SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1608.06993 · 2016

Densely Connected Convolutional Networks

Kilian Weinberger, Gao Huang, Laurens Van Der Maaten, Zhuang Liu

arXiv · PDF · Open in the Atlas

Code that ran

48 of 71 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 3 honoured the contract we drafted, 1 violated it and 27 have no contract verdict; 17 failures of our instrument; 23 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/1608.06993.json · badge: /api/ran/1608.06993.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
lionelmessi6410/ntga community (classifier label)
classifier basis not recorded on the link
7 ran · 3 no recorded run
Sakib1263/DenseNet-1D-2D-Tensorflow-Keras community (classifier label)
classifier basis not recorded on the link
5 ran · 0 no recorded run
GeoTrouvetout/Vehicle_ReID community (classifier label)
classifier basis not recorded on the link
4 ran · 1 no recorded run
0jason000/DenseNet community (classifier label)
classifier basis not recorded on the link
4 ran · 1 no recorded run
pudae/tensorflow-densenet community (classifier label)
classifier basis not recorded on the link
3 ran · 4 no recorded run
RaghavHub/rekuten_hack community (classifier label)
classifier basis not recorded on the link
3 ran · 1 no recorded run
jiweibo/imagenet reimplementation (classifier label)
classifier basis: classifier_2025
3 ran · 0 no recorded run
fengjiqiang/pretrainedmodel_pytorch reimplementation (classifier label)
classifier basis: classifier_2025
3 ran · 0 no recorded run
atapour/ransomware-classification community (classifier label)
classifier basis not recorded on the link
2 ran · 2 no recorded run
hycis/TensorGraph community (classifier label)
classifier basis not recorded on the link
2 ran · 2 no recorded run
seasonyc/densenet reimplementation (classifier label)
classifier basis: classifier_2025
2 ran · 0 no recorded run
okason97/DenseNet-Tensorflow2 reimplementation (classifier label)
classifier basis: classifier_2025
2 ran · 0 no recorded run
okason97/handshape-recognition community (classifier label)
classifier basis not recorded on the link
1 ran · 2 no recorded run
ApexPredator1/DenseNet_tensorflow reimplementation (classifier label)
classifier basis: classifier_2025
1 ran · 0 no recorded run
noahfl/densenet-sdr extension (classifier label)
classifier basis: classifier_2025
1 ran · 0 no recorded run
cges60809/DenseNet reimplementation (classifier label)
classifier basis: classifier_2025
1 ran · 0 no recorded run
ByakuyaUncia/DenseNetUncia reimplementation (classifier label)
classifier basis: classifier_2025
1 ran · 0 no recorded run
titu1994/DenseNet community (classifier label)
classifier basis not recorded on the link
1 ran · 0 no recorded run
bamos/densenet.pytorch community (classifier label)
classifier basis not recorded on the link
1 ran · 0 no recorded run
satishjasthi/Densenet_smplified reimplementation (classifier label)
classifier basis: classifier_2025
1 ran · 0 no recorded run
bfortuner/pytorch_tiramisu reimplementation (classifier label)
classifier basis: classifier_2025
0 ran · 3 no recorded run
DaikiTanak/manifold_mixup extension (classifier label)
classifier basis: classifier_2025
0 ran · 3 no recorded run
idobronstein/vision_networks community (classifier label)
classifier basis not recorded on the link
0 ran · 1 no recorded run
FunctionStatusWhere it lives
CNN Ran
ran on a synthesized input
lionelmessi6410/ntga/models/cnn.py
code served (permissive licence) · get_code("58b75b74e06f3539")
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 Functional. 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 tensorflow as tf

# Define arguments for the CNN function
input_shape = (28, 28, 3)  # Height, Width, Channels - small image dimensions
num_classes = 10  # Number of output classes (e.g., for CIFAR-10 or MNIST with color)

ARGS = {
    'input_shape': input_shape,
    'num_classes': num_classes
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Functional — a Functional output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Functional 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
Conv_1D_Block Ran
ran; our contract draft was wrong, not the code
Sakib1263/DenseNet-1D-2D-Tensorflow-Keras/Codes/DenseNet_1DCNN.py
code served (permissive licence) · get_code("fa041da57fc2fe89")
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 KerasTensor. 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 tensorflow as tf
import numpy as np

# Create a small input tensor
# Shape: (batch_size, sequence_length, input_channels)
# Using distinct dimensions: batch=2, seq_len=4, channels=3
x = tf.keras.Input(shape=(4, 3), batch_size=2)

# Define the parameters
model_width = 5  # output channels
kernel = 3  # kernel size
strides = 1  # stride

ARGS = {
    'x': x,
    'model_width': model_width,
    'kernel': kernel,
    'strides': strides
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type KerasTensor — a KerasTensor 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
Conv_2D_Block Ran
ran on a synthesized input
Sakib1263/DenseNet-1D-2D-Tensorflow-Keras/Codes/DenseNet_2DCNN.py
code served (permissive licence) · get_code("b660188e72021517")
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 KerasTensor. 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 tensorflow as tf
import numpy as np

# Create a small input tensor (batch=2, height=4, width=3, channels=1)
inputs = tf.keras.Input(shape=(4, 3, 1))

# Define parameters
model_width = 8
kernel = (3, 3)
strides = (1, 1)

# Build ARGS dictionary
ARGS = {
    'inputs': inputs,
    'model_width': model_width,
    'kernel': kernel,
    'strides': strides
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type KerasTensor — a KerasTensor 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
DNN_ReLU Ran
ran on a synthesized input
lionelmessi6410/ntga/models/dnn_relu.py
code served (permissive licence) · get_code("c4f13d6b734dbd11")
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 Functional. 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 tensorflow as tf
import numpy as np

