SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2304.05758 · 2023

Best Practices for 2-Body Pose Forecasting

arXiv · PDF · Open in the Atlas

Code that ran

16 of 17 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 13 have no contract verdict; 3 failures of our instrument; 1 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/2304.05758.json · badge: /api/ran/2304.05758.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
edodema/BestPractices2Body canonical (classifier label)
classifier basis: census
16 ran · 1 no recorded run
FunctionStatusWhere it lives
ang2joint Ran
ran; our contract draft was wrong, not the code
edodema/BestPractices2Body/src/utils/angle_to_joint.py
pointer only (licence: MIT); the stored copy get_code serves, from fanglaosi/skeleton-in-context, is pointer only (licence: NONE) · get_code("c0984e125450806e")
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 fanglaosi/skeleton-in-context, and GitHub detected no licence in that repository when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

extant_file Ran
ran on a synthesized input
edodema/BestPractices2Body/src/utils/pyt_utils.py
code served (permissive licence) · get_code("81e27319d084659d")
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 os
import tempfile
import argparse

def extant_file(x):
    """
    'Type' for argparse - checks that file exists but does not open.
    """
    if not os.path.exists(x):
        raise argparse.ArgumentTypeError('{0} does not exist'.format(x))
    return x


# Create a temporary file that exists
with tempfile.NamedTemporaryFile(delete=False) as tmp:
    temp_file_path = tmp.name

ARGS = {
    'x': temp_file_path
}
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 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
fkl_torch Ran
ran on a synthesized input
edodema/BestPractices2Body/src/utils/misc.py
code served (permissive licence) · get_code("1edbc792f56ab438")
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 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 torch
import numpy as np

torch.manual_seed(0)

# Understand what the function needs:
# - rotmat: tensor of shape (n, j_n*3*3) containing rotation matrices
# - parent: array of parent indices for each joint
# - offset: array of shape (j_n, 3) with joint offsets
# - rotInd: not used in function body
# - expmapInd: not used in function body

# Set up small dimensions
batch_size = 2
num_joints = 4
joint_dim = 3

# rotmat: (batch_size, num_joints * 3 * 3) = (2, 36)
# This will be reshaped to (n, j_n, 3, 3) in the function
rotmat = torch.eye(3).unsqueeze(0).repeat(batch_size * num_joints, 1, 1).reshape(batch_size, num_joints * 3 * 3).float()

# parent: array of parent joint indices, length num_joints
# parent[0] should be 0 or -1 (root), others > 0
parent = np.array([0, 1, 2, 3], dtype=np.int64)

# offset: (num_joints, 3) array of joint offsets
offset = np.random.randn(num_joints, joint_dim).astype(np.float32)

# rotInd: not used, can be any value
rotInd = np.arange(num_joints * 3).reshape(num_joints, 3)

# expmapInd: not used, can be any value
expmapInd = np.arange(num_joints * 3).reshape(num_joints, 3)

ARGS = {
    "rotmat": rotmat,
    "parent": parent,
    "offset": offset,
    "rotInd": rotInd,
    "expmapInd": expmapInd,
}
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 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
gen_velocity Ran
ran on a synthesized input
edodema/BestPractices2Body/src/utils/util.py
code served (permissive licence) · get_code("044ab013d7abd46a")
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, 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

torch.manual_seed(0)

# Create a small tensor with shape (batch=2, sequence=4)
# This represents a sequence of positions where we compute velocities
m = torch.randn(2, 4)

ARGS = {
    "m": m
}
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, 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
get_dct_matrix Ran
ran on a synthesized input
edodema/BestPractices2Body/src/utils/util.py
code served (permissive licence) · get_code("68de7a44b4f27b2d")
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, ndim 3, 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 numpy as np

ARGS = {
    "N": 8,
}
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, ndim 3, dtype family float — a tuple 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
normExPI_2p_by_frame Ran
ran on a synthesized input
edodema/BestPractices2Body/src/datasets/data_utils_pi3d.py
code served (permissive licence) · get_code("ba3289fec7009fc0")
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, 36) 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

