SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2410.04263 · ICML · 2025

DeFoG: Discrete Flow Matching for Graph Generation

Pascal Frossard, Manuel Madeira, Yiming Qin, Dorina Thanou

arXiv · PDF · Open in the Atlas

Code that ran

12 of 15 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, 3 violated it and 1 has no contract verdict; 7 failures of our instrument; 3 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/2410.04263.json · badge: /api/ran/2410.04263.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
manuelmlmadeira/DeFoG canonical (classifier label)
classifier basis: census
12 ran · 3 no recorded run
FunctionStatusWhere it lives
compare_hash Ran
ran, violated the contract we drafted
manuelmlmadeira/DeFoG/src/datasets/guacamol_dataset.py
code served (permissive licence) · get_code("ab10f4ba7e15c9c9")
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 hashlib
import tempfile
import os

# Create a temporary file with known content
temp_dir = tempfile.mkdtemp()
output_file = os.path.join(temp_dir, "test_smiles.txt")

# Write some test content to the file
test_content = b"CCO\nCC(C)O\nc1ccccc1"
with open(output_file, 'wb') as f:
    f.write(test_content)

# Compute the correct hash of this content
correct_hash = hashlib.md5(test_content).hexdigest()

ARGS = {
    "output_file": output_file,
    "correct_hash": correct_hash
}
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 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
degree_worker Ran
ran; our contract draft was wrong, not the code
manuelmlmadeira/DeFoG/src/analysis/spectre_utils.py
code served (permissive licence) · get_code("2830d89cd6960c9d")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned ndarray, ndim 1, dtype family int, shape (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 numpy as np
import networkx as nx

# Create a simple graph for testing
G = nx.Graph()
G.add_edges_from([(0, 1), (1, 2), (2, 3), (3, 0), (1, 3)])

ARGS = {
    "G": G
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
networkx, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 1, dtype family int, shape (4,) — shape (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
extract_smiles_from_csv Ran
ran; our contract draft was wrong, not the code
manuelmlmadeira/DeFoG/src/datasets/moses_dataset.py
code served (permissive licence) · get_code("ddbacf1b3a22d0c3")
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 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 pandas as pd
import tempfile
import os

# Create a temporary CSV file with SMILES data
temp_dir = tempfile.mkdtemp()
csv_path = os.path.join(temp_dir, "test_smiles.csv")

# Write a test CSV with SMILES column
df = pd.DataFrame({
    'SMILES': ['CC(=O)O', 'C1=CC=CC=C1', 'CCO', 'c1ccccc1']
})
df.to_csv(csv_path, index=False)

ARGS = {
    'csv_path': csv_path
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
pandas — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 1, 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
files_exist Ran
ran, violated the contract we drafted
manuelmlmadeira/DeFoG/src/datasets/guacamol_dataset.py
code served (permissive licence) · get_code("934e88324646e2d7")
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 os
import tempfile
import os.path as osp

# Create temporary files for testing
with tempfile.TemporaryDirectory() as tmpdir:
    # Create some test files
    test_file1 = osp.join(tmpdir, "file1.txt")
    test_file2 = osp.join(tmpdir, "file2.txt")
    
    with open(test_file1, 'w') as f:
        f.write("test1")
    with open(test_file2, 'w') as f:
        f.write("test2")
    
    # Define ARGS with a list of existing files
    ARGS = {
        "files": [test_file1, test_file2]
    }
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 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
get_smiles Ran
ran; our contract draft was wrong, not the code
manuelmlmadeira/DeFoG/src/datasets/moses_dataset.py
code served (permissive licence) · get_code("15ea5f6142f381aa")
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 os
import tempfile
import pandas as pd

# Create a temporary directory to hold test files
temp_dir = tempfile.mkdtemp()

# Create test CSV files for the non-filtered case
train_csv = os.path.join(temp_dir, 'train_moses.csv')
val_csv = os.path.join(temp_dir, 'val_moses.csv')
test_csv = os.path.join(temp_dir, 'test_moses.csv')

train_df = pd.DataFrame({'SMILES': ['CCO', 'CC(C)O', 'c1ccccc1']})
val_df = pd.DataFrame({'SMILES': ['CC', 'CCC']})
test_df = pd.DataFrame({'SMILES': ['CCCC']})

train_df.to_csv(train_csv, index=False)
val_df.to_csv(val_csv, index=False)
test_df.to_csv(test_csv, index=False)

