SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1904.01561 · 2019

Analyzing Learned Molecular Representations for Property Prediction

arXiv · PDF · Open in the Atlas

Code that ran

14 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; 6 honoured the contract we drafted, 0 violated it and 5 have no contract verdict; 3 failures of our instrument; 1 no recorded run), measured 2026-11-22; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/1904.01561.json · badge: /api/ran/1904.01561.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
tingweidaniel/chemprop_atomicfp_save_depth_final community (classifier label)
classifier basis not recorded on the link
8 ran · 1 no recorded run
anonymous20201002/chemprop community (classifier label)
classifier basis not recorded on the link
6 ran · 0 no recorded run
FunctionStatusWhere it lives
compute_gnorm Ran
ran, honoured the contract we drafted
anonymous20201002/chemprop/chemprop/nn_utils.py
pointer only (licence: MIT); the stored copy get_code serves, from zihao-wang/ReactionOOD, is pointer only (licence: NONE) · get_code("8efee7d886650920")
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 zihao-wang/ReactionOOD, 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.

compute_gnorm Ran
ran, honoured the contract we drafted
tingweidaniel/chemprop_atomicfp_save_depth_final/chemprop/nn_utils.py
code served (permissive licence) · get_code("ca7675a6350f0967")
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 torch.nn as nn
import math

torch.manual_seed(0)

# Create a simple model with parameters that have gradients
class SimpleModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear1 = nn.Linear(4, 3)
        self.linear2 = nn.Linear(3, 2)
    
    def forward(self, x):
        return self.linear2(self.linear1(x))

# Instantiate the model
model = SimpleModel()

# Create some dummy input and compute forward/backward to populate gradients
x = torch.randn(2, 4)
y = torch.randn(2, 2)
output = model(x)
loss = ((output - y) ** 2).mean()
loss.backward()

ARGS = {
    "model": model
}
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
compute_pnorm Ran
ran, honoured the contract we drafted
anonymous20201002/chemprop/chemprop/nn_utils.py
pointer only (licence: MIT); the stored copy get_code serves, from zihao-wang/ReactionOOD, is pointer only (licence: NONE) · get_code("c4eed975523e1ed4")
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 zihao-wang/ReactionOOD, 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.

compute_pnorm Ran
ran, honoured the contract we drafted
tingweidaniel/chemprop_atomicfp_save_depth_final/chemprop/nn_utils.py
code served (permissive licence) · get_code("e415ae10bb5a4e3b")
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 torch.nn as nn
import math

torch.manual_seed(0)

# Create a minimal nn.Module with some parameters
class SimpleModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear1 = nn.Linear(4, 3)
        self.linear2 = nn.Linear(3, 2)
    
    def forward(self, x):
        x = self.linear1(x)
        x = self.linear2(x)
        return x

model = SimpleModel()

ARGS = {
    "model": model
}
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_predictions Ran
ran on a synthesized input
tingweidaniel/chemprop_atomicfp_save_depth_final/chemprop/train/evaluate.py
code served (permissive licence) · get_code("37db24e9e0f38c01")
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 logging
from sklearn.metrics import accuracy_score, f1_score
import numpy as np

# Set up logger
logger = logging.getLogger("test_logger")
logger.setLevel(logging.INFO)
handler = logging.StreamHandler()
logger.addHandler(handler)

# Define predictions and targets
# 3 data points, 2 tasks
num_tasks = 2
preds = [
    [0.9, 0.1],   # data point 0: task 0 pred=0.9, task 1 pred=0.1
    [0.2, 0.8],   # data point 1: task 0 pred=0.2, task 1 pred=0.8
    [0.7, 0.3],   # data point 2: task 0 pred=0.7, task 1 pred=0.3
]

targets = [
    [1, 0],       # data point 0: task 0 target=1, task 1 target=0
    [0, 1],       # data point 1: task 0 target=0, task 1 target=1
    [1, None],    # data point 2: task 0 target=1, task 1 target=None (invalid)
]