ARGS = {
    'input_shape': (4,),
    'num_classes': 3
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Functional — a Functional output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Functional 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
augment_all_images Ran
ran on a synthesized input
RaghavHub/rekuten_hack/data_providers/cifar.py
code served (permissive licence) · get_code("bb510aa46f89d5e6")
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 4, dtype family float, shape (2, 4, 3, 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
import random

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

# Create a small batch of images
# Shape: (batch_size, height, width, channels)
# Using distinct dimensions: batch=2, height=4, width=3, channels=3
initial_images = np.random.rand(2, 4, 3, 3).astype(np.float32)

# Padding amount
pad = 4

ARGS = {
    "initial_images": initial_images,
    "pad": pad
}
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 4, dtype family float, shape (2, 4, 3, 3) — shape (2, 4, 3, 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
augment_image Ran
ran on a synthesized input
RaghavHub/rekuten_hack/data_providers/cifar.py
code served (permissive licence) · get_code("a33a7d8933a73deb")
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 3, dtype family float, shape (4, 5, 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 random
import numpy as np

# Set random seed for reproducibility
random.seed(42)
np.random.seed(42)

# Create a small test image: height=4, width=5, channels=3
image = np.random.randint(0, 256, size=(4, 5, 3), dtype=np.uint8)

# Padding amount
pad = 2

ARGS = {
    "image": image,
    "pad": pad,
}
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 3, dtype family float, shape (4, 5, 3) — shape (4, 5, 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
bytes_feature Ran
ran on a synthesized input
pudae/tensorflow-densenet/datasets/dataset_utils.py
code served (permissive licence) · get_code("9362d2135d7c4e5d")
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 Feature. 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 tensorflow as tf

# The function expects 'values' to be a string that will be encoded to bytes
# by tf.train.BytesList

ARGS = {
    "values": b"example_string_data"
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Feature — a Feature output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Feature 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
calculate_f1_score Ran
ran on a synthesized input
atapour/ransomware-classification/utils.py
code served (permissive licence) · get_code("564a2e986e0cf042")
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 numpy as np
import sklearn.metrics as metrics

torch.manual_seed(0)

# The function takes ground truth labels, predictions, and an average method
# gt and pred should be array-like with class labels
# For f1_score, we need integer class labels

# Create small ground truth labels (2 samples, 3 classes)
gt = np.array([0, 1])

# Create small prediction labels (2 samples, matching gt length)
pred = np.array([0, 2])

# average parameter is already set to default 'weighted'
average = 'weighted'

ARGS = {
    'gt': gt,
    'pred': pred,
    'average': average
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, sklearn, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type 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
calculate_precision_recall Ran
ran on a synthesized input
atapour/ransomware-classification/utils.py
code served (permissive licence) · get_code("5556a80909120efd")
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 numpy as np
import sklearn.metrics as metrics

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

# Create ground truth labels - multi-class classification
# Using 3 classes (0, 1, 2) with 10 samples total
gt = np.array([0, 0, 0, 1, 1, 1, 2, 2, 2, 1])

# Create predictions - some correct, some incorrect
pred = np.array([0, 0, 1, 1, 1, 2, 2, 2, 0, 1])

ARGS = {
    "gt": gt,
    "pred": pred
}
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 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
color_norm Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
seasonyc/densenet/cifar10-test.py
pointer only (licence: NONE) · get_code("ced15e9bed875da8")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, seasonyc/densenet 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 Ran
ran on a synthesized input
GeoTrouvetout/Vehicle_ReID/vehicle_reid/performance.py
code served (permissive licence) · get_code("4fb9f8b3cf01ab1a")
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 numpy as np

# Construct a valid ranking dict for the compute function
# The function expects a dict with a 'tracks' key containing track rankings

ARGS = {
    'ranking': {
        'tracks': {
            'track_1': {
                'class': 0,
                'ranking': {
                    '1': {'class': 0},
                    '2': {'class': 1},
                    '3': {'class': 0},
                    '4': {'class': 2},
                    '5': {'class': 0},
                }
            },
            'track_2': {
                'class': 1,
                'ranking': {
                    '1': {'class': 1},
                    '2': {'class': 0},
                    '3': {'class': 2},
                    '4': {'class': 1},
                    '5': {'class': 1},
                }
            },
            'track_3': {
                'class': 2,
                'ranking': {
                    '1': {'class': 0},
                    '2': {'class': 2},
                    '3': {'class': 1},
                    '4': {'class': 2},
                    '5': {'class': 0},
                }
            }
        }
    }
}
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 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
compute_aggregation Ran
ran on a synthesized input
GeoTrouvetout/Vehicle_ReID/vehicle_reid/ranking.py
code served (permissive licence) · get_code("03405811a26933c0")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned float64, ndim 0, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

# Create a list of sample distances
distances = [0.5, 1.2, 0.8, 1.5, 0.3, 0.9, 1.1, 0.7, 1.3, 0.6]

ARGS = {
    'distances': distances,
    'aggregation': 'mean50'
}
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 float64, ndim 0, dtype family float — no output shape is stated: observe_signature.py recorded float ndim 0, which disagrees with the fixture run
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
compute_metric Ran
ran on a synthesized input
GeoTrouvetout/Vehicle_ReID/vehicle_reid/ranking.py
code served (permissive licence) · get_code("0f3873bb6bde781e")
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

torch.manual_seed(0)

# Create small LR vectors for query and test
query = torch.randn(4)
test = torch.randn(4)

ARGS = {
    'query': query,
    'test': test,
    'metric': 'MCD'
}
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_residual Ran
ran on a synthesized input
GeoTrouvetout/Vehicle_ReID/vehicle_reid/ranking.py
code served (permissive licence) · get_code("6e041ab82714fe0e")
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 numpy as np
from sklearn.decomposition import SparseCoder

torch.manual_seed(0)

