SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2006.10738 · NeurIPS · 2020

Differentiable Augmentation for Data-Efficient GAN Training

Zhijian Liu, Ji Lin, Shengyu Zhao, Yan Zhu, Han Song, Mit

arXiv · PDF · Open in the Atlas

Code that ran

49 of 58 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 12 have no contract verdict; 33 failures of our instrument; 9 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/2006.10738.json · badge: /api/ran/2006.10738.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
mit-han-lab/data-efficient-gans canonical (classifier label)
classifier basis: classifier_2025
7 ran · 0 no recorded run
uzielroy/StyleGan_FewShot community (classifier label)
classifier basis not recorded on the link
10 ran · 4 no recorded run
milmor/LadaGAN-pytorch community (classifier label)
classifier basis not recorded on the link
8 ran · 2 no recorded run
eps696/stylegan2 community
Papers with Code archive
6 ran · 0 no recorded run
gaborvecsei/SLE-GAN community
Papers with Code archive
6 ran · 0 no recorded run
claim-berlin/3d_stylegan_circle_of_willis community
Papers with Code archive
4 ran · 1 no recorded run
milmor/ladagan community (classifier label)
classifier basis not recorded on the link
2 ran · 2 no recorded run
VITA-Group/Ultra-Data-Efficient-GAN-Training community
Papers with Code archive
2 ran · 0 no recorded run
milmor/TransGAN community
Papers with Code archive
2 ran · 0 no recorded run
milmor/self-supervised-gan community
Papers with Code archive
1 ran · 0 no recorded run
POSTECH-CVLab/PyTorch-StudioGAN community
Papers with Code archive
1 ran · 0 no recorded run
FunctionStatusWhere it lives
DiffAugment Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
mit-han-lab/data-efficient-gans/DiffAugment_pytorch.py
code served (permissive licence) · get_code("b0bbabd6de30f612")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1) and returned Tensor, ndim 4, dtype family float, shape (2, 3, 4, 5) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# Create a small image tensor: (batch_size=2, channels=3, height=4, width=5)
x = torch.randn(2, 3, 4, 5, dtype=torch.float32)

ARGS = {
    'x': x,
    'policy': 'color,translation,cutout',
    'channels_first': True
}
Setup block drafted by
a language model, Claude Haiku 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 Tensor, ndim 4, dtype family float, shape (2, 3, 4, 5) — shape (2, 3, 4, 5) 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
DiffAugment Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
mit-han-lab/data-efficient-gans/DiffAugment_pytorch.py
code served (permissive licence) · get_code("5a98bf7c8a33d10d")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1) and returned Tensor, ndim 4, dtype family float, shape (2, 3, 4, 5) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# Create a small batch of image tensors
# Shape: (batch_size, channels, height, width)
x = torch.randn(2, 3, 4, 5)

# Define arguments for DiffAugment function
ARGS = {
    'x': x,
    'policy': 'color,translation,cutout',
    'channels_first': True
}
Setup block drafted by
a language model, Claude Haiku 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 Tensor, ndim 4, dtype family float, shape (2, 3, 4, 5) — shape (2, 3, 4, 5) 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
DiffAugment Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
VITA-Group/Ultra-Data-Efficient-GAN-Training/BigGAN and DiffAugGAN/utils/diff_aug.py
code served (permissive licence) · get_code("b13ee69651d4b880")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1) and returned Tensor, ndim 4, dtype family float, shape (2, 3, 4, 5) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# Create a small batch of image tensors
# Shape: (batch_size, channels, height, width)
x = torch.randn(2, 3, 4, 5)

# Define arguments for DiffAugment function
ARGS = {
    'x': x,
    'policy': 'color,translation,cutout',
    'channels_first': True
}
Setup block drafted by
a language model, Claude Haiku 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 Tensor, ndim 4, dtype family float, shape (2, 3, 4, 5) — shape (2, 3, 4, 5) 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
DiffAugment Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
milmor/self-supervised-gan/diffaug.py
code served (permissive licence) · get_code("106e066ecaa411cb")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1) and returned EagerTensor, 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 tensorflow as tf
import numpy as np

# Create a small batch of image tensors
# Shape: (batch_size, height, width, channels)
batch_size = 2
height = 4
width = 3
channels = 3

# Create random image data in range [0, 1]
x = tf.constant(np.random.rand(batch_size, height, width, channels).astype(np.float32))

# Define ARGS dict with all parameters for DiffAugment
ARGS = {
    'x': x,
    'policy': 'color,translation,cutout',
    'channels_first': False
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
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, 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
DiffAugment Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
eps696/stylegan2/src/training/DiffAugment_tf.py
pointer only (licence: NOASSERTION) · get_code("719b769b10895fa0")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, eps696/stylegan2 (NOASSERTION). 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.

