, n_params, n_outputs), dtype=jnp.float32)

ARGS = {
    "matrix": matrix,
    "rhs": rhs,
    "weights": weights,
    "l2_regularizer": l2_regularizer,
    "l2_regularizer_rhs": l2_regularizer_rhs,
    "n_shared": n_shared,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
jax, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ArrayImpl, ndim 3, dtype family float, shape (2, 4, 3) — shape (2, 4, 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
mat2rotvec Ran
ran, violated the contract we drafted
isarandi/smplfitter/src/smplfitter/jax/bodyfitter.py
code served (permissive licence) · get_code("8239bd5db42a4f6d")
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 ArrayImpl, ndim 2, dtype family float, shape (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 jax.numpy as jnp
import numpy as np

# Create a valid rotation matrix
# A rotation matrix is 3x3 orthogonal with determinant 1
# For simplicity, use identity and a small rotation

def create_rotation_matrix():
    """Create a simple valid rotation matrix"""
    # Identity rotation
    return np.eye(3, dtype=np.float32)

def create_batch_rotation_matrices(batch_size=2):
    """Create a batch of rotation matrices"""
    matrices = []
    for i in range(batch_size):
        # Create identity rotations for simplicity
        matrices.append(create_rotation_matrix())
    return np.stack(matrices, axis=0)

# Create test rotation matrix input
# Shape: (2, 3, 3) - batch of 2, each 3x3 rotation matrix
rotmat_batch = create_batch_rotation_matrices(batch_size=2)
rotmat_jax = jnp.array(rotmat_batch)

ARGS = {
    "rotmat": rotmat_jax
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
jax, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ArrayImpl, ndim 2, dtype family float, shape (2, 3) — shape (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
new_rng Ran
ran; our contract draft was wrong, not the code
isarandi/nlf/nlf/common/util.py
code served (permissive licence) · get_code("6013ccdff44e576d")
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 Generator. 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 a base RNG that has the .integers() method
base_rng = np.random.Generator(np.random.PCG64(42))

# Define ARGS dict with the rng parameter
ARGS = {
    "rng": base_rng
}
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 Generator — a Generator output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Generator 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
proj_SO3 Ran
ran, violated the contract we drafted
isarandi/smplfitter/src/smplfitter/jax/bodyfitter.py
code served (permissive licence) · get_code("8905016ea221a614")
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 ArrayImpl, ndim 3, dtype family float, shape (2, 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 jax.numpy as jnp
import numpy as np

# Create a small batch of 3x3 matrices to project onto SO(3)
# Shape: (2, 3, 3) - batch of 2 matrices, each 3x3
batch_size = 2
A = jnp.array(np.random.randn(batch_size, 3, 3), dtype=jnp.float32)

ARGS = {
    "A": A
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
jax, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ArrayImpl, ndim 3, dtype family float, shape (2, 3, 3) — shape (2, 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
random_uniform_disc Ran
ran, honoured the contract we drafted
isarandi/nlf/nlf/common/util.py
code served (permissive licence) · get_code("9f151790ff5516e9")
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 float, 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 a random number generator
rng = np.random.RandomState(seed=42)

# Construct arguments for random_uniform_disc
ARGS = {
    "rng": rng
}
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 1, dtype family float, 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
rearrange_tf_to_pt Ran
ran, violated the contract we drafted
isarandi/nlf/nlf/pt/convert_ckpt_from_tf.py
code served (permissive licence) · get_code("697f5aa8d9395594")
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 4, dtype family float, shape (5, 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 einops

torch.manual_seed(0)

# Test case 1: 4D tensor with depthwise=False
value_4d = torch.randn(2, 3, 4, 5)  # h=2, w=3, c_in=4, c_out=5

# Test case 2: 4D tensor with depthwise=True
value_4d_dw = torch.randn(2, 3, 4, 5)  # h=2, w=3, c_in=4, c_out=5

# Test case 3: 2D tensor
value_2d = torch.randn(4, 5)  # c_in=4, c_out=5

# Test case 4: 1D tensor
value_1d = torch.randn(4)  # c_in=4

ARGS = {
    "value": value_4d,
    "depthwise": False
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
einops, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 4, dtype family float, shape (5, 4, 2, 3) — shape (5, 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
rotvec2mat Ran
ran, honoured the contract we drafted
isarandi/smplfitter/src/smplfitter/jax/bodyfitter.py
code served (permissive licence) · get_code("b1a2bb53aff45862")
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 ArrayImpl, ndim 3, dtype family float, shape (2, 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 jax.numpy as jnp