# Create small test tensors
# query: shape (1, 4) - represents query features
query = torch.randn(1, 4, dtype=torch.float32)

# test: shape (3, 4) - represents test/dictionary features
# This will be squeezed and used as the dictionary D
test = torch.randn(3, 4, dtype=torch.float32)

ARGS = {
    "query": query,
    "test": test,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, sklearn, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type 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
conv_block Ran
ran on a synthesized input
Sakib1263/DenseNet-1D-2D-Tensorflow-Keras/Codes/DenseNet_1DCNN.py
code served (permissive licence) · get_code("a324e3ecd95d62e4")
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 EagerTensor, ndim 3, dtype family float, shape (2, 4, 16) 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 tensorflow as tf
import numpy as np

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

# Create a small input tensor
# Shape: (batch_size=2, sequence_length=4, input_channels=3)
x = tf.random.normal((2, 4, 3))

# Define parameters for conv_block
num_filters = 16
bottleneck = True

ARGS = {
    'x': x,
    'num_filters': num_filters,
    'bottleneck': bottleneck
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type EagerTensor, ndim 3, dtype family float, shape (2, 4, 16) — shape (2, 4, 16) 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
conv_block Ran
ran on a synthesized input
Sakib1263/DenseNet-1D-2D-Tensorflow-Keras/Codes/DenseNet_2DCNN.py
code served (permissive licence) · get_code("97627c7cc3973ed6")
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 EagerTensor, ndim 4, dtype family float, shape (2, 4, 4, 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 tensorflow as tf
import numpy as np

# Create a small input tensor for the conv_block function
# Shape: (batch_size, height, width, channels)
batch_size = 2
height = 4
width = 4
channels = 3

inputs = tf.random.normal((batch_size, height, width, channels))

# Define the arguments dictionary
ARGS = {
    'x': inputs,
    'num_filters': 8,
    'bottleneck': True
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type EagerTensor, ndim 4, dtype family float, shape (2, 4, 4, 8) — shape (2, 4, 4, 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
conv_block Ran
ran on a synthesized input
lionelmessi6410/ntga/models/densenet.py
code served (permissive licence) · get_code("4e3b1fedf80792c3")
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 EagerTensor, ndim 4, dtype family float, shape (2, 4, 3, 28) 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 tensorflow as tf
import numpy as np

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

# Create a small input tensor for the conv_block function
# Shape: (batch_size, height, width, channels)
# batch=2, height=4, width=3, channels=16
x = tf.random.normal((2, 4, 3, 16))

# growth_rate parameter - controls feature map growth in DenseNet blocks
growth_rate = 12

# name parameter - string prefix for naming layers in the block
name = "conv_block_0"

ARGS = {
    "x": x,
    "growth_rate": growth_rate,
    "name": name
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type EagerTensor, ndim 4, dtype family float, shape (2, 4, 3, 28) — shape (2, 4, 3, 28) 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
dense_block Ran
ran on a synthesized input
lionelmessi6410/ntga/models/densenet.py
code served (permissive licence) · get_code("f22037ad697e2445")
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 EagerTensor, ndim 4, dtype family float, shape (2, 4, 4, 67) 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 tensorflow as tf
import tensorflow.keras.layers as layers
import numpy as np

# Create a small input tensor (batch=2, height=4, width=4, channels=3)
x = tf.random.normal((2, 4, 4, 3))

# Define parameters for dense_block
blocks = 2
name = 'dense_block_1'
growth_rate = 32

ARGS = {
    'x': x,
    'blocks': blocks,
    'name': name,
    'growth_rate': growth_rate
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type EagerTensor, ndim 4, dtype family float, shape (2, 4, 4, 67) — shape (2, 4, 4, 67) 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
densenet121 Ran
ran; our contract draft was wrong, not the code
jiweibo/imagenet/models/densenet.py
code served (permissive licence) · get_code("2cf8e05772f97227")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Sonnet 4.5, prompt python_fixture_v1) and returned DenseNet. 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 os

torch.manual_seed(0)

# The function densenet121 has two parameters:
# 1. pretrained (bool, default False)
# 2. **kwargs (additional keyword arguments passed to DenseNet constructor)

# Since pretrained=True would require loading a file from disk (which we want to avoid),
# we'll set pretrained=False

# The **kwargs can include any parameters that DenseNet.__init__ accepts:
# - growth_rate (default 32)
# - block_config (default (6, 12, 24, 16))
# - num_init_features (default 64)
# - bn_size (default 4)
# - drop_rate (default 0)
# - num_classes (default 1000)

# We can pass some of these as kwargs to test the function

ARGS = {
    'pretrained': False,
    'num_classes': 10,
    'drop_rate': 0.1
}
Setup block drafted by
a language model, Claude Sonnet 4.5, prompt python_fixture_v1
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type DenseNet — a DenseNet output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a DenseNet 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
densenet121 Ran
ran; our contract draft was wrong, not the code
fengjiqiang/pretrainedmodel_pytorch/densenet.py
code served (permissive licence) · get_code("bb3352f2e86fcbb2")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Sonnet 4.5, prompt python_fixture_v1) and returned DenseNet. 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)

# The densenet121 function signature is:
# def densenet121(pretrained=False, **kwargs):

# Looking at the function, it will create a DenseNet model with specific parameters
# The pretrained parameter controls whether to load pretrained weights
# **kwargs are passed to the DenseNet constructor

# For pretrained=False, no model loading happens, so we can safely call it
# For pretrained=True, it would try to download weights, which we want to avoid

# The function doesn't actually use any parameters except pretrained and **kwargs
# which are passed to DenseNet constructor