# Define a simple metric function for regression
def metric_func_regression(targets_list, preds_list):
    """Simple metric: mean absolute error"""
    return np.mean(np.abs(np.array(targets_list) - np.array(preds_list)))

# Define a metric function for classification
def metric_func_classification(targets_list, preds_list):
    """Simple metric: accuracy"""
    preds_binary = [1 if p > 0.5 else 0 for p in preds_list]
    return accuracy_score(targets_list, preds_binary)

# Define a metric function for multiclass
def metric_func_multiclass(targets_list, preds_list, labels=None):
    """Simple metric: accuracy for multiclass"""
    preds_binary = [np.argmax(p) for p in preds_list]
    return accuracy_score(targets_list, preds_binary)

ARGS = {
    "preds": preds,
    "targets": targets,
    "num_tasks": num_tasks,
    "metric_func": metric_func_regression,
    "dataset_type": "regression",
    "logger": logger,
}
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 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
get_atom_fdim Ran
ran on a synthesized input
tingweidaniel/chemprop_atomicfp_save_depth_final/chemprop/features/featurization.py
code served (permissive licence) · get_code("513750ee5fcea171")
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 int, ndim 0, dtype family int, 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.

from argparse import Namespace

torch_seed = 0
import torch
torch.manual_seed(torch_seed)

ARGS = {
    'args': Namespace()
}
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 int, ndim 0, dtype family int, 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_bond_fdim Ran
ran on a synthesized input
tingweidaniel/chemprop_atomicfp_save_depth_final/chemprop/features/featurization.py
code served (permissive licence) · get_code("963d26c20eb3e4da")
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 int, ndim 0, dtype family int, 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.

from argparse import Namespace

ARGS = {
    "args": Namespace()
}
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 int, ndim 0, dtype family int, 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_checkpoint_paths Ran
ran; our contract draft was wrong, not the code
anonymous20201002/chemprop/chemprop/args.py
code served (permissive licence) · get_code("8fb802f2499b4ccb")
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 os
import tempfile
from typing import List, Optional

# Create a temporary directory with some test checkpoint files
temp_dir = tempfile.mkdtemp()
test_checkpoint_1 = os.path.join(temp_dir, "checkpoint_1.pt")
test_checkpoint_2 = os.path.join(temp_dir, "checkpoint_2.pt")
test_checkpoint_3 = os.path.join(temp_dir, "other_file.txt")

# Create dummy files
open(test_checkpoint_1, 'w').close()
open(test_checkpoint_2, 'w').close()
open(test_checkpoint_3, 'w').close()

# Create subdirectory with checkpoint
subdir = os.path.join(temp_dir, "subdir")
os.makedirs(subdir)
test_checkpoint_3 = os.path.join(subdir, "checkpoint_3.pt")
open(test_checkpoint_3, 'w').close()

ARGS = {
    "checkpoint_path": None,
    "checkpoint_paths": None,
    "checkpoint_dir": temp_dir,
    "ext": ".pt"
}
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 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
load_features Ran
ran on a synthesized input
tingweidaniel/chemprop_atomicfp_save_depth_final/chemprop/features/utils.py
code served (permissive licence) · get_code("5045ed0a87b443be")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned ndarray, ndim 2, dtype family float, shape (5, 8) 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 csv
import pickle
import tempfile
import numpy as np
from scipy import sparse

# Create temporary directory for test files
tmpdir = tempfile.mkdtemp()

# Create test .npy file
npy_path = os.path.join(tmpdir, "test_features.npy")
np.save(npy_path, np.random.rand(5, 8))

# Create test .npz file
npz_path = os.path.join(tmpdir, "test_features.npz")
np.savez(npz_path, features=np.random.rand(5, 8))