# The function rotvec2mat takes a rotation vector and converts it to a rotation matrix
# It expects rotvec to be an array with shape (..., 3) where the last dimension is 3D rotation vectors

# Create a small batch of rotation vectors
# Shape: (2, 3) - batch of 2 rotation vectors, each with 3 components
rotvec = jnp.array([
    [0.1, 0.2, 0.3],
    [0.0, 0.0, 0.0]
], dtype=jnp.float32)

ARGS = {
    "rotvec": rotvec
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
jax — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ArrayImpl, ndim 3, dtype family float, shape (2, 3, 3) — shape (2, 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
scale_align Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
isarandi/nlf/nlf/common/util3d.py
code served (permissive licence) · get_code("4a1c4949c642e768")
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 (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 numpy as np

# Create a small poses array with distinct dimensions for clarity
# Shape: (batch=2, num_poses=3, coordinates=2)
# This allows us to test the axis reduction properly
poses = np.random.randn(2, 3, 2).astype(np.float32)

ARGS = {
    "poses": poses
}
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 3, dtype family float, 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
BodyFitter Not yet run
unverified (harvested, no recorded run)
isarandi/smplfitter/src/smplfitter/jax/bodyfitter.py
code served (permissive licence) · get_code("d437f0383a441aa2")
How it ran

no invocation recorded.

procrustes Not yet run
unverified (harvested, no recorded run)
isarandi/nlf/nlf/common/procrustes.py
code served (permissive licence) · get_code("b9bd6d4461e049e9")
How it ran

no invocation recorded.

Also present in other papers' code

Of 20 samples lifted from this paper's linked repositories, 20 carry a function identity and 0 do not. Of the 20 distinct functions with an identity, 2 are also present in other papers whose code we hold (2 with at least one paper dated earlier, 0 the same month, 0 later) and 18 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
fixed_padding_layer 1 other paper
1 dated earlier
2212.14474 (2022-12)
Learning 3D Human Pose Estimation from Dozens of Datasets using a Geometry-Aware Autoencoder to Bridge Between Skeleton Formats
rearrange_tf_to_pt 1 other paper
1 dated earlier
2212.14474 (2022-12)
Learning 3D Human Pose Estimation from Dozens of Datasets using a Geometry-Aware Autoencoder to Bridge Between Skeleton Formats

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

With the explosive growth of available training data, single-image 3D human modeling is ahead of a transition to a data-centric paradigm. A key to successfully exploiting data scale is to design flexible models that can be supervised from various heterogeneous data sources produced by different researchers or vendors. To this end, we propose a simple yet powerful paradigm for seamlessly unifying different human pose and shape-related tasks and datasets. Our formulation is centered on the ability -both at training and test time -to query any arbitrary point of the human volume, and obtain its estimated location in 3D. We achieve this by learning a continuous neural field of body point localizer functions, each of which is a differently parameterized 3D heatmap-based convolutional point localizer (detector). For generating parametric output, we propose an efficient post-processing step for fitting SMPL-family body models to nonparametric joint and vertex predictions. With this approach, we can naturally exploit differently annotated data sources including mesh, 2D/3D skeleton and dense pose, without having to convert between them, and thereby train large-scale 3D human mesh and skeleton estimation models that considerably outperform the state-of-the-art on several public benchmarks including 3DPW, EMDB, EHF, SSP-3D and AGORA.

For agents

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

get_harvested_code_for_paper("2407.07532")
get_code_for_paper("2407.07532")
have("2407.07532")

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

curl https://syntology.ai/api/ran/2407.07532.json

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

[![Syntology run record](https://syntology.ai/api/ran/2407.07532.svg)](https://syntology.ai/paper/2407.07532)

Connect an agent — have() is free.