ARGS = {
    'pretrained': False,
    # We can pass additional kwargs that would go to DenseNet constructor
    # but they're optional, so we can leave them out or include some
}
Setup block drafted by
a language model, Claude Sonnet 4.5, prompt python_fixture_v1
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type DenseNet — a DenseNet output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a DenseNet 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
densenet169 Ran
ran; our contract draft was wrong, not the code
jiweibo/imagenet/models/densenet.py
code served (permissive licence) · get_code("5ed3a0cb0ce1988f")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Sonnet 4.5, prompt python_fixture_v1) and returned DenseNet. 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 os
import re
from collections import OrderedDict

torch.manual_seed(0)

# The function densenet169 has two parameters:
# 1. pretrained (bool, default=False)
# 2. **kwargs (additional keyword arguments passed to DenseNet constructor)

# Since pretrained=True would require loading a file from disk (which we want to avoid),
# we'll set pretrained=False

# For **kwargs, we can pass any valid DenseNet constructor arguments like:
# - growth_rate (default 32)
# - block_config (default (6, 12, 24, 16))
# - num_init_features (default 64)
# - bn_size (default 4)
# - drop_rate (default 0)
# - num_classes (default 1000)

# We'll use default values or simple overrides

ARGS = {
    'pretrained': False,
    # Optional kwargs that could be passed to DenseNet:
    'num_classes': 10,  # Override default 1000 to a smaller value
    'drop_rate': 0.0,
}
Setup block drafted by
a language model, Claude Sonnet 4.5, prompt python_fixture_v1
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type DenseNet — a DenseNet output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a DenseNet 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
densenet169 Ran
ran; our contract draft was wrong, not the code
fengjiqiang/pretrainedmodel_pytorch/densenet.py
code served (permissive licence) · get_code("a5e43271fe68a229")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Sonnet 4.5, prompt python_fixture_v1) and returned DenseNet. 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)

# The function densenet169 takes:
# - pretrained: bool (default False)
# - **kwargs: additional keyword arguments that get passed to DenseNet constructor

# Looking at the function signature, it only has 'pretrained' as explicit parameter
# and **kwargs for any additional DenseNet constructor parameters

# For a minimal setup, we just need to provide the pretrained parameter
# and optionally any kwargs that would be passed to DenseNet

ARGS = {
    'pretrained': False,
    # We can also pass DenseNet constructor kwargs if needed, but they have defaults
    # For example: 'num_classes': 10, 'growth_rate': 32, etc.
}
Setup block drafted by
a language model, Claude Sonnet 4.5, prompt python_fixture_v1
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type DenseNet — a DenseNet output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a DenseNet 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
densenet201 Ran
ran; our contract draft was wrong, not the code
jiweibo/imagenet/models/densenet.py
code served (permissive licence) · get_code("ed4f95207d5e697e")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Sonnet 4.5, prompt python_fixture_v1) and returned DenseNet. 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 os

torch.manual_seed(0)

# The function densenet201 takes two parameters:
# 1. pretrained (bool) - whether to load pretrained weights
# 2. **kwargs - additional keyword arguments passed to DenseNet constructor

# Since pretrained=True would require loading a file from disk (which we want to avoid),
# we'll set pretrained=False

# The **kwargs can include any parameters that DenseNet.__init__ accepts:
# - growth_rate (default 32)
# - block_config (default (6, 12, 24, 16))
# - num_init_features (default 64)
# - bn_size (default 4)
# - drop_rate (default 0)
# - num_classes (default 1000)

# We'll provide minimal valid arguments
ARGS = {
    'pretrained': False,
    # We can optionally pass kwargs, but they're not required
    # Let's pass a simple one to show it works
    'num_classes': 10
}
Setup block drafted by
a language model, Claude Sonnet 4.5, prompt python_fixture_v1
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type DenseNet — a DenseNet output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a DenseNet 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
densenet201 Ran
ran; our contract draft was wrong, not the code
fengjiqiang/pretrainedmodel_pytorch/densenet.py
code served (permissive licence) · get_code("cc30af9ad8ab8701")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Sonnet 4.5, prompt python_fixture_v1) and returned DenseNet. 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)

# The densenet201 function signature is:
# def densenet201(pretrained=False, **kwargs):

# Looking at the function, it would typically:
# 1. Create a DenseNet model with specific config for densenet201
# 2. If pretrained=True, load weights from model_zoo
# 3. Pass any **kwargs to the DenseNet constructor

# Since we need to avoid network downloads, we set pretrained=False
# We can pass additional kwargs that would go to DenseNet.__init__

ARGS = {
    'pretrained': False,
    # Optional kwargs that could be passed to DenseNet constructor:
    # 'num_classes': 10,  # example override
    # 'drop_rate': 0.1,   # example override
}
Setup block drafted by
a language model, Claude Sonnet 4.5, prompt python_fixture_v1
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type DenseNet — a DenseNet output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a DenseNet 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
float_parameter Ran
ran, honoured the contract we drafted
0jason000/DenseNet/src/transforms.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from chenchenzong/dpc, is pointer only (licence: NONE) · get_code("40d3c39a403ae1f7")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from chenchenzong/dpc, 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.

get_aug_str Ran
ran; our contract draft was wrong, not the code
seasonyc/densenet/cifar10-test.py
pointer only (licence: NONE) · get_code("a1f85ee41ec15a37")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, seasonyc/densenet when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

get_train_params_by_name Ran
ran; our contract draft was wrong, not the code
ApexPredator1/DenseNet_tensorflow/run_dense_net.py
code served (permissive licence) · get_code("9c721ee6d67b41e8")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Sonnet 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 torch
import numpy as np

torch.manual_seed(0)

# The function takes a single parameter 'name' and returns a dictionary
# based on the name provided. We need to provide a valid name string.

