SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1902.08570 · 2019

ParticleNet: Jet Tagging via Particle Clouds

arXiv · PDF · Open in the Atlas

Code that ran

19 of 23 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; 1 honoured the contract we drafted, 0 violated it and 15 have no contract verdict; 3 failures of our instrument; 4 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/1902.08570.json · badge: /api/ran/1902.08570.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
hqucms/weaver-core canonical (classifier label)
classifier basis: census
14 ran · 4 no recorded run
StefReck/MEdgeConv community (classifier label)
classifier basis not recorded on the link
4 ran · 0 no recorded run
Jai2500/particlenet reimplementation (classifier label)
classifier basis: classifier_2025
1 ran · 0 no recorded run
FunctionStatusWhere it lives
confusion_matrix Ran
ran on a synthesized input
hqucms/weaver-core/weaver/utils/nn/metrics.py
code served (permissive licence) · get_code("2279ff8f4830ab0f")
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 torch
import numpy as np

torch.manual_seed(0)

# y_true: ground truth labels (1D array-like)
# y_score: predicted scores or probabilities (1D or 2D array-like)

# Case 1: Binary classification with 1D y_score (will be thresholded at 0.5)
y_true = np.array([0, 1, 1, 0, 1])
y_score = np.array([0.2, 0.8, 0.9, 0.3, 0.7])

ARGS = {
    "y_true": y_true,
    "y_score": y_score,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, sklearn, torch — 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
delta_phi Ran
ran on a synthesized input
hqucms/weaver-core/weaver/nn/model/ParticleNeXt.py
code served (permissive licence) · get_code("6a60eb8afd0c968f")
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 float, 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 math

def delta_phi(a, b):
    return (a - b + math.pi) % (2 * math.pi) - math.pi

# The function delta_phi takes two numeric arguments a and b
# and computes an angular difference wrapped to [-pi, pi]
# Both arguments should be numeric values (int, float, or similar)

ARGS = {
    "a": 2.5,
    "b": 1.0
}
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 float, 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
delta_r2 Ran
ran on a synthesized input
hqucms/weaver-core/weaver/nn/model/ParticleNeXt.py
code served (permissive licence) · get_code("a1f3a8f5a58b8a82")
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 float, 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
import math

torch.manual_seed(0)

# The function delta_r2 takes 4 scalar parameters: eta1, phi1, eta2, phi2
# It computes a distance metric in (eta, phi) space
# All parameters are used in arithmetic operations, so they should be numeric values

# Create small numeric values for the parameters
eta1 = 1.5
phi1 = 2.3
eta2 = 0.8
phi2 = 1.2

ARGS = {
    "eta1": eta1,
    "phi1": phi1,
    "eta2": eta2,
    "phi2": phi2
}
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 float, 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
evaluate_metrics Ran
ran on a synthesized input
hqucms/weaver-core/weaver/utils/nn/metrics.py
code served (permissive licence) · get_code("6424b6b029d386f2")
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 dict. 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
from functools import partial
import sklearn.metrics as _m

# Set up y_true and y_score for multi-class classification
np.random.seed(0)

# Create true labels for 4 samples, 3 classes
y_true = np.array([0, 1, 2, 1], dtype=np.int64)

# Create scores for each class (4 samples x 3 classes)
y_score = np.array([
    [0.7, 0.2, 0.1],
    [0.1, 0.8, 0.1],
    [0.1, 0.2, 0.7],
    [0.3, 0.6, 0.1]
], dtype=np.float32)

# Define eval_metrics with various metric types
eval_metrics = ['roc_auc_score', 'confusion_matrix']

ARGS = {
    'y_true': y_true,
    'y_score': y_score,
    'eval_metrics': eval_metrics
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, sklearn — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict 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
flops_to_string Ran
ran; our contract draft was wrong, not the code
hqucms/weaver-core/weaver/utils/flops_counter.py
code served (permissive licence) · get_code("acc45731e5dadf85")
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 torch
import numpy as np

# Based on the function signature and body analysis:
# - flops: int (required) - number of floating point operations
# - units: Optional[str] (optional) - one of 'GMac', 'MMac', 'KMac', or None
# - precision: int (optional) - number of decimal places for rounding

