SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2003.00982 · 2020

Benchmarking Graph Neural Networks

Yoshua Bengio, Anh Luu, Thomas Laurent, Xavier Bresson, Chaitanya Joshi, Vijay Prakash Dwivedi, Chaitanya Dwivedi, Vijay ©2022, Anh Joshi, Thomas Luu, Yoshua Laurent, Xavier Bengio, and 1 more

arXiv · PDF · Open in the Atlas

Code that ran

10 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; 0 honoured the contract we drafted, 1 violated it and 8 have no contract verdict; 1 failure of our instrument; 13 no recorded run), measured 2026-11-21; 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/2003.00982.json · badge: /api/ran/2003.00982.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
graphdeeplearning/benchmarking-gnns canonical (classifier label)
classifier basis not recorded on the link
3 ran · 0 no recorded run
liketheflower/jimmy_dgl community (classifier label)
classifier basis not recorded on the link
5 ran · 13 no recorded run
yashkotadia/GatedGCN-PATTERN community (classifier label)
classifier basis not recorded on the link
1 ran · 0 no recorded run
chenfengzhao/gnnhls community (classifier label)
classifier basis not recorded on the link
1 ran · 0 no recorded run
FunctionStatusWhere it lives
accuracy_SBM Ran
ran on a synthesized input
yashkotadia/GatedGCN-PATTERN/train/metrics.py
code served (permissive licence) · get_code("4c77cfa0f2d2f37d")
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 float64, ndim 0, 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 numpy as np

torch.manual_seed(0)

# Create small tensors for testing
batch_size = 2
num_classes = 4

# scores: logits tensor that will be passed through softmax
# Shape: (batch_size, num_classes)
scores = torch.randn(batch_size, num_classes)

# targets: ground truth class indices
# Shape: (batch_size,)
targets = torch.tensor([0, 2], dtype=torch.long)

ARGS = {
    "scores": scores,
    "targets": targets,
}
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 float64, ndim 0, dtype family float — no output shape is stated: observe_signature.py recorded float ndim 0, which disagrees with the fixture run
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
diag_offdiag_maxpool Ran
ran on a synthesized input
graphdeeplearning/benchmarking-gnns/layers/three_wl_gnn_layers.py
code served (permissive licence) · get_code("15bb4b29cca9b4a1")
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, 6) 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 a tensor input with shape that allows:
# - input.shape[-1] to get N (last dimension size)
# - torch.diagonal(input, dim1=-2, dim2=-1) to work (needs at least 2D for the last two dims)
# - Indexing with dim=2 after diagonal, so needs at least 3 dimensions
# - Operations that preserve batch dimensions

# Looking at the function:
# - diagonal extracts diagonal across last 2 dims, result has shape [..., N]
# - max(..., dim=2)[0] suggests the result before max has at least 3 dims
# - The output concatenates on dim=1, suggesting (batch, channels, ...)
# - min_mat is reshaped to (1, 1, N, N), then subtracted from input
# - max operations on dim=3 and dim=2 suggest input is 4D: (batch, channels, N, N)

# So input should be: (batch_size, num_channels, N, N)
N = 4
batch_size = 2
num_channels = 3

input_tensor = torch.randn(batch_size, num_channels, N, N)

ARGS = {
    "input": input_tensor
}
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, 6) — shape (2, 6) 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
find_src_dst_ntypes Ran
ran on a synthesized input
liketheflower/jimmy_dgl/python/dgl/heterograph.py
code served (permissive licence) · get_code("61782edfe2cb4a92")
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 numpy as np

class MockGraphIndex:
    """Mock GraphIndex that supports the edges() method."""
    def __init__(self, src_ids, dst_ids):
        self.src_ids = src_ids
        self.dst_ids = dst_ids
    
    def edges(self):
        """Returns (src, dst, edge_ids) where src and dst have tonumpy() method."""
        src_tensor = MockTensor(self.src_ids)
        dst_tensor = MockTensor(self.dst_ids)
        edge_ids = MockTensor(np.arange(len(self.src_ids)))
        return (src_tensor, dst_tensor, edge_ids)

class MockTensor:
    """Mock tensor-like object with tonumpy() method."""
    def __init__(self, data):
        self.data = np.array(data)
    
    def tonumpy(self):
        return self.data
    
    def __iter__(self):
        return iter(self.data)

torch.manual_seed(0)