# Analyze the function:
# normExPI_2p_by_frame takes seq: np.ndarray
# It expects seq.shape to be (nb, dim) where nb is number of frames, dim is flattened dimension
# For each frame i, it reshapes seq[i] to (-1, 3), so dim must be divisible by 3
# It accesses img[10], img[11], img[3], so we need at least 12 points (indices 0-11)
# Each point has 3 coordinates, so minimum 12*3 = 36 dimensions
# Let's use 12 points * 3 coords = 36 dims, and 2 frames for batch

# Create a small ExPI sequence dataset
# 2 frames, each flattened to 36 dimensions (12 3D points)
nb_frames = 2
nb_points = 12
dims_per_point = 3
total_dim = nb_points * dims_per_point

# Create synthetic ExPI data
np.random.seed(0)
seq = np.random.randn(nb_frames, total_dim).astype(np.float64)

# Ensure the data is valid by creating reasonable 3D points
# Initialize with points that form a reasonable plane structure
for i in range(nb_frames):
    img_points = seq[i].reshape((nb_points, 3))
    # Make points roughly coplanar and structured
    for j in range(nb_points):
        img_points[j, 0] = np.random.randn()  # x
        img_points[j, 1] = np.random.randn()  # y
        img_points[j, 2] = np.random.randn() * 0.1  # z (small variance)
    seq[i] = img_points.reshape(total_dim)

ARGS = {
    "seq": seq
}
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, 36) — shape (2, 36) 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
normExPI_xoz Ran
ran on a synthesized input
edodema/BestPractices2Body/src/datasets/data_utils_pi3d.py
code served (permissive licence) · get_code("d349fb0b9c8391b7")
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, 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

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

# Create small test data for the function
# img: Input ExPI data - should be a 2D array where each row is a 3D point (x, y, z)
img = np.array([
    [0.5, 0.2, 0.3],
    [1.2, 0.8, 0.5]
], dtype=np.float64)

# P0: Origin point (3D)
P0 = np.array([0.0, 0.0, 0.0], dtype=np.float64)

# P1: Point defining x-axis direction (3D)
# Should be different from P0
P1 = np.array([1.0, 0.0, 0.0], dtype=np.float64)

# P2: Point defining the xoz plane (3D)
# Should be different from P0 and P1, and not collinear
P2 = np.array([0.0, 0.0, 1.0], dtype=np.float64)

ARGS = {
    'img': img,
    'P0': P0,
    'P1': P1,
    'P2': P2
}
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, 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
readCSVasFloat Ran
ran on a synthesized input
edodema/BestPractices2Body/src/datasets/data_utils_pi3d.py
code served (permissive licence) · get_code("2382970d969a42d1")
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 tempfile
import os
import numpy as np

# Create a temporary CSV file with sample data
temp_dir = tempfile.mkdtemp()
csv_filename = os.path.join(temp_dir, "test_data.csv")

# Write a sample CSV file with header and data
with open(csv_filename, 'w') as f:
    f.write("col1,col2,col3,col4\n")
    f.write("1.0,2.0,3.0,4.0\n")
    f.write("5.0,6.0,7.0,8.0\n")
    f.write("9.0,10.0,11.0,12.0\n")

ARGS = {
    "filename": csv_filename,
    "with_key": True
}
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
rigid_align Ran
ran on a synthesized input
edodema/BestPractices2Body/src/utils/rigid_align.py
code served (permissive licence) · get_code("6f649cf48e2412ce")
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 (5, 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 simple 3D point clouds for testing
# A: source points, B: target points
# Use distinct sizes: 5 points with 3D coordinates

np.random.seed(0)

# Source points (5 points in 3D space)
A = np.random.randn(5, 3).astype(np.float32)

# Target points (5 points in 3D space)
# Create B as a transformed version of A to ensure alignment is meaningful
rotation_angle = np.pi / 6
cos_a, sin_a = np.cos(rotation_angle), np.sin(rotation_angle)
rotation_matrix = np.array([
    [cos_a, -sin_a, 0],
    [sin_a, cos_a, 0],
    [0, 0, 1]
], dtype=np.float32)
translation = np.array([1.0, 2.0, 3.0], dtype=np.float32)
B = (A @ rotation_matrix.T) + translation