ARGS = {
    "flops": 5_000_000_000,  # 5 billion MACs - will trigger GMac output
    "units": None,            # Let function auto-detect units
    "precision": 2            # Default precision of 2 decimal places
}
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 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
focal_loss Ran
ran on a synthesized input
hqucms/weaver-core/weaver/nn/loss/focal.py
code served (permissive licence) · get_code("22cdb78ad8aed2f6")
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 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
import torch.nn as nn

torch.manual_seed(0)

# Create small input tensor: shape (batch_size=2, num_classes=4)
# This is logits from a model
input_tensor = torch.randn(2, 4, requires_grad=True)

# Create target tensor: shape (batch_size=2,) with class indices
# Values must be in range [0, num_classes)
target_tensor = torch.tensor([0, 2], dtype=torch.long)

# Parameters for focal_loss
alpha = 0.25
gamma = 2.0
reduction = 'mean'
eps = 1e-08

ARGS = {
    'input': input_tensor,
    'target': target_tensor,
    'alpha': alpha,
    'gamma': gamma,
    'reduction': reduction,
    'eps': eps,
}
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 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
get_autocast_config Ran
ran on a synthesized input
hqucms/weaver-core/weaver/utils/nn/tools.py
code served (permissive licence) · get_code("93fd5621a8e99074")
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

class Args:
    def __init__(self):
        self.use_amp = True
        self.amp_dtype = 'bf16'

ARGS = {'args': 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 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
get_graph_ids Ran
ran on a synthesized input
StefReck/MEdgeConv/medgeconv/legacy/util.py
code served (permissive licence) · get_code("3a6889623a4463cb")
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 EagerTensor, 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 tensorflow as tf

# Create a small is_valid tensor
# This should be a 1D tensor of shape (None,) with binary values (0s and 1s)
is_valid = tf.constant([[1], [0], [1], [0]], dtype=tf.int32)

ARGS = {
    "is_valid": is_valid
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type EagerTensor, 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
get_xixj Ran
ran on a synthesized input
StefReck/MEdgeConv/medgeconv/util.py
code served (permissive licence) · get_code("8cdf24d6f598d4e3")
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 4, 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 tensorflow as tf
import numpy as np

# Set seed for reproducibility
tf.random.set_seed(0)

# Create small tensors for testing
batch_size = 2
n_features = 4
k = 3

# nodes_disjoint: shape (batch_size, n_features)
nodes_disjoint = tf.constant(
    np.random.randn(batch_size, n_features).astype(np.float32)
)

# knn: shape (batch_size, k) with int32 indices
# These are indices into the nodes_disjoint tensor
knn = tf.constant(
    np.array([[1, 0, 1], [0, 1, 0]], dtype=np.int32)
)

# k: number of nearest neighbors
k_val = 3

ARGS = {
    "nodes_disjoint": nodes_disjoint,
    "knn": knn,
    "k": k_val,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 4, 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
import_module Ran
ran on a synthesized input
hqucms/weaver-core/weaver/utils/import_tools.py
code served (permissive licence) · get_code("5fd748f3941e18cd")
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 module. 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 tempfile
import os

# Create a temporary Python module file
temp_dir = tempfile.mkdtemp()
module_file = os.path.join(temp_dir, "test_module.py")

# Write a simple valid Python module to the file
with open(module_file, 'w') as f:
    f.write("""
def hello():
    return "Hello from imported module"

class TestClass:
    def __init__(self):
        self.value = 42
""")

ARGS = {
    'path': module_file,
    'name': '_mod'
}
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 module — a module output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a module 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
kernel_nn Ran
ran on a synthesized input
StefReck/MEdgeConv/medgeconv/layers.py
code served (permissive licence) · get_code("7dae7c334c065a24")
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 KerasTensor. 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 tensorflow.keras as ks
import tensorflow as tf
import numpy as np

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

# Create a small input tensor (batch_size=2, features=4)
input_tensor = tf.keras.Input(shape=(4,))

ARGS = {
    'x': input_tensor,
    'units': [8, 6],
    'activation': 'relu',
    'kernel_initializer': 'glorot_uniform'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type KerasTensor — a KerasTensor 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
knn Ran
ran on a synthesized input
hqucms/weaver-core/weaver/nn/model/ParticleNet.py
code served (permissive licence) · get_code("c3cde8788ee9ab7e")
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 int, shape (2, 3, 2) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# Create a small input tensor
# x should be shape (batch, features, num_points) based on how it's used in pairwise_distance
# transpose(2, 1) suggests 3D tensor with dims (batch, features, points)
x = torch.randn(2, 4, 3)