# Create a uni-bipartite metagraph
# SRC node types: 0, 1
# DST node types: 2, 3
# Node type 4 is isolated
src_node_type_ids = np.array([0, 0, 1, 1])
dst_node_type_ids = np.array([2, 3, 2, 3])

metagraph = MockGraphIndex(src_node_type_ids, dst_node_type_ids)

# Define node types
ntypes = ["user", "item", "product", "category", "isolated"]

ARGS = {
    "ntypes": ntypes,
    "metagraph": metagraph,
}
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 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
from_csr Ran
ran on a synthesized input
liketheflower/jimmy_dgl/python/dgl/graph_index.py
code served (permissive licence) · get_code("61a050f11d9e1f7e")
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 GraphIndex. 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

torch.manual_seed(0)

# Mock the utils module and _CAPI_DGLGraphCSRCreate since they're from DGL
class MockIndex:
    def __init__(self, data):
        self.data = data
    
    def todgltensor(self):
        # Return a mock tensor representation
        return self.data

class utils:
    @staticmethod
    def toindex(x):
        if isinstance(x, torch.Tensor):
            return MockIndex(x)
        return MockIndex(torch.tensor(x))

# Mock the C API function
def _CAPI_DGLGraphCSRCreate(indptr_tensor, indices_tensor, direction):
    # Return a mock GraphIndex
    class GraphIndex:
        def __init__(self, indptr, indices, direction):
            self.indptr = indptr
            self.indices = indices
            self.direction = direction
    return GraphIndex(indptr_tensor, indices_tensor, direction)

# Inject mocks into globals so from_csr can access them
import sys
current_module = sys.modules[__name__]
setattr(current_module, 'utils', utils)
setattr(current_module, '_CAPI_DGLGraphCSRCreate', _CAPI_DGLGraphCSRCreate)

# Create valid CSR format tensors
# For a graph with 3 nodes and some edges:
# indptr: [0, 2, 3, 4] - points to where each node's edges start in indices
# indices: [1, 2, 0, 1] - column indices of edges

indptr_tensor = torch.tensor([0, 2, 3, 4], dtype=torch.int64)
indices_tensor = torch.tensor([1, 2, 0, 1], dtype=torch.int64)
direction_str = "out"

ARGS = {
    "indptr": indptr_tensor,
    "indices": indices_tensor,
    "direction": direction_str
}
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 GraphIndex — a GraphIndex output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a GraphIndex 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
gpu_setup Ran
ran; our contract draft was wrong, not the code
graphdeeplearning/benchmarking-gnns/main_COLLAB_edge_classification.py
pointer only (licence: MIT); the stored copy get_code serves, from trustworthygnn/mia-gnn, is pointer only (licence: NONE) · get_code("f72434c0dcd5d11b")
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 trustworthygnn/mia-gnn, 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.

is_all Ran
ran, violated the contract we drafted
liketheflower/jimmy_dgl/python/dgl/base.py
code served (permissive licence) · get_code("48c452849a73cec5")
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
import torch.nn as nn
import numpy as np

torch.manual_seed(0)

ALL = '__ALL__'

def is_all(arg):
    """Return true if the argument is a special symbol for all nodes or edges."""
    return isinstance(arg, str) and arg == ALL

ARGS = {
    'arg': ALL
}
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 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
is_unibipartite Ran
ran on a synthesized input
liketheflower/jimmy_dgl/python/dgl/heterograph.py
code served (permissive licence) · get_code("0fccc3114c12388a")
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
import numpy as np