# Create test .csv file
csv_path = os.path.join(tmpdir, "test_features.csv")
with open(csv_path, 'w', newline='') as f:
    writer = csv.writer(f)
    writer.writerow(['feat1', 'feat2', 'feat3'])
    for _ in range(5):
        writer.writerow([str(np.random.rand()) for _ in range(3)])

# Create test .txt file
txt_path = os.path.join(tmpdir, "test_features.txt")
with open(txt_path, 'w', newline='') as f:
    writer = csv.writer(f)
    writer.writerow(['feat1', 'feat2', 'feat3'])
    for _ in range(5):
        writer.writerow([str(np.random.rand()) for _ in range(3)])

# Create test .pkl file with sparse array
pkl_path = os.path.join(tmpdir, "test_features.pkl")
sparse_features = [sparse.csr_matrix(np.random.rand(1, 8)) for _ in range(5)]
with open(pkl_path, 'wb') as f:
    pickle.dump(sparse_features, f)

# Use the .npy path as the main test case
ARGS = {
    'path': npy_path
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
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 2, dtype family float, shape (5, 8) — shape (5, 8) 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
onek_encoding_unk Ran
ran; our contract draft was wrong, not the code
tingweidaniel/chemprop_atomicfp_save_depth_final/chemprop/features/featurization.py
code served (permissive licence) · get_code("3e07b9e0170215ca")
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

torch.manual_seed(0)

# Construct arguments for onek_encoding_unk function
# Parameters: value (int), choices (List[int])

# Test case 1: value is in choices
value_1 = 2
choices_1 = [1, 2, 3, 4]

# Test case 2: value is not in choices (should trigger unknown encoding)
value_2 = 5
choices_2 = [1, 2, 3, 4]

# We'll use the first test case for ARGS
ARGS = {
    "value": value_1,
    "choices": choices_1
}
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
onek_encoding_unk Ran
ran; our contract draft was wrong, not the code
anonymous20201002/chemprop/chemprop/features/featurization.py
code served (permissive licence) · get_code("49ebe299656c7302")
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

# Setup for onek_encoding_unk function
# The function takes:
# - value: an int that we want to encode
# - choices: a List[int] of possible values

# Test case 1: value is in choices
value_1 = 5
choices_1 = [1, 3, 5, 7, 9]

# Test case 2: value is not in choices (will use the unknown category)
value_2 = 10
choices_2 = [1, 3, 5, 7, 9]

# We'll use the first test case for the ARGS
ARGS = {
    "value": value_1,
    "choices": choices_1
}
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
param_count Ran
ran, honoured the contract we drafted
anonymous20201002/chemprop/chemprop/nn_utils.py
pointer only (licence: MIT); the stored copy get_code serves, from zihao-wang/ReactionOOD, is pointer only (licence: NONE) · get_code("f1cc5264ce058bab")
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 zihao-wang/ReactionOOD, 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.

param_count Ran
ran, honoured the contract we drafted
tingweidaniel/chemprop_atomicfp_save_depth_final/chemprop/nn_utils.py
code served (permissive licence) · get_code("d68ff6a82a950b49")
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 int, ndim 0, dtype family int, 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)

# Build a simple model with trainable parameters
class SimpleModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear1 = nn.Linear(4, 3)
        self.linear2 = nn.Linear(3, 2)
    
    def forward(self, x):
        x = self.linear1(x)
        x = self.linear2(x)
        return x

model = SimpleModel()

ARGS = {
    "model": model
}
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 int, ndim 0, dtype family int, 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
predict Ran
ran on a synthesized input
anonymous20201002/chemprop/chemprop/sklearn_train.py
code served (permissive licence) · get_code("c9a8d90e903cc480")
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 numpy as np
from sklearn.ensemble import RandomForestClassifier, RandomForestRegressor
from sklearn.svm import SVC, SVR

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

# Create small training data
X_train = np.random.randn(10, 4)
y_train_regression = np.random.randn(10)
y_train_binary = np.random.randint(0, 2, 10)