ARGS = {
    "A": A,
    "B": B
}
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 (5, 3) — shape (5, 3) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
rigid_align_torch Ran
ran on a synthesized input
edodema/BestPractices2Body/src/utils/rigid_align.py
code served (permissive licence) · get_code("718c13d3298cc856")
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 (2, 4, 36, 3) as a later run recorded it. The image it ran in was not recorded. The code mentions cuda, and the harness aliases cuda to the CPU for every sample; whether this run needed that is not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import numpy as np

torch.manual_seed(0)

# Create small test tensors
# Shape: (batch_size, num_frames, num_keypoints, 3)
# Using batch=2, num_frames=4, num_keypoints=36, coords=3
A = torch.randn(2, 4, 36, 3, dtype=torch.float32)
B = torch.randn(2, 4, 36, 3, dtype=torch.float32)

ARGS = {
    "A": A,
    "B": B,
}
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 4, dtype family float, shape (2, 4, 36, 3) — shape (2, 4, 36, 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
rigid_transform_3D Ran
ran on a synthesized input
edodema/BestPractices2Body/src/utils/rigid_align.py
code served (permissive licence) · get_code("010398fec721a847")
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 source and target point clouds
# A: source points (4 points in 3D space)
A = np.array([
    [1.0, 2.0, 3.0],
    [4.0, 5.0, 6.0],
    [7.0, 8.0, 9.0],
    [10.0, 11.0, 12.0]
], dtype=np.float64)

# B: target points (4 points in 3D space)
# Apply a known rigid transformation to A to create B
rotation_true = np.array([
    [0.866, -0.5, 0.0],
    [0.5, 0.866, 0.0],
    [0.0, 0.0, 1.0]
], dtype=np.float64)
translation_true = np.array([1.0, 2.0, 3.0], dtype=np.float64)

B = (A @ rotation_true.T) + translation_true

ARGS = {
    "A": A,
    "B": B
}
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
rodrigues Ran
ran; our contract draft was wrong, not the code
edodema/BestPractices2Body/src/utils/angle_to_joint.py
pointer only (licence: MIT); the stored copy get_code serves, from fanglaosi/skeleton-in-context, is pointer only (licence: NONE) · get_code("d7032acefdc1b55a")
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 fanglaosi/skeleton-in-context, and GitHub detected no licence in that repository when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

rotmat2euler_torch Ran
ran on a synthesized input
edodema/BestPractices2Body/src/utils/misc.py
code served (permissive licence) · get_code("13250dfe061caa67")
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, 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 batch of rotation matrices
# We need N x 3 x 3 tensors
# Let's create 2 rotation matrices (batch size 2)

# Create some valid rotation matrices
# For simplicity, we'll create identity and a rotated matrix
batch_size = 2
R1 = torch.eye(3).float().unsqueeze(0)  # Identity matrix
R2 = torch.tensor([
    [0.866, -0.5, 0.0],
    [0.5, 0.866, 0.0],
    [0.0, 0.0, 1.0]
]).float().unsqueeze(0)  # Rotation around z-axis

R = torch.cat([R1, R2], dim=0)  # Shape: (2, 3, 3)

ARGS = {
    'R': R
}
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, 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
rotmat2quat_torch Ran
ran on a synthesized input
edodema/BestPractices2Body/src/utils/misc.py
code served (permissive licence) · get_code("963cf42da1883209")
How it ran

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

import torch

torch.manual_seed(0)

# Create a batch of 2 rotation matrices (3x3 each)
# Valid rotation matrices should be orthogonal, but for testing purposes
# we'll create matrices that are close to valid rotations
batch_size = 2

# Create simple rotation-like matrices
R = torch.tensor([
    [[1.0, 0.0, 0.0],
     [0.0, 1.0, 0.0],
     [0.0, 0.0, 1.0]],
    [[0.866, -0.5, 0.0],
     [0.5, 0.866, 0.0],
     [0.0, 0.0, 1.0]]
], dtype=torch.float32)

ARGS = {
    "R": R
}
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
update_lr_multistep Ran
ran on a synthesized input
edodema/BestPractices2Body/src/utils/util.py
code served (permissive licence) · get_code("ebbcbb831700b43b")
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, 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 torch
import torch.nn as nn

torch.manual_seed(0)