# Create a minimal GraphIndex stand-in class that implements the edges() method
class GraphIndex:
    def __init__(self, src, dst):
        self.src = src
        self.dst = dst
    
    def edges(self):
        """Returns (src, dst, edge_ids) tuple"""
        # Create a simple tensor wrapper that has tonumpy() method
        return (TensorWrapper(self.src), TensorWrapper(self.dst), None)

class TensorWrapper:
    """Minimal wrapper that provides tonumpy() method"""
    def __init__(self, data):
        self.data = data
    
    def tonumpy(self):
        if isinstance(self.data, np.ndarray):
            return self.data
        return np.array(self.data)

# Create a uni-bipartite graph: src and dst have no overlap
# src nodes: {0, 1}
# dst nodes: {2, 3}
src_nodes = np.array([0, 1], dtype=np.int64)
dst_nodes = np.array([2, 3], dtype=np.int64)

graph = GraphIndex(src_nodes, dst_nodes)

ARGS = {
    "graph": graph
}
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 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
make_canonical_etypes Ran
ran on a synthesized input
liketheflower/jimmy_dgl/python/dgl/heterograph.py
code served (permissive licence) · get_code("3d5e4058d7d4c01b")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned list, ndim 2, 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 numpy as np

class GraphIndex:
    """Stand-in class for metagraph GraphIndex"""
    def __init__(self, num_nodes, num_edges, src, dst):
        self._num_nodes = num_nodes
        self._num_edges = num_edges
        self._src = src
        self._dst = dst
    
    def number_of_nodes(self):
        return self._num_nodes
    
    def number_of_edges(self):
        return self._num_edges
    
    def edges(self):
        return self._src, self._dst, np.arange(len(self._src))

# Create a simple metagraph with 3 node types and 2 edge types
# Edge 0: node 0 -> node 1
# Edge 1: node 1 -> node 2
metagraph = GraphIndex(
    num_nodes=3,
    num_edges=2,
    src=np.array([0, 1]),
    dst=np.array([1, 2])
)

etypes = ["follows", "likes"]
ntypes = ["user", "post", "comment"]

ARGS = {
    "etypes": etypes,
    "ntypes": ntypes,
    "metagraph": metagraph
}
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 list, ndim 2, dtype family float — a list output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
ops_2_to_2 Ran
ran on a synthesized input
graphdeeplearning/benchmarking-gnns/layers/ring_gnn_equiv_layer.py
code served (permissive licence) · get_code("8a8a1e482dac9901")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned list, ndim 5, 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

torch.manual_seed(0)

# The function expects:
# - inputs: a 4D tensor (batch_size, ..., dim, dim) where last two dims are square matrices
# - dim: an integer representing the dimension of the square matrices
# - normalization: a string or None
# - normalization_val: a float (not used in the function body)

# Create a small 4D tensor: batch_size=2, intermediate=3, dim=4, dim=4
inputs = torch.randn(2, 3, 4, 4)

# dim parameter - used for repeat operations and must match last two dimensions of inputs
dim = 4

# normalization parameter
normalization = 'inf'

# normalization_val parameter
normalization_val = 1.0

ARGS = {
    'inputs': inputs,
    'dim': dim,
    'normalization': normalization,
    'normalization_val': normalization_val,
}
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 list, ndim 5, dtype family float — a list output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
ops_2_to_2 Ran
ran on a synthesized input
chenfengzhao/gnnhls/benchmarking-gnns/layers/ring_gnn_equiv_layer.py
code served (permissive licence) · get_code("98f002b1572282d2")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned list, ndim 5, 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

torch.manual_seed(0)

# Create a small 4D tensor with shape (batch=2, channels=3, height=4, width=4)
# The function expects inputs to have diagonal operations on dims 2 and 3
inputs = torch.randn(2, 3, 4, 4)

