SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2406.16715 · NeurIPS · 2025

GC4NC: A Benchmark Framework for Graph Condensation on Node Classification with New Insights

Carl Yang, Wei Jin, Noveen Sachdeva, Shengbo Gong, Juntong Ni

arXiv · PDF · Open in the Atlas

Code that ran

10 of 10 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; 3 honoured the contract we drafted, 2 violated it and 0 have no contract verdict; 5 failures of our instrument; 0 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/2406.16715.json · badge: /api/ran/2406.16715.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
emory-melody/graphslim canonical (classifier label)
classifier basis: census
7 ran · 0 no recorded run
chandlerbang/gcond canonical (classifier label)
classifier basis: census
3 ran · 0 no recorded run
FunctionStatusWhere it lives
dict2obj Ran
ran; our contract draft was wrong, not the code
emory-melody/graphslim/graphslim/config.py
code served (permissive licence) · get_code("57b72777b363affe")
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 Obj. 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 json
from pprint import pformat

class Obj(object):

    def __init__(self, dict_):
        self.__dict__.update(dict_)

    def __repr__(self):
        return pformat(self.__dict__, compact=True)

def dict2obj(d):
    return json.loads(json.dumps(d), object_hook=Obj)

# Construct valid arguments for dict2obj
# The function takes a dictionary and converts it to an object with nested objects

test_dict = {
    "name": "test",
    "value": 42,
    "nested": {
        "key1": "value1",
        "key2": [1, 2, 3],
        "deeper": {
            "x": 10,
            "y": 20
        }
    },
    "list_of_dicts": [
        {"a": 1},
        {"b": 2}
    ]
}

ARGS = {
    "d": test_dict
}
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 Obj — a Obj output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Obj 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
format_args_tables Ran
ran; our contract draft was wrong, not the code
emory-melody/graphslim/graphslim/config.py
code served (permissive licence) · get_code("e7915897d68f4cda")
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 logging
from argparse import Namespace

torch_imported = False
try:
    import torch
    torch.manual_seed(0)
    torch_imported = True
except ImportError:
    pass

# Create a simple args object with various parameter types
args = Namespace(
    dataset='dataset_name',
    method='random',
    gpu_id=0,
    device='cpu',
    setting='transductive',
    split='full',
    run_reduction=True,
    run_eval=True,
    run_inter_eval=False,
    eval_interval=10,
    hidden=64,
    condense_model=None,
    epochs=100,
    multi_label=False,
    dis_metric='euclidean',
    lr_adj=0.01,
    lr_feat=0.01,
    optim='adam',
    threshold=0.5,
    dropout=0.5,
    ntrans=2,
    with_bn=True,
    save_path='/tmp/model.pt',
    load_path=None,
    eval_whole=True,
    with_structure=True,
    lr=0.001,
    weight_decay=1e-5,
    pre_norm=False,
    outer_loop=1,
    inner_loop=5,
    reduction_rate=0.5,
    seed=42,
    nlayers=3,
    verbose=1,
    soft_label=False,
    init='xavier',
    checkpoints=[],
    logger=logging.getLogger('test_logger'),
    metric='accuracy',
    wandb=False,
    wandb_project='project',
    wandb_run_name='run1',
    wandb_required=False,
    agg='mean',
    eval_epochs=10,
    eval_model='mlp',
    final_eval_model='mlp',
    eval_wd=0.0,
    eval_loss='cross_entropy',
    activation='relu',
    alpha=0.5,
    ptb_r=0.1,
    attack='none',
    coarsen_strategy='heavy_edge',
    batch_adj=False,
    mx_size=1000,
    ts=2,
    feat_alpha=0.1,
    smoothness_alpha=0.1,
    eigen_k=10,
    ratio=0.5,
    lr_eigenvec=0.01,
    gamma=0.9,
    test_epochs=5,
)

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 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
get_loops Ran
ran, honoured the contract we drafted
chandlerbang/gcond/gcond_agent_induct.py
pointer only (licence: NONE) · get_code("e74da5f66093b26e")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, chandlerbang/gcond 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_loops Ran
ran, honoured the contract we drafted
chandlerbang/gcond/gcond_agent_transduct.py
pointer only (licence: NONE) · get_code("86deaed779072495")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, chandlerbang/gcond 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_dgl_graph Ran
ran, violated the contract we drafted
emory-melody/graphslim/graphslim/compat.py
code served (permissive licence) · get_code("9539d9d5d4de1c0e")
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

# Create a mock DGL graph object
class MockDGLGraph:
    pass

# Set the module path to make it appear as a DGL graph
MockDGLGraph.__module__ = 'dgl.graph'

# Create an instance of the mock DGL graph
dgl_graph = MockDGLGraph()

ARGS = {
    'graph': dgl_graph
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type bool, ndim 0, dtype family float, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
is_pyg_data Ran
ran, violated the contract we drafted
emory-melody/graphslim/graphslim/compat.py
code served (permissive licence) · get_code("c602aee068ee9339")
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
from torch_geometric.data import Data

torch.manual_seed(0)