# k parameter - number of nearest neighbors to find
k = 2

# distance_fn parameter - function to compute pairwise distances
def custom_distance_fn(x):
    inner = -2 * torch.matmul(x.transpose(2, 1), x)
    xx = torch.sum(x ** 2, dim=1, keepdim=True)
    return (xx.transpose(2, 1) + inner + xx).clamp(min=0)

distance_fn = custom_distance_fn

# exclude_self parameter - whether to exclude self in knn search
exclude_self = True

ARGS = {
    "x": x,
    "k": k,
    "distance_fn": distance_fn,
    "exclude_self": exclude_self
}
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 3, dtype family int, shape (2, 3, 2) — shape (2, 3, 2) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
pad_array Ran
ran, honoured the contract we drafted
Jai2500/particlenet/particlenet.py
pointer only (licence: NONE) · get_code("8d12109420ecf2ca")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, Jai2500/particlenet 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.

pairwise_distance Ran
ran on a synthesized input
hqucms/weaver-core/weaver/nn/model/ParticleNet.py
code served (permissive licence) · get_code("399f9925e569e331")
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, 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 torch

torch.manual_seed(0)

# The function expects x to be a tensor with shape (batch, features, points)
# Based on the operations:
# - x.transpose(2, 1): swaps dimensions 1 and 2
# - torch.matmul(x.transpose(2, 1), x): matrix multiplication
# - torch.sum(x ** 2, dim=1, keepdim=True): sum over dimension 1
# So x should have 3 dimensions: (batch_size, num_features, num_points)

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

ARGS = {
    "x": x
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 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
params_to_string Ran
ran; our contract draft was wrong, not the code
hqucms/weaver-core/weaver/utils/flops_counter.py
code served (permissive licence) · get_code("44f6804749deb60e")
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 torch
import numpy as np

ARGS = {
    'params_num': 1500000,
    'units': None,
    'precision': 2,
}
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 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
reduce_mean_valid_disjoint Ran
ran on a synthesized input
StefReck/MEdgeConv/medgeconv/legacy/util.py
code served (permissive licence) · get_code("972f6f317f2e5fe5")
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 EagerTensor, 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 tensorflow as tf
import numpy as np

# Create small tensors for testing
# nodes: shape (None, n_features) - flattened nodes from multiple graphs
# We'll create a tensor representing nodes from 2 graphs with 4 features each
# Graph 0 has 3 valid nodes, Graph 1 has 2 valid nodes
nodes_data = np.array([
    [1.0, 2.0, 3.0, 4.0],      # node 0, graph 0
    [5.0, 6.0, 7.0, 8.0],      # node 1, graph 0
    [9.0, 10.0, 11.0, 12.0],   # node 2, graph 0
    [13.0, 14.0, 15.0, 16.0],  # node 3, graph 1
    [17.0, 18.0, 19.0, 20.0],  # node 4, graph 1
], dtype=np.float32)
nodes = tf.constant(nodes_data)

# is_valid: shape (bs, n_points) int32 tensor
# bs=2 (batch size / number of graphs)
# n_points=3 (max nodes per graph, padded)
# Values are 1 for valid nodes, 0 for padding
is_valid_data = np.array([
    [1, 1, 1],  # graph 0: 3 valid nodes
    [1, 1, 0],  # graph 1: 2 valid nodes (last one is padding)
], dtype=np.int32)
is_valid = tf.constant(is_valid_data)

ARGS = {
    "nodes": nodes,
    "is_valid": is_valid,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type EagerTensor, 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
roc_auc_score_ovo Ran
ran on a synthesized input
hqucms/weaver-core/weaver/utils/nn/metrics.py
code served (permissive licence) · get_code("89fe702ab2009e77")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned ndarray, ndim 2, dtype family float, shape (3, 3) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np
import sklearn.metrics as _m

def roc_auc_score_ovo(y_true, y_score):
    if y_score.ndim == 1:
        return _m.roc_auc_score(y_true, y_score)
    else:
        num_classes = y_score.shape[1]
        result = np.zeros((num_classes, num_classes), dtype='float32')
        for i in range(num_classes):
            for j in range(i + 1, num_classes):
                weights = np.logical_or(y_true == i, y_true == j)
                truth = y_true == j
                score = y_score[:, j] / np.maximum(y_score[:, i] + y_score[:, j], 1e-06)
                result[i, j] = _m.roc_auc_score(truth, score, sample_weight=weights)
    return result