# dim parameter - used for repeat operations and represents the spatial dimension
dim = 4

# normalization parameter - can be 'inf' or None
normalization = 'inf'

# normalization_val parameter
normalization_val = 1.0

ARGS = {
    'inputs': inputs,
    'dim': dim,
    'normalization': normalization,
    'normalization_val': normalization_val
}
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 list, ndim 5, dtype family float — a list output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
bipartite Not yet run
unverified (harvested, no recorded run)
liketheflower/jimmy_dgl/python/dgl/convert.py
code served (permissive licence) · get_code("08ec4232dc9afa5b")
How it ran

no invocation recorded.

create_heterograph_from_relations Not yet run
unverified (harvested, no recorded run)
liketheflower/jimmy_dgl/python/dgl/heterograph_index.py
code served (permissive licence) · get_code("0e0d8b04e0a30007")
How it ran

no invocation recorded.

create_unitgraph_from_coo Not yet run
unverified (harvested, no recorded run)
liketheflower/jimmy_dgl/python/dgl/heterograph_index.py
code served (permissive licence) · get_code("6926e1ccdf433cce")
How it ran

no invocation recorded.

create_unitgraph_from_csr Not yet run
unverified (harvested, no recorded run)
liketheflower/jimmy_dgl/python/dgl/heterograph_index.py
code served (permissive licence) · get_code("da2d1d33aa78c510")
How it ran

no invocation recorded.

frame_like Not yet run
unverified (harvested, no recorded run)
liketheflower/jimmy_dgl/python/dgl/frame.py
code served (permissive licence) · get_code("9ab8241d3faddf35")
How it ran

no invocation recorded.

from_coo Not yet run
unverified (harvested, no recorded run)
liketheflower/jimmy_dgl/python/dgl/graph_index.py
code served (permissive licence) · get_code("4ae9983fc353b1f9")
How it ran

no invocation recorded.

from_shared_mem_graph_index Not yet run
unverified (harvested, no recorded run)
liketheflower/jimmy_dgl/python/dgl/graph_index.py
code served (permissive licence) · get_code("e8e462381451afcd")
How it ran

no invocation recorded.

graph Not yet run
unverified (harvested, no recorded run)
liketheflower/jimmy_dgl/python/dgl/convert.py
code served (permissive licence) · get_code("5759d6efe4e482a4")
How it ran

no invocation recorded.

hetero_from_relations Not yet run
unverified (harvested, no recorded run)
liketheflower/jimmy_dgl/python/dgl/convert.py
code served (permissive licence) · get_code("fa65de5b1c005eee")
How it ran

no invocation recorded.

infer_scheme Not yet run
unverified (harvested, no recorded run)
liketheflower/jimmy_dgl/python/dgl/frame.py
code served (permissive licence) · get_code("41db54384fdd01d6")
How it ran

no invocation recorded.

mutation Not yet run
unverified (harvested, no recorded run)
liketheflower/jimmy_dgl/python/dgl/graph.py
code served (permissive licence) · get_code("63102c67281002a7")
How it ran

no invocation recorded.

rand_graph Not yet run
unverified (harvested, no recorded run)
liketheflower/jimmy_dgl/python/dgl/generators.py
code served (permissive licence) · get_code("a8fbb0de9230475b")
How it ran

no invocation recorded.