# Create a valid PyG Data object
node_features = torch.randn(5, 4)
edge_index = torch.tensor([[0, 1, 2, 3], [1, 2, 3, 4]], dtype=torch.long)
pyg_graph = Data(x=node_features, edge_index=edge_index)

ARGS = {
    "graph": pyg_graph
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch, torch_geometric — 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
load_config Ran
ran; our contract draft was wrong, not the code
chandlerbang/gcond/configs.py
pointer only (licence: NONE) · get_code("a2d97b7ae1e59b60")
How it ran

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

mx_inv_sqrt Ran
ran, honoured the contract we drafted
emory-melody/graphslim/graphslim/models/ignr.py
pointer only (licence: MIT); the stored copy get_code serves, from twistedcubic/copt, is pointer only (licence: NONE) · get_code("43cda63f18aa6233")
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 twistedcubic/copt, 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.

mx_tr Ran
ran; our contract draft was wrong, not the code
emory-melody/graphslim/graphslim/models/ignr.py
code served (permissive licence) · get_code("c0c3857f67d2ea9e")
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

torch.manual_seed(0)

# mx_tr expects a matrix-like object with .diag() method and .sum() method
# This is a torch tensor
mx = torch.randn(4, 4)

ARGS = {
    "mx": mx
}
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
to_pyg_data Ran
ran; our contract draft was wrong, not the code
emory-melody/graphslim/graphslim/compat.py
code served (permissive licence) · get_code("3ad62857b188d6b0")
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 Data. 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
from torch_geometric.data import Data

torch.manual_seed(0)

# Create a minimal PyG-like object with the attributes the function checks for
class PyGLikeGraph:
    def __init__(self):
        self.edge_index = torch.tensor([[0, 1, 1, 2], [1, 0, 2, 1]], dtype=torch.long)
        self.x = torch.randn(3, 4)
        self.y = torch.tensor([0, 1, 0])

graph = PyGLikeGraph()

ARGS = {
    'graph': graph,
    'node_feature': 'feat',
    'node_label': 'label'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch, torch_geometric — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Data — a Data 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

Also present in other papers' code

Of 10 samples lifted from this paper's linked repositories, 10 carry a function identity and 0 do not. Of the 10 distinct functions with an identity, 4 are also present in other papers whose code we hold (4 with at least one paper dated earlier, 0 the same month, 1 later) and 6 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
load_config 2 other papers
1 dated earlier, 1 later
2310.09192 (2023-10)
Does Graph Distillation See Like Vision Dataset Counterpart?
2407.00615 (2024-07)
GC-Bench: An Open and Unified Benchmark for Graph Condensation
mx_inv_sqrt 2 other papers
2 dated earlier
2003.03892 (2020-03)
COPT: Coordinated Optimal Transport for Graph Sketching
2310.09192 (2023-10)
Does Graph Distillation See Like Vision Dataset Counterpart?
get_loops 1 other paper
1 dated earlier
2402.04924 (2024-02)
Two Trades is not Baffled: Condensing Graph via Crafting Rational Gradient Matching
get_loops 1 other paper
1 dated earlier
2310.09192 (2023-10)
Does Graph Distillation See Like Vision Dataset Counterpart?

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

Graph condensation (GC) is an emerging technique designed to learn a significantly smaller graph that retains the essential information of the original graph. This condensed graph has shown promise in accelerating graph neural networks while preserving performance comparable to those achieved with the original, larger graphs. Additionally, this technique facilitates downstream applications like neural architecture search and deepens our understanding of redundancies in large graphs. Despite the rapid development of GC methods, particularly for node classification, a unified evaluation framework is still lacking to systematically compare different GC methods or clarify key design choices for improving their effectiveness. To bridge these gaps, we introduce GC4NC, a comprehensive framework for evaluating diverse GC methods on node classification across multiple dimensions including performance, efficiency, privacy preservation, denoising ability, NAS effectiveness, and transferability. Our systematic evaluation offers novel insights into how condensed graphs behave and the critical design choices that drive their success. These findings pave the way for future advancements in GC methods, enhancing both performance and expanding their real-world applications.

For agents

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

get_harvested_code_for_paper("2406.16715")
get_code_for_paper("2406.16715")
have("2406.16715")

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

curl https://syntology.ai/api/ran/2406.16715.json

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

[![Syntology run record](https://syntology.ai/api/ran/2406.16715.svg)](https://syntology.ai/paper/2406.16715)

Connect an agent — have() is free.