# Setup code
np.random.seed(0)

# Create y_true: multiclass labels with 3 classes
y_true = np.array([0, 1, 2, 0, 1, 2, 0, 1])

# Create y_score: probability scores for 3 classes, 8 samples
# Using softmax-like normalized scores
y_score = np.array([
    [0.7, 0.2, 0.1],
    [0.1, 0.8, 0.1],
    [0.1, 0.2, 0.7],
    [0.6, 0.3, 0.1],
    [0.2, 0.7, 0.1],
    [0.1, 0.1, 0.8],
    [0.5, 0.4, 0.1],
    [0.2, 0.6, 0.2]
], dtype='float32')

ARGS = {
    'y_true': y_true,
    'y_score': y_score
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, sklearn — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 2, dtype family float, shape (3, 3) — shape (3, 3) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
to_pt2 Ran
ran; our contract draft was wrong, not the code
hqucms/weaver-core/weaver/nn/model/ParticleNeXt.py
code served (permissive licence) · get_code("779b98ee0e7c9f1d")
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, 1) 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, features=4)
# The function accesses x[:, :2], so we need at least 2 columns
x = torch.randn(2, 4)

ARGS = {
    'x': x,
    'eps': 1e-08
}
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, 1) — shape (2, 1) 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
to_pt2 Ran
ran on a synthesized input
hqucms/weaver-core/weaver/nn/model/ParticleTransformer.py
code served (permissive licence) · get_code("9d3613b4ffa9d368")
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, 1) 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 batch size 2 and at least 2 dimensions
# The function accesses x[:, :2] and calls .size(1), so we need at least 2 columns
x = torch.randn(2, 4)

# eps parameter - use the default value
eps = 1e-08

ARGS = {
    "x": x,
    "eps": eps,
}
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, 1) — shape (2, 1) 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_model_complexity_info Not yet run
unverified (harvested, no recorded run)
hqucms/weaver-core/weaver/utils/flops_counter.py
code served (permissive licence) · get_code("abf919ccaf86558d")
How it ran

no invocation recorded.

one_hot Not yet run
unverified (harvested, no recorded run)
hqucms/weaver-core/weaver/nn/loss/focal.py
code served (permissive licence) · get_code("5daf197a89a09cce")
How it ran

no invocation recorded.

pairwise_distance_etaphi Not yet run
unverified (harvested, no recorded run)
hqucms/weaver-core/weaver/nn/model/ParticleNet.py
code served (permissive licence) · get_code("cbfecd3ae2772e1f")
How it ran

no invocation recorded.

unwrap_model Not yet run
unverified (harvested, no recorded run)
hqucms/weaver-core/weaver/utils/nn/tools.py
code served (permissive licence) · get_code("464cf959f947b7c4")
How it ran

no invocation recorded.

Also present in other papers' code

Of 23 samples lifted from this paper's linked repositories, 23 carry a function identity and 0 do not. Of the 23 distinct functions with an identity, 3 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 3 later) and 20 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
flops_to_string 1 other paper
1 later
2210.05144 (2022-10)
Mixture of Attention Heads: Selecting Attention Heads Per Token
params_to_string 1 other paper
1 later
2210.05144 (2022-10)
Mixture of Attention Heads: Selecting Attention Heads Per Token
to_pt2 1 other paper
1 later
2407.08682 (2024-07)
Jet Tagging with More-Interaction Particle Transformer

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

How to represent a jet is at the core of machine learning on jet physics. Inspired by the notion of point clouds, we propose a new approach that considers a jet as an unordered set of its constituent particles, effectively a "particle cloud". Such a particle cloud representation of jets is efficient in incorporating raw information of jets and also explicitly respects the permutation symmetry. Based on the particle cloud representation, we propose ParticleNet, a customized neural network architecture using Dynamic Graph Convolutional Neural Network for jet tagging problems. The ParticleNet architecture achieves state-of-the-art performance on two representative jet tagging benchmarks and is improved significantly over existing methods.

For agents

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

get_harvested_code_for_paper("1902.08570")
get_code_for_paper("1902.08570")
have("1902.08570")

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

curl https://syntology.ai/api/ran/1902.08570.json

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

[![Syntology run record](https://syntology.ai/api/ran/1902.08570.svg)](https://syntology.ai/paper/1902.08570)

Connect an agent — have() is free.