SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1409.5185 · 2014

Deeply-Supervised Nets

arXiv · PDF · Open in the Atlas

Code that ran

20 of 21 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, 0 violated it and 20 have no contract verdict; 0 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/1409.5185.json · badge: /api/ran/1409.5185.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
ellisdg/3DUnetCNN community (classifier label)
classifier basis not recorded on the link
20 ran · 1 no recorded run
FunctionStatusWhere it lives
batch_loss Ran
ran on a synthesized input
ellisdg/3DUnetCNN/unet3d/train/training_utils.py
code served (permissive licence) · get_code("e0f6de538e2f1e18")
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. The image it ran in was not recorded. The code mentions cuda, and the harness aliases cuda to the CPU for every sample; whether this run needed that is 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 torch.nn as nn

torch.manual_seed(0)

# Define a minimal model that takes images and outputs predictions
class SimpleModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear = nn.Linear(4, 3)
    
    def forward(self, x):
        # x shape: (batch_size, 4)
        return self.linear(x)

# Define a minimal inferer that takes images and model
class SimpleInferer:
    def __call__(self, images, model):
        return model(images)

# Create model
model = SimpleModel()

# Create small tensors (batch=2, feature_dim=4 for images, num_classes=3 for target)
images = torch.randn(2, 4)
target = torch.randint(0, 3, (2,))

# Create loss criterion
criterion = nn.CrossEntropyLoss()

# Create inferer
inferer = SimpleInferer()

ARGS = {
    "model": model,
    "images": images,
    "target": target,
    "criterion": criterion,
    "n_gpus": 0,
    "use_amp": False,
    "inferer": inferer,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple — a tuple 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
calculate_origin_offset Ran
ran on a synthesized input
ellisdg/3DUnetCNN/unet3d/utils/affine.py
code served (permissive licence) · get_code("48624962e9942c9f")
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 1, dtype family float, 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 torch

torch.manual_seed(0)

# The function takes two parameters: new_spacing and old_spacing
# Looking at the function body, it uses torch.divide, torch.subtract operations
# These work on torch tensors, so both parameters should be tensors
# The function performs element-wise operations, so they should have compatible shapes

new_spacing = torch.tensor([2.0, 3.0, 4.0, 5.0])
old_spacing = torch.tensor([1.0, 1.5, 2.0, 2.5])

ARGS = {
    "new_spacing": new_spacing,
    "old_spacing": old_spacing,
}
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 1, dtype family float, 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
epoch_training Ran
ran on a synthesized input
ellisdg/3DUnetCNN/unet3d/train/training_utils.py
code served (permissive licence) · get_code("d5f1b71fdb23285a")
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. The code mentions cuda, and the harness aliases cuda to the CPU for every sample; whether this run needed that is 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 torch.optim as optim

torch.manual_seed(0)

# Create a simple model
class SimpleModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.fc = nn.Linear(4, 2)
    
    def forward(self, x):
        return self.fc(x)

# Create a simple criterion
criterion = nn.CrossEntropyLoss()

# Create model and optimizer
model = SimpleModel()
optimizer = optim.SGD(model.parameters(), lr=0.01)

# Create a simple data loader (list of dicts with 'image' and 'label' keys)
train_loader = [
    {
        'image': torch.randn(2, 4),
        'label': torch.randint(0, 2, (2,))
    },
    {
        'image': torch.randn(2, 4),
        'label': torch.randint(0, 2, (2,))
    }
]

ARGS = {
    'train_loader': train_loader,
    'model': model,
    'criterion': criterion,
    'optimizer': optimizer,
    'epoch': 0,
    'n_gpus': None,
    'print_frequency': 1,
    'print_gpu_memory': False,
    'scaler': None,
    'samples_per_epoch': None,
    'iteration': 1
}
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
epoch_validation Ran
ran on a synthesized input
ellisdg/3DUnetCNN/unet3d/train/training_utils.py
code served (permissive licence) · get_code("c0efd1b82154d3f9")
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. The code mentions cuda, and the harness aliases cuda to the CPU for every sample; whether this run needed that is 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 minimal model
class SimpleModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.fc = nn.Linear(4, 3)
    
    def forward(self, x):
        return self.fc(x)