# Create test .smiles files for the filtered case
train_smiles_file = os.path.join(temp_dir, 'new_train.smiles')
val_smiles_file = os.path.join(temp_dir, 'new_val.smiles')
test_smiles_file = os.path.join(temp_dir, 'new_test.smiles')

with open(train_smiles_file, 'w') as f:
    f.write('CCO\nCC(C)O\nc1ccccc1\n')

with open(val_smiles_file, 'w') as f:
    f.write('CC\nCCC\n')

with open(test_smiles_file, 'w') as f:
    f.write('CCCC\n')

ARGS = {
    'raw_dir': temp_dir,
    'filter_dataset': False
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
pandas — 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
l2 Ran
ran, honoured the contract we drafted
manuelmlmadeira/DeFoG/src/analysis/dist_helper.py
pointer only (licence: MIT); the stored copy get_code serves, from lf-wen/hgdm, is pointer only (licence: NONE) · get_code("039fc6561a9516bc")
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 lf-wen/hgdm, 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.

load_pickle Ran
ran; our contract draft was wrong, not the code
manuelmlmadeira/DeFoG/src/datasets/dataset_utils.py
code served (permissive licence) · get_code("70db7c9cbe9e680d")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1) and returned 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 pickle
import tempfile
import os

# Create a temporary file with pickled data
temp_dir = tempfile.mkdtemp()
temp_path = os.path.join(temp_dir, "test.pkl")

# Create some sample data to pickle
sample_data = {
    "key1": "value1",
    "key2": [1, 2, 3],
    "key3": {"nested": "dict"}
}

# Write the pickled data to the temporary file
with open(temp_path, 'wb') as f:
    pickle.dump(sample_data, f)

ARGS = {
    "path": temp_path
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
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 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
normalize Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
manuelmlmadeira/DeFoG/src/utils.py
code served (permissive licence) · get_code("e067f48b33fff70a")
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 PlaceHolder. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# Create small test tensors
batch_size = 2
num_nodes = 4
num_node_features = 3
num_edge_features = 5

# X: (batch_size, num_nodes, num_node_features)
X = torch.randn(batch_size, num_nodes, num_node_features)

# E: (batch_size, num_nodes, num_nodes, num_edge_features)
E = torch.randn(batch_size, num_nodes, num_nodes, num_edge_features)

# Make E symmetric (as required by the mask function)
E = (E + E.transpose(1, 2)) / 2

# y: (batch_size, 1) - target values
y = torch.randn(batch_size, 1)

# norm_values: normalization scaling factors for X, E, y
norm_values = [
    torch.ones(num_node_features),
    torch.ones(num_edge_features),
    torch.ones(1)
]

# norm_biases: normalization bias terms for X, E, y
norm_biases = [
    torch.zeros(num_node_features),
    torch.zeros(num_edge_features),
    torch.zeros(1)
]

# node_mask: (batch_size, num_nodes) - binary mask indicating valid nodes
node_mask = torch.ones(batch_size, num_nodes, dtype=torch.bool)
# Mark some nodes as invalid
node_mask[0, 2:] = False
node_mask[1, 3:] = False

ARGS = {
    'X': X,
    'E': E,
    'y': y,
    'norm_values': norm_values,
    'norm_biases': norm_biases,
    'node_mask': node_mask
}
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 PlaceHolder — a PlaceHolder output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a PlaceHolder 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
spectral_worker Ran
ran; our contract draft was wrong, not the code
manuelmlmadeira/DeFoG/src/analysis/spectre_utils.py
code served (permissive licence) · get_code("d06299b76725edd9")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned ndarray, ndim 1, dtype family float, shape (200,) 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 networkx as nx
from scipy.linalg import eigvalsh

# Create a simple graph for testing
G = nx.karate_club_graph()

# Define the arguments dictionary
ARGS = {
    "G": G,
    "n_eigvals": 10
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
networkx, numpy, scipy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 1, dtype family float, shape (200,) — shape (200,) 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
symmetrize_and_mask_diag Ran
ran on a synthesized input
manuelmlmadeira/DeFoG/src/utils.py
code served (permissive licence) · get_code("06bef51bb1d6d083")
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, 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 3D tensor (batch_size=2, height=4, width=4)
E_3d = torch.randn(2, 4, 4)

# Create a 4D tensor (batch_size=2, height=4, width=4, channels=3)
E_4d = torch.randn(2, 4, 4, 3)

# Test with 3D tensor
ARGS = {
    "E": E_3d
}
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, 4, 4) — shape (2, 4, 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
to_list Ran
ran, violated the contract we drafted
manuelmlmadeira/DeFoG/src/datasets/guacamol_dataset.py
code served (permissive licence) · get_code("85832bcd5531c861")
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 1, dtype family int; 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
from typing import Sequence

torch.manual_seed(0)