ARGS = {
    'name': 'C10'
}
Setup block drafted by
a language model, Claude Sonnet 4.5, prompt python_fixture_v1
Imports
numpy, torch — 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
get_train_params_by_name Ran
ran; our contract draft was wrong, not the code
noahfl/densenet-sdr/run_dense_net.py
code served (permissive licence) · get_code("19a94862e2f61a05")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Sonnet 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 torch
import numpy as np

torch.manual_seed(0)

# The function takes a single parameter 'name' and returns a dictionary
# based on the name provided. We need to provide a valid name string.

ARGS = {
    'name': 'C10'
}
Setup block drafted by
a language model, Claude Sonnet 4.5, prompt python_fixture_v1
Imports
numpy, torch — 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
get_train_params_by_name Ran
ran; our contract draft was wrong, not the code
cges60809/DenseNet/run_dense_net.py
code served (permissive licence) · get_code("81af5b85426f6c99")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Sonnet 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 torch
import numpy as np

torch.manual_seed(0)

# The function takes a single parameter 'name' and returns a dictionary
# based on the name provided. We need to test with one of the valid names.

name = 'C10'

ARGS = {
    'name': name
}
Setup block drafted by
a language model, Claude Sonnet 4.5, prompt python_fixture_v1
Imports
numpy, torch — 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
get_train_params_by_name Ran
ran on a synthesized input
RaghavHub/rekuten_hack/run_dense_net.py
code served (permissive licence) · get_code("14f2096a9b9dc093")
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 torch
import torch.nn as nn

torch.manual_seed(0)

ARGS = {
    'name': 'C10'
}
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 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
image_to_tfexample Ran
ran on a synthesized input
pudae/tensorflow-densenet/datasets/dataset_utils.py
code served (permissive licence) · get_code("ef2b2fdbc99cb13a")
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 Example. 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 tensorflow as tf

# The function expects:
# - image_data: bytes (image data)
# - image_format: bytes (image format like b'jpeg')
# - height: int (image height)
# - width: int (image width)
# - class_id: int (class label)

ARGS = {
    'image_data': b'\x89PNG\r\n\x1a\n\x00\x00\x00\rIHDR\x00\x00\x00\x01',
    'image_format': b'jpeg',
    'height': 64,
    'width': 32,
    'class_id': 5
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Example — a Example output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Example 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
init_group_params Ran
ran on a synthesized input
0jason000/DenseNet/src/optim.py
code served (permissive licence) · get_code("698131638ff8d4ad")
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 torch
import torch.nn as nn

torch.manual_seed(0)

class Param:
    def __init__(self, name):
        self.name = name
        self.data = torch.randn(2, 4)

params = [
    Param('weight'),
    Param('bias'),
    Param('beta'),
    Param('gamma'),
    Param('kernel'),
    Param('scale')
]

weight_decay = 0.01

ARGS = {
    'params': params,
    'weight_decay': weight_decay
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 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
int64_feature Ran
ran; our contract draft was wrong, not the code
pudae/tensorflow-densenet/datasets/dataset_utils.py
code served (permissive licence) · get_code("24f10eaa5e2a2113")
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 Feature. 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 tensorflow as tf

# The function int64_feature takes a 'values' parameter
# It can be either a scalar or a list/tuple of values
# Let's test with a list of integers

ARGS = {
    "values": [1, 2, 3, 4, 5]
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Feature — a Feature output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Feature 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
int_parameter Ran
ran, honoured the contract we drafted
0jason000/DenseNet/src/transforms.py
code served (permissive licence) · get_code("e5a370ee6856bd8d")
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 numpy as np

ARGS = {
    "level": 5,
    "maxval": 100
}
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 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
is_model Ran
ran, violated the contract we drafted
0jason000/DenseNet/src/model.py
code served (permissive licence) · get_code("b85a6569fd4f3890")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned bool, ndim 0, dtype family float, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

# Set up the global model entrypoints dictionary with some test models
_model_entrypoints = {
    'resnet18': 'dummy_model_1',
    'vit_base': 'dummy_model_2',
    'bert_small': 'dummy_model_3'
}

# Define arguments for is_model function
ARGS = {
    'model_name': 'resnet18'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type bool, ndim 0, dtype family float, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
make_basic_block_layer Ran
ran; our contract draft was wrong, not the code
lionelmessi6410/ntga/models/residual_block.py
code served (permissive licence) · get_code("c06de04ffb006a10")
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 Sequential. 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 tensorflow as tf

# Set random seed for reproducibility
tf.random.set_seed(0)

# Define ARGS dictionary with valid arguments for make_basic_block_layer
ARGS = {
    'filter_num': 64,
    'blocks': 3,
    'stride': 1
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Sequential — a Sequential output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Sequential 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
make_bottleneck_layer Ran
ran on a synthesized input
lionelmessi6410/ntga/models/residual_block.py
code served (permissive licence) · get_code("f4298d3aa73e6097")
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 Sequential. 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 tensorflow as tf
import numpy as np

# Set random seeds for reproducibility
tf.random.set_seed(0)
np.random.seed(0)

# Define ARGS dictionary with all parameters for make_bottleneck_layer
ARGS = {
    "filter_num": 64,
    "blocks": 3,
    "stride": 1
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Sequential — a Sequential output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Sequential 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
mse Ran
ran on a synthesized input
hycis/TensorGraph/tensorgraph/cost.py
code served (permissive licence) · get_code("bd9a2c9a7ed3dff5")
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 EagerTensor, 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 tensorflow as tf
import numpy as np