# Build a minimal criterion
criterion = nn.CrossEntropyLoss()

# Build a minimal val_loader (list of dicts with 'image' and 'label' keys)
val_loader = [
    {'image': torch.randn(2, 4), 'label': torch.randint(0, 3, (2,))},
    {'image': torch.randn(2, 4), 'label': torch.randint(0, 3, (2,))},
]

# Build model
model = SimpleModel()

# Build inferer (callable that takes images and model, returns output)
class SimpleInferer:
    def __call__(self, images, model):
        return model(images)

inferer = SimpleInferer()

ARGS = {
    'val_loader': val_loader,
    'model': model,
    'criterion': criterion,
    'n_gpus': 0,
    'print_freq': 1,
    'use_amp': False,
    'inferer': inferer,
}
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
extract_gifti_array Ran
ran on a synthesized input
ellisdg/3DUnetCNN/unet3d/utils/cifti.py
code served (permissive licence) · get_code("d7f84b2efefc9200")
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 (10, 3) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

class GiftiArray:
    def __init__(self, data, metadata):
        self.data = data
        self.metadata = metadata

class GiftiObject:
    def __init__(self, darrays):
        self.darrays = darrays

# Create sample data arrays with metadata
array1_data = np.random.randn(10, 3).astype(np.float32)
array1_metadata = {
    'Name': 'CoordinateSystemXformNode',
    'GeometricType': 'Anatomical',
    'AnatomicalStructurePrimary': 'CortexLeft',
    'AnatomicalStructureSecondary': 'Gyrus'
}

array2_data = np.random.randn(8, 4).astype(np.float32)
array2_metadata = {
    'Name': 'Surfaces',
    'GeometricType': 'Functional',
    'AnatomicalStructurePrimary': 'CortexRight',
    'AnatomicalStructureSecondary': 'Sulcus'
}

gifti_arrays = [
    GiftiArray(array1_data, array1_metadata),
    GiftiArray(array2_data, array2_metadata)
]

gifti_object = GiftiObject(gifti_arrays)

ARGS = {
    'gifti_object': gifti_object,
    'index': 0,
    'geometric_type': 'Anatomical',
    'primary_anatomical_structure': 'CortexLeft',
    'secondary_anatomical_structure': 'Gyrus'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — 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 (10, 3) — shape (10, 3) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
extract_gifti_surface_vertices Ran
ran on a synthesized input
ellisdg/3DUnetCNN/unet3d/utils/cifti.py
code served (permissive licence) · get_code("ec8a60718108f746")
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 (10, 3) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

class MockGiftiArray:
    def __init__(self, data, name, geometric_type='Anatomical', primary_structure='CortexLeft', secondary_structure=None):
        self.data = data
        self.metadata = {
            'Name': name,
            'GeometricType': geometric_type,
            'AnatomicalStructurePrimary': primary_structure,
            'AnatomicalStructureSecondary': secondary_structure
        }

class MockGiftiObject:
    def __init__(self, darrays):
        self.darrays = darrays

# Create mock GIFTI surface object with vertex data
vertex_data = np.random.randn(10, 3).astype(np.float32)
array0 = MockGiftiArray(
    data=vertex_data,
    name='CoordinateSystemXYZ',
    geometric_type='Anatomical',
    primary_structure='CortexLeft',
    secondary_structure=None
)

array1 = MockGiftiArray(
    data=np.random.randn(10, 3).astype(np.int32),
    name='Normal',
    geometric_type='Anatomical',
    primary_structure='CortexLeft',
    secondary_structure=None
)

surface = MockGiftiObject(darrays=[array0, array1])

ARGS = {
    'surface': surface,
    'index': 0,
    'geometric_type': 'Anatomical',
    'primary_anatomical_structure': 'CortexLeft',
    'secondary_anatomical_structure': None
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — 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 (10, 3) — shape (10, 3) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
flip_image Ran
ran on a synthesized input
ellisdg/3DUnetCNN/unet3d/utils/augment.py
code served (permissive licence) · get_code("e3f4fafafb4fd349")
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 Image. 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)

class Image:
    def __init__(self, data):
        self.data = data
    
    def detach(self):
        return self.data.detach()
    
    def make_similar(self, data):
        return Image(data)

image = Image(torch.randn(2, 3, 4, 5))
axis = [2, 3]