# Test features
X_test = np.random.randn(2, 4)

# Create a trained RandomForestRegressor
rf_regressor = RandomForestRegressor(n_estimators=2, random_state=0)
rf_regressor.fit(X_train, y_train_regression)

# Create a trained RandomForestClassifier
rf_classifier = RandomForestClassifier(n_estimators=2, random_state=0)
rf_classifier.fit(X_train, y_train_binary)

# Create a trained SVM Regressor
svm_regressor = SVR(kernel='linear')
svm_regressor.fit(X_train, y_train_regression)

# Create a trained SVM Classifier
svm_classifier = SVC(kernel='linear', probability=False)
svm_classifier.fit(X_train, y_train_binary)

# Test case 1: RandomForestRegressor with regression
ARGS = {
    'model': rf_regressor,
    'model_type': 'random_forest',
    'dataset_type': 'regression',
    'features': X_test
}
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 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
load_args Not yet run
unverified (harvested, no recorded run)
tingweidaniel/chemprop_atomicfp_save_depth_final/chemprop/utils.py
code served (permissive licence) · get_code("76e27ba5baa89e9f")
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 11 distinct functions with an identity, 6 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 6 later) and 5 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
param_count 3 other papers
3 later
2005.10036 (2020-05)
Uncertainty Quantification Using Neural Networks for Molecular Property Prediction
2310.03152 (2023-10)
Towards out-of-distribution generalizable predictions of chemical kinetic properties
2502.12979 (2025-02)
Electron flow matching for generative reaction mechanism prediction obeying conservation laws
compute_gnorm 2 other papers
2 later
2005.10036 (2020-05)
Uncertainty Quantification Using Neural Networks for Molecular Property Prediction
2310.03152 (2023-10)
Towards out-of-distribution generalizable predictions of chemical kinetic properties
compute_pnorm 2 other papers
2 later
2005.10036 (2020-05)
Uncertainty Quantification Using Neural Networks for Molecular Property Prediction
2310.03152 (2023-10)
Towards out-of-distribution generalizable predictions of chemical kinetic properties
onek_encoding_unk 2 other papers
2 later
2305.01520 (2023-05)
Conditional Graph Information Bottleneck for Molecular Relational Learning
2305.18451 (2023-05)
Shift-Robust Molecular Relational Learning with Causal Substructure
get_checkpoint_paths 1 other paper
1 later
2310.03152 (2023-10)
Towards out-of-distribution generalizable predictions of chemical kinetic properties
load_args 1 other paper
1 later
2005.10036 (2020-05)
Uncertainty Quantification Using Neural Networks for Molecular Property Prediction

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-22.

Repositories linked to this paper

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

Abstract

Advancements in neural machinery have led to a wide range of algorithmic solutions for molecular property prediction. Two classes of models in particular have yielded promising results: neural networks applied to computed molecular fingerprints or expert-crafted descriptors, and graph convolutional neural networks that construct a learned molecular representation by operating on the graph structure of the molecule. However, recent literature has yet to clearly determine which of these two methods is superior when generalizing to new chemical space. Furthermore, prior research has rarely examined these new models in industry research settings in comparison to existing employed models. In this paper, we benchmark models extensively on 19 public and 16 proprietary industrial datasets spanning a wide variety of chemical endpoints. In addition, we introduce a graph convolutional model that consistently matches or outperforms models using fixed molecular descriptors as well as previous graph neural architectures on both public and proprietary datasets. Our empirical findings indicate that while approaches based on these representations have yet to reach the level of experimental reproducibility, our proposed model nevertheless offers significant improvements over models currently used in industrial workflows.

For agents

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

get_harvested_code_for_paper("1904.01561")
get_code_for_paper("1904.01561")
have("1904.01561")

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

curl https://syntology.ai/api/ran/1904.01561.json

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

[![Syntology run record](https://syntology.ai/api/ran/1904.01561.svg)](https://syntology.ai/paper/1904.01561)

Connect an agent — have() is free.