# Create small TensorFlow tensors for ytrue and ypred
ytrue = tf.constant([[1.0, 2.0, 3.0, 4.0],
                      [5.0, 6.0, 7.0, 8.0]], dtype=tf.float32)

ypred = tf.constant([[1.1, 2.1, 2.9, 4.2],
                      [4.9, 6.1, 7.1, 7.9]], dtype=tf.float32)

ARGS = {
    'ytrue': ytrue,
    'ypred': ypred
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type EagerTensor, 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
np_load_func Ran
ran on a synthesized input
hycis/TensorGraph/tensorgraph/data_iterator.py
code served (permissive licence) · get_code("abf7602f79476ae0")
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 (2, 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 tempfile
import os

# Create a temporary directory and a temporary numpy file
temp_dir = tempfile.mkdtemp()
temp_file = os.path.join(temp_dir, "test_array.npy")

# Create a small numpy array and save it
test_array = np.array([[1, 2, 3, 4], [5, 6, 7, 8]], dtype=np.float32)
np.save(temp_file, test_array)

ARGS = {
    "path": temp_file
}
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 (2, 4) — shape (2, 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
parse_args Ran
ran; our contract draft was wrong, not the code
ByakuyaUncia/DenseNetUncia/train_eval_image_classifier.py
pointer only (licence: NONE) · get_code("1214ac6284444e74")
How it ran

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

preprocess_config Ran
ran; our contract draft was wrong, not the code
okason97/DenseNet-Tensorflow2/src/scripts/train/run_train.py
code served (permissive licence) · get_code("5145248018144008")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Sonnet 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.

c = {
    'data.batch_size': '32',
    'data.episodes': '100',
    'data.gpu': '0',
    'data.cuda': '1',
    'data.train_size': '0.8',
    'data.test_size': '0.2',
    'data.rotation_range': '15.0',
    'data.width_shift_range': '0.1',
    'data.height_shift_range': '0.1',
    'model.growth_rate': '12',
    'model.reduction': '0.5',
    'model.nb_layers': '6,12,24,16',
    'train.epochs': '50',
    'train.patience': '10',
    'train.lr': '0.001',
    'other.param': 'some_string_value'
}

ARGS = {'c': c}
Setup block drafted by
a language model, Claude Sonnet 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
preprocess_config Ran
ran; our contract draft was wrong, not the code
okason97/DenseNet-Tensorflow2/src/scripts/eval/run_eval.py
code served (permissive licence) · get_code("34795ee835344936")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Sonnet 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.

c = {
    'data.batch_size': '32',
    'data.episodes': '100',
    'data.gpu': '0',
    'data.cuda': '1',
    'model.growth_rate': '12',
    'data.train_size': '0.8',
    'data.test_size': '0.2',
    'data.rotation_range': '15.0',
    'data.width_shift_range': '0.1',
    'data.height_shift_range': '0.1',
    'model.reduction': '0.5',
    'model.nb_layers': '4,8,12',
    'data.dataset': 'cifar10',
    'model.name': 'densenet',
    'optimizer': 'adam'
}

ARGS = {'c': c}
Setup block drafted by
a language model, Claude Sonnet 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
preprocess_input Ran
ran on a synthesized input
titu1994/DenseNet/densenet.py
code served (permissive licence) · get_code("2b2390bcd64e4f4f")
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 4, dtype family float, shape (2, 4, 3, 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
import keras.backend as K

# Set up a small batch of image data
# The function expects a 4D tensor for batch processing or 3D for single image
# Using channels_last format (height, width, channels) or channels_first (channels, height, width)
# Create a small batch: 2 images, 4x3 spatial dims, 3 channels (RGB)

# Using channels_last format (default in TensorFlow backend)
x_channels_last = np.random.rand(2, 4, 3, 3).astype(np.float32) * 255

ARGS = {
    'x': x_channels_last,
    'data_format': 'channels_last'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
keras, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 4, dtype family float, shape (2, 4, 3, 3) — shape (2, 4, 3, 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
rolling Ran
ran on a synthesized input
bamos/densenet.pytorch/plot.py
code served (permissive licence) · get_code("76cc7194ae64e0dc")
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 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

# Analyzing the function:
# - N: int, used as array index and divisor
# - i: array-like, indexed with slice [N-1:]
# - loss: array-like, passed to np.convolve
# - err: array-like, passed to np.convolve

# Create test data
N = 3
i = np.array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9])  # At least N elements
loss = np.array([0.5, 0.4, 0.3, 0.2, 0.1, 0.05, 0.02, 0.01, 0.005, 0.002])
err = np.array([0.9, 0.8, 0.7, 0.6, 0.5, 0.4, 0.3, 0.2, 0.1, 0.05])

ARGS = {
    'N': N,
    'i': i,
    'loss': loss,
    'err': err
}
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 tuple, ndim 2, 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
schedule_fn Ran
ran, honoured the contract we drafted
satishjasthi/Densenet_smplified/cifar10_densenet_classification.py
pointer only (licence: NONE) · get_code("d1a1f72a1675802c")
How it ran

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

stem Ran
ran on a synthesized input
Sakib1263/DenseNet-1D-2D-Tensorflow-Keras/Codes/DenseNet_1DCNN.py
code served (permissive licence) · get_code("050e4107164ed015")
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 KerasTensor. 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 tensorflow as tf
import numpy as np

# Create a small input tensor
# The stem function expects inputs to be a tensor with shape (batch, timesteps, channels)
# Using batch=2, timesteps=4, channels=3 for a tiny valid input
batch_size = 2
timesteps = 4
channels = 3

inputs = tf.keras.Input(shape=(timesteps, channels), batch_size=batch_size)

num_filters = 8

ARGS = {
    'inputs': inputs,
    'num_filters': num_filters,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type KerasTensor — a KerasTensor 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
train_test_split_balanced Ran
ran on a synthesized input
okason97/handshape-recognition/src/utils/model_selection/selection.py
code served (permissive licence) · get_code("4052dd4aed95d311")
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 numpy as np