# Test case 1: A sequence (list)
test_list = [1, 2, 3, 4]

# Test case 2: A sequence (tuple)
test_tuple = (5, 6, 7)

# Test case 3: A non-sequence value (int)
test_int = 42

# Test case 4: A non-sequence value (float)
test_float = 3.14

# Test case 5: A string (which is a Sequence but should be treated specially)
test_string = "hello"

# Test case 6: A numpy array (is a Sequence)
test_numpy = np.array([1, 2, 3])

# Test case 7: A torch tensor (is a Sequence)
test_tensor = torch.tensor([1, 2, 3, 4])

# Test case 8: A dict (not a Sequence)
test_dict = {"a": 1, "b": 2}

# We'll test with one representative case
ARGS = {
    "value": [1, 2, 3, 4]
}
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 1, dtype family int — 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
unnormalize Ran
ran; our contract draft was wrong, not the code
manuelmlmadeira/DeFoG/src/utils.py
code served (permissive licence) · get_code("1739cf39d80953b0")
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 PlaceHolder. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# Create small tensors for node features (X), edge features (E), and global features (y)
batch_size = 2
num_nodes = 4
node_feature_dim = 3
edge_feature_dim = 5
global_feature_dim = 2

# X: [batch_size, num_nodes, node_feature_dim]
X = torch.randn(batch_size, num_nodes, node_feature_dim)

# E: [batch_size, num_nodes, num_nodes, edge_feature_dim]
E = torch.randn(batch_size, num_nodes, num_nodes, edge_feature_dim)
# Make E symmetric (transpose dimensions 1 and 2)
E = (E + E.transpose(1, 2)) / 2

# y: [batch_size, global_feature_dim]
y = torch.randn(batch_size, global_feature_dim)

# norm_values: list of normalization scales for X, E, y
norm_values = [
    torch.ones(1, 1, node_feature_dim) * 2.0,  # scale for X
    torch.ones(1, 1, 1, edge_feature_dim) * 3.0,  # scale for E
    torch.ones(1, global_feature_dim) * 1.5,  # scale for y
]

# norm_biases: list of normalization biases for X, E, y
norm_biases = [
    torch.ones(1, 1, node_feature_dim) * 0.1,  # bias for X
    torch.ones(1, 1, 1, edge_feature_dim) * 0.2,  # bias for E
    torch.ones(1, global_feature_dim) * 0.05,  # bias for y
]

# node_mask: [batch_size, num_nodes] - indicates valid nodes (1 = valid, 0 = invalid)
node_mask = torch.ones(batch_size, num_nodes)
node_mask[0, 3] = 0  # mask out last node in first batch

ARGS = {
    "X": X,
    "E": E,
    "y": y,
    "norm_values": norm_values,
    "norm_biases": norm_biases,
    "node_mask": node_mask,
    "collapse": False,
}
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 PlaceHolder — a PlaceHolder output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a PlaceHolder 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
mol2smiles Not yet run
unverified (harvested, no recorded run)
manuelmlmadeira/DeFoG/src/analysis/rdkit_functions.py
code served (permissive licence) · get_code("0c9e13c0bf546794")
How it ran

no invocation recorded.

mol2smilesWithNoSanitize Not yet run
unverified (harvested, no recorded run)
manuelmlmadeira/DeFoG/src/analysis/rdkit_functions.py
code served (permissive licence) · get_code("85807cf49bba7f49")
How it ran

no invocation recorded.

remove_hydrogens Not yet run
unverified (harvested, no recorded run)
manuelmlmadeira/DeFoG/src/datasets/dataset_utils.py
code served (permissive licence) · get_code("dc5bc1ae656e5ea3")
How it ran

no invocation recorded.