ARGS = {
    "image": image,
    "axis": axis,
}
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 Image — a Image output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Image output the recorded ndim 4 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
generate_hcp_filenames Ran
ran on a synthesized input
ellisdg/3DUnetCNN/unet3d/utils/filenames.py
code served (permissive licence) · get_code("e6578b6b8b776391")
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. 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 shutil

# Create a temporary directory structure for testing
temp_dir = tempfile.mkdtemp()

# Create some test subject directories and files
subject_ids = ['100307', '100408']
for subject_id in subject_ids:
    subject_dir = os.path.join(temp_dir, subject_id)
    os.makedirs(subject_dir, exist_ok=True)
    
    # Create dummy feature files
    with open(os.path.join(subject_dir, 'features.nii.gz'), 'w') as f:
        f.write('dummy feature data')
    
    # Create dummy metric files
    with open(os.path.join(subject_dir, 'metrics.nii.gz'), 'w') as f:
        f.write('dummy metric data')
    
    # Create dummy surface files (using template format)
    for hemi in ['L', 'R']:
        surface_file = os.path.join(subject_dir, f'surface_{hemi}_{subject_id}.surf.gii')
        with open(surface_file, 'w') as f:
            f.write('dummy surface data')

ARGS = {
    'directory': temp_dir,
    'surface_basename_template': 'surface_{hemi}_{subject_id}.surf.gii',
    'target_basenames': 'metrics.nii.gz',
    'feature_basenames': 'features.nii.gz',
    'subject_ids': subject_ids,
    'hemispheres': ['L', 'R']
}
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 — a list 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
get_feature_filename_and_subject_id Ran
ran on a synthesized input
ellisdg/3DUnetCNN/unet3d/predict/utils.py
code served (permissive licence) · get_code("6db950092e32b901")
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 MockDataset:
    def __init__(self):
        self.feature_index = 0
        self.epoch_filenames = [
            ['data/subject_001/session_01/features.npy', 'subject_001'],
            ['data/subject_002/session_01/features.npy', 'subject_002'],
            ['data/subject_001/session_02/features.npy', 'subject_001'],
        ]

dataset = MockDataset()
idx = 0

ARGS = {
    'dataset': dataset,
    'idx': idx,
    'verbose': False,
}
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
get_image Ran
ran on a synthesized input
ellisdg/3DUnetCNN/unet3d/utils/image.py
code served (permissive licence) · get_code("50a36c3ef53a39bf")
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 Image. 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
from monai.data.meta_tensor import MetaTensor

torch.manual_seed(0)

# Create a small tensor for image data
data = torch.randn(2, 4, 3, 5)

# Create an affine matrix (4x4)
affine = torch.eye(4)

# Optional kwargs
kwargs = {}

ARGS = {
    "data": data,
    "affine": affine,
    "kwargs": kwargs
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
monai, nibabel, numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Image — a Image 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
get_spacing_from_affine Ran
ran on a synthesized input
ellisdg/3DUnetCNN/unet3d/utils/affine.py
code served (permissive licence) · get_code("ac1b3032b0af0eeb")
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 1, dtype family float, shape (3,) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# The function expects an affine transformation matrix
# It accesses affine[:3, :3] which is a 3x3 slice
# So we need at least a 3x3 matrix, but we can make it larger (4x4 is common for affine transforms)
affine = torch.tensor([
    [1.5, 0.1, 0.2, 10.0],
    [0.0, 2.0, 0.1, 20.0],
    [0.1, 0.0, 3.5, 30.0],
    [0.0, 0.0, 0.0, 1.0]
], dtype=torch.float32)

ARGS = {
    "affine": affine
}
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 1, dtype family float, shape (3,) — shape (3,) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
infer_subject_id Ran
ran on a synthesized input
ellisdg/3DUnetCNN/unet3d/predict/utils.py
code served (permissive licence) · get_code("7855782a9942e2e4")
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.