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

# Create a small balanced dataset
# 2 classes, 6 samples per class = 12 total samples
n_samples_per_class = 6
n_features = 4
n_classes = 2

# Create feature data (data array)
data = np.random.randn(n_samples_per_class * n_classes, n_features).astype(np.float32)

# Create balanced target labels
target = np.repeat(np.arange(n_classes), n_samples_per_class)

ARGS = {
    "data": data,
    "target": target,
    "test_size": 0.2,
    "train_size": 0,
    "n_train_per_class": 0,
    "n_test_per_class": 0,
}
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 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
transition_block Ran
ran on a synthesized input
lionelmessi6410/ntga/models/densenet.py
code served (permissive licence) · get_code("75303155bc2bdb0c")
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 EagerTensor, ndim 4, dtype family float, shape (2, 4, 4, 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 tensorflow as tf
import numpy as np

# Create a small input tensor with shape (batch, height, width, channels)
# batch=2, height=8, width=8, channels=16
x = tf.random.normal((2, 8, 8, 16))

# reduction parameter: controls how much to reduce filters
reduction = 0.5

# name parameter: string prefix for layer naming
name = "transition"

ARGS = {
    "x": x,
    "reduction": reduction,
    "name": name
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type EagerTensor, ndim 4, dtype family float, shape (2, 4, 4, 8) — shape (2, 4, 4, 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
DNN Not yet run
unverified (harvested, no recorded run)
lionelmessi6410/ntga/models/dnn.py
code served (permissive licence) · get_code("6e049069292303d8")
How it ran

no invocation recorded.

FCDenseNet103 Not yet run
unverified (harvested, no recorded run)
bfortuner/pytorch_tiramisu/models/tiramisu.py
code served (permissive licence) · get_code("a31003fad13ae67c")
How it ran

no invocation recorded.

FCDenseNet57 Not yet run
unverified (harvested, no recorded run)
bfortuner/pytorch_tiramisu/models/tiramisu.py
code served (permissive licence) · get_code("f1271d09e9ba1ad7")
How it ran

no invocation recorded.

FCDenseNet67 Not yet run
unverified (harvested, no recorded run)
bfortuner/pytorch_tiramisu/models/tiramisu.py
code served (permissive licence) · get_code("cfa9b6051f647903")
How it ran

no invocation recorded.

adv_loss Not yet run
unverified (harvested, no recorded run)
lionelmessi6410/ntga/generate_attack.py
code served (permissive licence) · get_code("166e542afd9fd8cb")
How it ran

no invocation recorded.

binary_f1 Not yet run
unverified (harvested, no recorded run)
hycis/TensorGraph/tensorgraph/cost.py
code served (permissive licence) · get_code("2db2a94e821fe395")
How it ran

no invocation recorded.

compute Not yet run
unverified (harvested, no recorded run)
GeoTrouvetout/Vehicle_ReID/vehicle_reid/latent_representation.py
code served (permissive licence) · get_code("408652244f294883")
How it ran

no invocation recorded.

densenet121 Not yet run
unverified (harvested, no recorded run)
DaikiTanak/manifold_mixup/densenet_mixup.py
pointer only (licence: NONE) · get_code("5215dafd01cc9248")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, DaikiTanak/manifold_mixup 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. No invocation recorded.

densenet169 Not yet run
unverified (harvested, no recorded run)
DaikiTanak/manifold_mixup/densenet_mixup.py
pointer only (licence: NONE) · get_code("b97695b36002e117")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, DaikiTanak/manifold_mixup 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. No invocation recorded.

densenet201 Not yet run
unverified (harvested, no recorded run)
DaikiTanak/manifold_mixup/densenet_mixup.py
pointer only (licence: NONE) · get_code("0ca25406530d858d")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, DaikiTanak/manifold_mixup 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. No invocation recorded.

entropy Not yet run
unverified (harvested, no recorded run)
hycis/TensorGraph/tensorgraph/cost.py
code served (permissive licence) · get_code("3489cf036722364c")
How it ran

no invocation recorded.

get_data_provider_by_name Not yet run
unverified (harvested, no recorded run)
RaghavHub/rekuten_hack/data_providers/utils.py
code served (permissive licence) · get_code("a2a43a28044c501d")
How it ran

no invocation recorded.

Showing the first 60. An agent gets every one from get_harvested_code_for_paper("1608.06993").