zero_initializer Not yet run
unverified (harvested, no recorded run)
liketheflower/jimmy_dgl/python/dgl/init.py
code served (permissive licence) · get_code("c7a98eaf8183d806")
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, 13 are also present in other papers whose code we hold (11 with at least one paper dated earlier, 0 the same month, 4 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
gpu_setup 8 other papers
8 later
2009.11746 (2020-09)
Learning Graph Normalization for Graph Neural Networks
2010.02863 (2020-10)
Directional Graph Networks
2012.10258 (2020-12)
An Experimental Study of the Transferability of Spectral Graph Networks
and 5 more not named here
diag_offdiag_maxpool 2 other papers
1 dated earlier, 1 later
1905.11136 (2019-05)
Provably Powerful Graph Networks
2106.06707 (2021-06)
Graph Neural Networks with Local Graph Parameters
is_all 2 other papers
1 dated earlier, 1 later
1909.01315 (2019-09)
Deep Graph Library: A Graph-Centric, Highly-Performant Package for Graph Neural Networks
2009.03509 (2020-09)
Masked Label Prediction: Unified Message Passing Model for Semi-Supervised Classification
create_heterograph_from_relations 1 other paper
1 dated earlier
1909.01315 (2019-09)
Deep Graph Library: A Graph-Centric, Highly-Performant Package for Graph Neural Networks
frame_like 1 other paper
1 dated earlier
1909.01315 (2019-09)
Deep Graph Library: A Graph-Centric, Highly-Performant Package for Graph Neural Networks
from_coo 1 other paper
1 dated earlier
1909.01315 (2019-09)
Deep Graph Library: A Graph-Centric, Highly-Performant Package for Graph Neural Networks
from_csr 1 other paper
1 dated earlier
1909.01315 (2019-09)
Deep Graph Library: A Graph-Centric, Highly-Performant Package for Graph Neural Networks
from_shared_mem_graph_index 1 other paper
1 dated earlier
1909.01315 (2019-09)
Deep Graph Library: A Graph-Centric, Highly-Performant Package for Graph Neural Networks
infer_scheme 1 other paper
1 dated earlier
1909.01315 (2019-09)
Deep Graph Library: A Graph-Centric, Highly-Performant Package for Graph Neural Networks
is_unibipartite 1 other paper
1 dated earlier
1909.01315 (2019-09)
Deep Graph Library: A Graph-Centric, Highly-Performant Package for Graph Neural Networks
make_canonical_etypes 1 other paper
1 dated earlier
1909.01315 (2019-09)
Deep Graph Library: A Graph-Centric, Highly-Performant Package for Graph Neural Networks
ops_2_to_2 1 other paper
1 later
2106.06707 (2021-06)
Graph Neural Networks with Local Graph Parameters
zero_initializer 1 other paper
1 dated earlier
1909.01315 (2019-09)
Deep Graph Library: A Graph-Centric, Highly-Performant Package for Graph Neural Networks

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 2026-11-21.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

In the last few years, graph neural networks (GNNs) have become the standard toolkit for analyzing and learning from data on graphs. This emerging field has witnessed an extensive growth of promising techniques that have been applied with success to computer science, mathematics, biology, physics and chemistry. But for any successful field to become mainstream and reliable, benchmarks must be developed to quantify progress. This led us in March 2020 to release a benchmark framework that i) comprises of a diverse collection of mathematical and real-world graphs, ii) enables fair model comparison with the same parameter budget to identify key architectures, iii) has an open-source, easy-to-use and reproducible code infrastructure, and iv) is flexible for researchers to experiment with new theoretical ideas. As of December 2022, the GitHub repository 1 has reached 2,000 stars and 380 forks, which demonstrates the utility of the proposed open-source framework through the wide usage by the GNN community. In this paper, we present an updated version of our benchmark with a concise presentation of the aforementioned framework characteristics, an additional medium-sized molecular dataset AQSOL, similar to the popular ZINC, but with a real-world measured chemical target, and discuss how this framework can be leveraged to explore new GNN designs and insights. As a proof of value of our benchmark, we study the case of graph positional encoding (PE) in GNNs, which was introduced with this benchmark and has since spurred interest of exploring more powerful PE for Transformers and GNNs in a robust experimental setting.

For agents

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

get_harvested_code_for_paper("2003.00982")
get_code_for_paper("2003.00982")
have("2003.00982")

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

curl https://syntology.ai/api/ran/2003.00982.json

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

[![Syntology run record](https://syntology.ai/api/ran/2003.00982.svg)](https://syntology.ai/paper/2003.00982)

Connect an agent — have() is free.