Also present in other papers' code

Of 15 samples lifted from this paper's linked repositories, 15 carry a function identity and 0 do not. Of the 15 distinct functions with an identity, 11 are also present in other papers whose code we hold (11 with at least one paper dated earlier, 0 the same month, 10 later) and 4 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
l2 10 other papers
8 dated earlier, 2 later
1910.00760 (2019-10)
Efficient Graph Generation with Graph Recurrent Attention Networks
2204.01613 (2022-04)
SPECTRE: Spectral Conditioning Helps to Overcome the Expressivity Limits of One-shot Graph Generators
2305.19337 (2023-05)
HiGen: Hierarchical Graph Generative Networks
and 7 more not named here
load_pickle 9 other papers
8 dated earlier, 1 later
1907.00456 (2019-07)
Way Off-Policy Batch Deep Reinforcement Learning of Implicit Human Preferences in Dialog
2005.03545 (2020-05)
MISA: Modality-Invariant and -Specific Representations for Multimodal Sentiment Analysis
and 6 more not named here
degree_worker 6 other papers
5 dated earlier, 1 later
1910.00760 (2019-10)
Efficient Graph Generation with Graph Recurrent Attention Networks
2204.01613 (2022-04)
SPECTRE: Spectral Conditioning Helps to Overcome the Expressivity Limits of One-shot Graph Generators
2305.19337 (2023-05)
HiGen: Hierarchical Graph Generative Networks
and 3 more not named here
to_list 6 other papers
5 dated earlier, 1 later
2303.08137 (2023-03)
LayoutDM: Discrete Diffusion Model for Controllable Layout Generation
2310.15817 (2023-10)
Discriminator Guidance for Autoregressive Diffusion Models
2312.17397 (2023-12)
Classifier-free graph diffusion for molecular property targeting
and 3 more not named here
normalize 4 other papers
3 dated earlier, 1 later
2312.17397 (2023-12)
Classifier-free graph diffusion for molecular property targeting
2401.13858 (2024-01)
Graph Diffusion Transformers for Multi-Conditional Molecular Generation
2405.11416 (2024-05)
Discrete-state Continuous-time Diffusion for Graph Generation
and 1 more not named here
unnormalize 4 other papers
3 dated earlier, 1 later
2312.17397 (2023-12)
Classifier-free graph diffusion for molecular property targeting
2401.13858 (2024-01)
Graph Diffusion Transformers for Multi-Conditional Molecular Generation
2405.11416 (2024-05)
Discrete-state Continuous-time Diffusion for Graph Generation
and 1 more not named here
compare_hash 3 other papers
2 dated earlier, 1 later
2311.02142 (2023-11)
Sparse Training of Discrete Diffusion Models for Graph Generation
2312.17397 (2023-12)
Classifier-free graph diffusion for molecular property targeting
2603.28572 (2026-03)
Unrestrained Simplex Denoising for Discrete Data A Non-Markovian Approach Applied to Graph Generation
files_exist 3 other papers
2 dated earlier, 1 later
2312.17397 (2023-12)
Classifier-free graph diffusion for molecular property targeting
2403.06020 (2024-03)
Multi-conditioned Graph Diffusion for Neural Architecture Search
2603.28572 (2026-03)
Unrestrained Simplex Denoising for Discrete Data A Non-Markovian Approach Applied to Graph Generation
mol2smiles 2 other papers
1 dated earlier, 1 later
2312.17397 (2023-12)
Classifier-free graph diffusion for molecular property targeting
2603.28572 (2026-03)
Unrestrained Simplex Denoising for Discrete Data A Non-Markovian Approach Applied to Graph Generation
spectral_worker 2 other papers
1 dated earlier, 1 later
2312.17397 (2023-12)
Classifier-free graph diffusion for molecular property targeting
2603.28572 (2026-03)
Unrestrained Simplex Denoising for Discrete Data A Non-Markovian Approach Applied to Graph Generation
remove_hydrogens 1 other paper
1 dated earlier
2311.02142 (2023-11)
Sparse Training of Discrete Diffusion Models for Graph Generation

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 generative models are essential across diverse scientific domains by capturing complex distributions over relational data. Among them, graph diffusion models achieve superior performance but face inefficient sampling and limited flexibility due to the tight coupling between training and sampling stages. We introduce DeFoG, a novel graph generative framework that disentangles sampling from training, enabling a broader design space for more effective and efficient model optimization. DeFoG employs a discrete flow-matching formulation that respects the inherent symmetries of graphs. We theoretically ground this disentangled formulation by explicitly relating the training loss to the sampling algorithm and showing that DeFoG faithfully replicates the ground truth graph distribution. Building on these foundations, we thoroughly investigate DeFoG's design space and propose novel sampling methods that significantly enhance performance and reduce the required number of refinement steps. Extensive experiments demonstrate state-of-the-art performance across synthetic, molecular, and digital pathology datasets, covering both unconditional and conditional generation settings. It also outperforms most diffusion-based models with just 5-10% of their sampling steps.

For agents

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

get_harvested_code_for_paper("2410.04263")
get_code_for_paper("2410.04263")
have("2410.04263")

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

curl https://syntology.ai/api/ran/2410.04263.json

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

[![Syntology run record](https://syntology.ai/api/ran/2410.04263.svg)](https://syntology.ai/paper/2410.04263)

Connect an agent — have() is free.