Also present in other papers' code

Of 71 samples lifted from this paper's linked repositories, 71 carry a function identity and 0 do not. Of the 70 distinct functions with an identity, 21 are also present in other papers whose code we hold (8 with at least one paper dated earlier, 0 the same month, 21 later) and 49 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
is_model 22 other papers
2 dated earlier, 20 later
1409.4842 (2014-09)
Going deeper with convolutions
1512.00567 (2015-12)
Rethinking the Inception Architecture for Computer Vision
1911.11907 (2019-11)
GhostNet: More Features from Cheap Operations
and 19 more not named here
float_parameter 15 other papers
2 dated earlier, 13 later
1409.4842 (2014-09)
Going deeper with convolutions
1512.00567 (2015-12)
Rethinking the Inception Architecture for Computer Vision
1904.12848 (2019-04)
Unsupervised Data Augmentation for Consistency Training
and 12 more not named here
int_parameter 8 other papers
2 dated earlier, 6 later
1409.4842 (2014-09)
Going deeper with convolutions
1512.00567 (2015-12)
Rethinking the Inception Architecture for Computer Vision
1911.11907 (2019-11)
GhostNet: More Features from Cheap Operations
and 5 more not named here
densenet121 6 other papers
6 later
1812.00332 (2018-12)
ProxylessNAS: Direct Neural Architecture Search on Target Task and Hardware
1904.03441 (2019-04)
Iterative Normalization: Beyond Standardization towards Efficient Whitening
1908.03706 (2019-08)
Exploiting temporal consistency for real-time video depth estimation
and 3 more not named here
densenet169 6 other papers
6 later
1812.00332 (2018-12)
ProxylessNAS: Direct Neural Architecture Search on Target Task and Hardware
1904.03441 (2019-04)
Iterative Normalization: Beyond Standardization towards Efficient Whitening
1908.03706 (2019-08)
Exploiting temporal consistency for real-time video depth estimation
and 3 more not named here
densenet201 6 other papers
6 later
1812.00332 (2018-12)
ProxylessNAS: Direct Neural Architecture Search on Target Task and Hardware
1904.03441 (2019-04)
Iterative Normalization: Beyond Standardization towards Efficient Whitening
1908.03706 (2019-08)
Exploiting temporal consistency for real-time video depth estimation
and 3 more not named here
init_group_params 5 other papers
2 dated earlier, 3 later
1409.4842 (2014-09)
Going deeper with convolutions
1512.00567 (2015-12)
Rethinking the Inception Architecture for Computer Vision
1911.11907 (2019-11)
GhostNet: More Features from Cheap Operations
and 2 more not named here
model_entrypoint 5 other papers
2 dated earlier, 3 later
1409.4842 (2014-09)
Going deeper with convolutions
1512.00567 (2015-12)
Rethinking the Inception Architecture for Computer Vision
1911.11907 (2019-11)
GhostNet: More Features from Cheap Operations
and 2 more not named here
optimize_clones 4 other papers
4 later
1701.06487 (2017-01)
Dirty Pixels: Towards End-to-End Image Processing and Perception
1804.10745 (2018-04)
Generalizing Across Domains via Cross-Gradient Training
1807.06521 (2018-07)
CBAM: Convolutional Block Attention Module
and 1 more not named here
Conv_1D_Block 3 other papers
2 dated earlier, 1 later · 1 only as an assembly brick
1502.03167 (2015-02)
Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
1512.03385 (2015-12)
Deep Residual Learning for Image Recognition
1611.05431 (2016-11)
Aggregated Residual Transformations for Deep Neural Networks
int64_feature 2 other papers
2 later
1904.07850 (2019-04)
Objects as Points
2106.09309 (2021-06)
Layer Folding: Neural Network Depth Reduction using Activation Linearization
make_basic_block_layer 2 other papers
1 dated earlier, 1 later · 1 only as an assembly brick
1512.03385 (2015-12)
Deep Residual Learning for Image Recognition
2006.07733 (2020-06)
Bootstrap your own latent: A new approach to self-supervised Learning
preprocess_image 2 other papers
2 later
1905.02231 (2019-05)
A Geometric Approach to Obtain a Bird's Eye View from an Image
1908.07410 (2019-08)
ViSiL: Fine-grained Spatio-Temporal Video Similarity Learning
rolling 2 other papers
1 dated earlier, 1 later
1606.04797 (2016-06)
V-Net: Fully Convolutional Neural Networks for Volumetric Medical Image Segmentation
1703.00443 (2017-03)
OptNet: Differentiable Optimization as a Layer in Neural Networks
FCDenseNet103 1 other paper
1 later
1611.09326 (2016-11)
The One Hundred Layers Tiramisu: Fully Convolutional DenseNets for Semantic Segmentation
FCDenseNet57 1 other paper
1 later
1611.09326 (2016-11)
The One Hundred Layers Tiramisu: Fully Convolutional DenseNets for Semantic Segmentation
FCDenseNet67 1 other paper
1 later
1611.09326 (2016-11)
The One Hundred Layers Tiramisu: Fully Convolutional DenseNets for Semantic Segmentation
densenet121 1 other paper
1 later
1806.05236 (2018-06)
Manifold Mixup: Better Representations by Interpolating Hidden States
densenet169 1 other paper
1 later
1806.05236 (2018-06)
Manifold Mixup: Better Representations by Interpolating Hidden States
densenet201 1 other paper
1 later
1806.05236 (2018-06)
Manifold Mixup: Better Representations by Interpolating Hidden States

Shown: the 20 functions present in the most other papers. 1 more shared functions are not listed here.

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

More are linked than shown here.

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

Abstract

Recent work has shown that convolutional networks can be substantially deeper, more accurate, and efficient to train if they contain shorter connections between layers close to the input and those close to the output. In this paper, we embrace this observation and introduce the Dense Convolutional Network (DenseNet), which connects each layer to every other layer in a feed-forward fashion. Whereas traditional convolutional networks with L layers have L connections-one between each layer and its subsequent layer-our network has L(L+1) 2 direct connections. For each layer, the feature-maps of all preceding layers are used as inputs, and its own feature-maps are used as inputs into all subsequent layers. DenseNets have several compelling advantages: they alleviate the vanishing-gradient problem, strengthen feature propagation, encourage feature reuse, and substantially reduce the number of parameters. We evaluate our proposed architecture on four highly competitive object recognition benchmark tasks (CIFAR-10, CIFAR-100, SVHN, and ImageNet). DenseNets obtain significant improvements over the state-of-the-art on most of them, whilst requiring less computation to achieve high performance. Code and pre-trained models are available at https://github.com/liuzhuang13/DenseNet.

For agents

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

get_harvested_code_for_paper("1608.06993")
get_code_for_paper("1608.06993")
have("1608.06993")

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

curl https://syntology.ai/api/ran/1608.06993.json

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

[![Syntology run record](https://syntology.ai/api/ran/1608.06993.svg)](https://syntology.ai/paper/1608.06993)

Connect an agent — have() is free.