# Create a simple model
model = nn.Linear(4, 3)

# Create an optimizer
optimizer = torch.optim.SGD(model.parameters(), lr=0.0003)

# Define arguments for update_lr_multistep
ARGS = {
    'nb_iter': 15000,
    'optimizer': optimizer,
}
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 tuple, ndim 1, dtype family float — a tuple 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
with_zeros Ran
ran; our contract draft was wrong, not the code
edodema/BestPractices2Body/src/utils/angle_to_joint.py
pointer only (licence: MIT); the stored copy get_code serves, from fanglaosi/skeleton-in-context, is pointer only (licence: NONE) · get_code("6588622485d7bdc2")
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 fanglaosi/skeleton-in-context, and GitHub detected no licence in that repository when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

get_logger Not yet run
unverified (harvested, no recorded run)
edodema/BestPractices2Body/src/utils/logger.py
code served (permissive licence) · get_code("7660a1c90472fa55")
How it ran

no invocation recorded.

Also present in other papers' code

Of 17 samples lifted from this paper's linked repositories, 17 carry a function identity and 0 do not. Of the 17 distinct functions with an identity, 7 are also present in other papers whose code we hold (7 with at least one paper dated earlier, 0 the same month, 3 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
ang2joint 3 other papers
1 dated earlier, 2 later
2207.01567 (2022-07)
Back to MLP: A Simple Baseline for Human Motion Prediction
2308.04808 (2023-08)
Joint-Relation Transformer for Multi-Person Motion Prediction
2312.03703 (2023-12)
Skeleton-in-Context: Unified Skeleton Sequence Modeling with In-Context Learning
rodrigues 3 other papers
1 dated earlier, 2 later
2207.01567 (2022-07)
Back to MLP: A Simple Baseline for Human Motion Prediction
2308.04808 (2023-08)
Joint-Relation Transformer for Multi-Person Motion Prediction
2312.03703 (2023-12)
Skeleton-in-Context: Unified Skeleton Sequence Modeling with In-Context Learning
with_zeros 3 other papers
1 dated earlier, 2 later
2207.01567 (2022-07)
Back to MLP: A Simple Baseline for Human Motion Prediction
2308.04808 (2023-08)
Joint-Relation Transformer for Multi-Person Motion Prediction
2312.03703 (2023-12)
Skeleton-in-Context: Unified Skeleton Sequence Modeling with In-Context Learning
extant_file 1 other paper
1 dated earlier
2207.01567 (2022-07)
Back to MLP: A Simple Baseline for Human Motion Prediction
fkl_torch 1 other paper
1 dated earlier
2207.01567 (2022-07)
Back to MLP: A Simple Baseline for Human Motion Prediction
rotmat2euler_torch 1 other paper
1 dated earlier
2207.01567 (2022-07)
Back to MLP: A Simple Baseline for Human Motion Prediction
rotmat2quat_torch 1 other paper
1 dated earlier
2207.01567 (2022-07)
Back to MLP: A Simple Baseline for Human Motion Prediction

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

The task of collaborative human pose forecasting stands for predicting the future poses of multiple interacting people, given those in previous frames. Predicting two people in interaction, instead of each separately, promises better performance, due to their body-body motion correlations. But the task has remained so far primarily unexplored. In this paper, we review the progress in human pose forecasting and provide an in-depth assessment of the single-person practices that perform best for 2-body collaborative motion forecasting. Our study confirms the positive impact of frequency input representations, space-time separable and fully-learnable interaction adjacencies for the encoding GCN and FC decoding. Other single-person practices do not transfer to 2-body, so the proposed best ones do not include hierarchical body modeling or attention-based interaction encoding. We further contribute a novel initialization procedure for the 2-body spatial interaction parameters of the encoder, which benefits performance and stability. Altogether, our proposed 2-body pose forecasting best practices yield a performance improvement of 21.9% over the state-of-the-art on the most recent ExPI dataset, whereby the novel initialization accounts for 3.5%. See our project page at https://www.pinlab.org/bestpractices2body

For agents

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

get_harvested_code_for_paper("2304.05758")
get_code_for_paper("2304.05758")
have("2304.05758")

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

curl https://syntology.ai/api/ran/2304.05758.json

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

[![Syntology run record](https://syntology.ai/api/ran/2304.05758.svg)](https://syntology.ai/paper/2304.05758)

Connect an agent — have() is free.