ARGS = {
    "filename": "path/to/subject_001/data/file.txt",
    "ref_filename": "path/to/reference/data/file.txt"
}
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 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
load_images_from_dataset Ran
ran on a synthesized input
ellisdg/3DUnetCNN/unet3d/predict/volumetric.py
code served (permissive licence) · get_code("d175ff824428b980")
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. 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 torch.nn as nn

torch.manual_seed(0)

class MockDataset:
    def __init__(self):
        pass
    
    def get_feature_image(self, idx, return_unmodified=False):
        x_image = torch.randn(2, 4)
        if return_unmodified:
            ref_image = torch.randn(3, 5)
            return x_image, ref_image
        else:
            return x_image

dataset = MockDataset()
idx = 0
resample_predictions = True

ARGS = {
    "dataset": dataset,
    "idx": idx,
    "resample_predictions": resample_predictions,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple — a tuple 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
load_state_dict Ran
ran on a synthesized input
ellisdg/3DUnetCNN/unet3d/models/build.py
code served (permissive licence) · get_code("fb1d0885dc8670c6")
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 SimpleModel. 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 torch.nn as nn
import math

torch.manual_seed(0)

# Create a simple model
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

# Create model instance
model = SimpleModel()

# Create a state dict with slightly different shapes for testing shape matching
state_dict = {
    'linear1.weight': torch.randn(3, 4),  # Different from model's (3, 4)
    'linear1.bias': torch.randn(3),
    'linear2.weight': torch.randn(2, 3),
    'linear2.bias': torch.randn(2),
}

# Setup arguments
ARGS = {
    'model': model,
    'state_dict': state_dict,
    'n_gpus': 1,
    'strict': 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 SimpleModel — a SimpleModel output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a SimpleModel 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
load_volumetric_model Ran
ran on a synthesized input
ellisdg/3DUnetCNN/unet3d/predict/volumetric.py
code served (permissive licence) · get_code("a952576fb1f2c26a")
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 MinimalUNet3D. 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 torch.nn as nn
from unittest.mock import Mock, patch, MagicMock

torch.manual_seed(0)

# Create a minimal mock for build_or_load_model
class MinimalUNet3D(nn.Module):
    def __init__(self):
        super().__init__()
        self.conv1 = nn.Conv3d(1, 8, kernel_size=3)
        self.conv2 = nn.Conv3d(8, 16, kernel_size=3)
    
    def forward(self, x):
        x = self.conv1(x)
        x = self.conv2(x)
        return x

def mock_build_or_load_model(model_name, model_filename, n_gpus, strict, **kwargs):
    """Mock implementation of build_or_load_model"""
    model = MinimalUNet3D()
    return model

# Patch the import before defining ARGS
import sys
from unittest.mock import MagicMock

# Create a mock unet3d module structure
mock_unet3d = MagicMock()
mock_unet3d.models.build.build_or_load_model = mock_build_or_load_model
sys.modules['unet3d'] = mock_unet3d
sys.modules['unet3d.models'] = mock_unet3d.models
sys.modules['unet3d.models.build'] = mock_unet3d.models.build

ARGS = {
    'model_name': 'unet3d',
    'model_filename': 'model.pt',
    'n_gpus': 1,
    'strict': True,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch, unet3d — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type MinimalUNet3D — a MinimalUNet3D output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a MinimalUNet3D 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
pytorch_predict_batch_array Ran
ran on a synthesized input
ellisdg/3DUnetCNN/unet3d/predict/utils.py
code served (permissive licence) · get_code("0a4754cb9d311c12")
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, 3) as a later run recorded it. The image it ran in was not recorded. The code mentions cuda, and the harness aliases cuda to the CPU for every sample; whether this run needed that is not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn

torch.manual_seed(0)

# Create a minimal model that satisfies the function's requirements
class SimpleModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear = nn.Linear(4, 3)
    
    def forward(self, x):
        return self.linear(x)

# Instantiate the model
model = SimpleModel()

# Create batch of tiny tensors (2 samples, 4 features each)
batch = [
    torch.randn(4),
    torch.randn(4)
]

# Define arguments for pytorch_predict_batch_array
ARGS = {
    'model': model,
    'batch': batch,
    'n_gpus': 0
}
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, 3) — shape (2, 3) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
random_flip_dimensions Ran
ran on a synthesized input
ellisdg/3DUnetCNN/unet3d/utils/augment.py
code served (permissive licence) · get_code("360dfc67628c771e")
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 numpy as np

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

# The function random_flip_dimensions takes one parameter: n_dimensions
# It iterates through range(n_dimensions) and randomly decides whether to append each dimension index to a list
# So n_dimensions should be an integer representing the number of dimensions

n_dimensions = 5

ARGS = {
    "n_dimensions": n_dimensions
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — 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
random_scale_factor Ran
ran on a synthesized input
ellisdg/3DUnetCNN/unet3d/utils/augment.py
code served (permissive licence) · get_code("c746ac1567201dd2")
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 1, dtype family float, shape (3,) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

ARGS = {
    "n_dim": 3,
    "mean": 1.0,
    "std": 0.25,
}
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 1, dtype family float, shape (3,) — shape (3,) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
set_affine_spacing Ran
ran on a synthesized input
ellisdg/3DUnetCNN/unet3d/utils/affine.py
code served (permissive licence) · get_code("aed17f756b9cc8c3")
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 (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 4x4 affine transformation matrix
# The upper-left 3x3 block contains rotation, zoom, and shear (RZS)
# The rightmost column contains translation
affine = torch.tensor([
    [1.5, 0.0, 0.0, 10.0],
    [0.0, 2.0, 0.0, 20.0],
    [0.0, 0.0, 1.5, 30.0],
    [0.0, 0.0, 0.0, 1.0]
], dtype=torch.float32)

# Create spacing values (one for each dimension)
spacing = torch.tensor([1.0, 2.0, 1.5], dtype=torch.float32)

ARGS = {
    "affine": affine,
    "spacing": spacing
}
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 (4, 4) — shape (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
wrapped_partial Ran
ran on a synthesized input
ellisdg/3DUnetCNN/unet3d/utils/filenames.py
code served (permissive licence) · get_code("1e0de6b9c650bc32")
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 partial. 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 torch.nn as nn
from functools import partial, update_wrapper

torch.manual_seed(0)

# Define a simple function to wrap
def add_numbers(a, b, c=0):
    """Simple function that adds numbers."""
    return a + b + c

# Define another function for testing
def multiply_and_add(x, y, z=1):
    """Multiply x by y and add z."""
    return x * y + z

# Define a tensor-based function
def tensor_operation(tensor, scale=1.0, offset=0.0):
    """Apply scale and offset to a tensor."""
    return tensor * scale + offset

# Test with a simple function
test_func = add_numbers
test_args = (5,)
test_kwargs = {"c": 3}

# Test with tensor function
tensor_func = tensor_operation
tensor_arg = torch.randn(2, 4)
tensor_kwargs = {"scale": 2.0, "offset": 1.0}

# Construct the ARGS dictionary
ARGS = {
    "func": test_func,
    "args": test_args,
    "kwargs": test_kwargs,
}
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 partial — a partial output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a partial 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
extract_gifti_array_names Not yet run
unverified (harvested, no recorded run)
ellisdg/3DUnetCNN/unet3d/utils/cifti.py
code served (permissive licence) · get_code("a613b779df9b18fb")
How it ran

no invocation recorded.

Also present in other papers' code

Of 21 samples lifted from this paper's linked repositories, 21 carry a function identity and 0 do not. Of the 21 distinct functions with an identity, 1 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 1 later) and 20 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

FunctionAlso present inEarliest-datedOthers named
wrapped_partial 1 other paper
1 later
1910.06849 (2019-10)
DeepGCNs: Making GCNs Go as Deep as CNNs

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

Our proposed deeply-supervised nets (DSN) method simultaneously minimizes classification error while making the learning process of hidden layers direct and transparent. We make an attempt to boost the classification performance by studying a new formulation in deep networks. Three aspects in convolutional neural networks (CNN) style architectures are being looked at: (1) transparency of the intermediate layers to the overall classification; (2) discriminativeness and robustness of learned features, especially in the early layers; (3) effectiveness in training due to the presence of the exploding and vanishing gradients. We introduce "companion objective" to the individual hidden layers, in addition to the overall objective at the output layer (a different strategy to layer-wise pre-training). We extend techniques from stochastic gradient methods to analyze our algorithm. The advantage of our method is evident and our experimental result on benchmark datasets shows significant performance gain over existing methods (e.g. all state-of-the-art results on MNIST, CIFAR-10, CIFAR-100, and SVHN).

For agents

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

get_harvested_code_for_paper("1409.5185")
get_code_for_paper("1409.5185")
have("1409.5185")

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

curl https://syntology.ai/api/ran/1409.5185.json

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

[![Syntology run record](https://syntology.ai/api/ran/1409.5185.svg)](https://syntology.ai/paper/1409.5185)

Connect an agent — have() is free.