DiffAugment Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
milmor/TransGAN/diffaug.py
code served (permissive licence) · get_code("9f13704fe37f3a7f")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1) and returned EagerTensor, 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 tensorflow as tf
import numpy as np

# Create a small batch of image tensors
# Shape: (batch_size, height, width, channels)
batch_size = 2
height = 4
width = 3
channels = 3

# Create random image data in range [0, 1]
x = tf.constant(np.random.rand(batch_size, height, width, channels).astype(np.float32))

ARGS = {
    'x': x,
    'policy': 'color,translation,cutout',
    'channels_first': False
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
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, 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
DiffAugment Ran
ran; our contract draft was wrong, not the code
claim-berlin/3d_stylegan_circle_of_willis/diff_augment.py
pointer only (licence: NOASSERTION) · get_code("aff3959c75a4b707")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, claim-berlin/3d_stylegan_circle_of_willis (NOASSERTION). 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.

apply_diffaug Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
POSTECH-CVLab/PyTorch-StudioGAN/src/utils/diffaug.py
pointer only (licence: NOASSERTION) · get_code("11fe0ba0c9c1c2ba")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, POSTECH-CVLab/PyTorch-StudioGAN (NOASSERTION). 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.

calc_distances Ran
ran on a synthesized input
uzielroy/StyleGan_FewShot/save_images_monet.py
code served (permissive licence) · get_code("591df4f8315785cd")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned ndarray, ndim 1, dtype family float, shape (2,) 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

torch.manual_seed(0)

# Create small numpy arrays for the function parameters
# p0: a single point (1D array)
# points: multiple points (2D array where each row is a point)

p0 = np.array([1.0, 2.0, 3.0, 4.0])
points = np.array([
    [1.5, 2.5, 3.5, 4.5],
    [0.5, 1.5, 2.5, 3.5]
])

ARGS = {
    "p0": p0,
    "points": points
}
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 ndarray, ndim 1, dtype family float, shape (2,) — shape (2,) 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_frechet_distance Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
uzielroy/StyleGan_FewShot/metric/fid_score.py
pointer only (licence: MIT); the stored copy get_code serves, from shpda/text-to-image-gen-model-acgan, is pointer only (licence: NONE) · get_code("4cee7c1861294d59")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from shpda/text-to-image-gen-model-acgan, 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.

calculate_frechet_distance Ran
ran on a synthesized input
milmor/ladagan/tensorflow/fid.py
code served (permissive licence) · get_code("7bd28928af1bae66")
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
from scipy import linalg
import warnings

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

# Create small but distinct dimensions for mu and sigma
dim = 4

# Generate sample mean vectors (1D arrays)
mu1 = np.random.randn(dim)
mu2 = np.random.randn(dim)

# Generate sample covariance matrices (2D arrays)
# Create positive semi-definite matrices by A @ A.T
A1 = np.random.randn(dim, dim)
sigma1 = A1 @ A1.T

A2 = np.random.randn(dim, dim)
sigma2 = A2 @ A2.T

# Build the arguments dictionary
ARGS = {
    'mu1': mu1,
    'sigma1': sigma1,
    'mu2': mu2,
    'sigma2': sigma2,
    'eps': 1e-06
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, scipy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type 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
create_loader Ran
ran on a synthesized input
milmor/LadaGAN-pytorch/image_datasets.py
code served (permissive licence) · get_code("0c5eecc2f5f4579c")
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 _MultiProcessingDataLoaderIter. 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
import tempfile
import os
from PIL import Image

torch.manual_seed(0)
np.random.seed(0)

# Create a temporary directory with some test images
temp_dir = tempfile.mkdtemp()

# Create a few small test images
for i in range(3):
    img = Image.new('RGB', (8, 8), color=(i*50, i*60, i*70))
    img.save(os.path.join(temp_dir, f'test_image_{i}.jpg'))

ARGS = {
    'data_dir': temp_dir,
    'img_size': 4,
    'batch_size': 2,
    'workers': 0,
    'pin_memory': False
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
PIL, numpy, torch, torchvision — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type _MultiProcessingDataLoaderIter — a _MultiProcessingDataLoaderIter output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a _MultiProcessingDataLoaderIter 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
d_logistic_loss Ran
ran on a synthesized input
milmor/LadaGAN-pytorch/utils.py
code served (permissive licence) · get_code("dfe4839c7c03b7fa")
How it ran

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

import torch
import torch.nn.functional as F

torch.manual_seed(0)

# Construct small tensors for real_pred and fake_pred
# The function applies F.softplus and then takes mean()
# So we need 2D or higher tensors that support these operations
real_pred = torch.randn(2, 4)
fake_pred = torch.randn(2, 4)

ARGS = {
    "real_pred": real_pred,
    "fake_pred": fake_pred,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 0, dtype family float, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
d_r1_loss Ran
ran; our contract draft was wrong, not the code
milmor/LadaGAN-pytorch/utils.py
code served (permissive licence) · get_code("0143fef228211c97")
How it ran

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

import torch
from torch import autograd

torch.manual_seed(0)

# real_pred needs to:
# - support .sum()
# - be differentiable w.r.t. real_img
# - have a shape that allows reshape(shape[0], -1)

# real_img needs to:
# - be a tensor with requires_grad=True
# - support autograd.grad with create_graph=True

batch_size = 2
spatial_dims = 4

real_img = torch.randn(batch_size, 3, spatial_dims, spatial_dims, requires_grad=True)

# Create real_pred by applying some operation to real_img so it's differentiable
real_pred = (real_img ** 2).sum(dim=[2, 3])  # Shape: (batch_size, 3)

ARGS = {
    "real_pred": real_pred,
    "real_img": real_img,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 0, dtype family float, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
denormalize Ran
ran on a synthesized input
milmor/LadaGAN-pytorch/utils.py
code served (permissive licence) · get_code("05e9b8206464518b")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 2, dtype family float, shape (2, 4) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import numpy as np

torch.manual_seed(0)

# The function denormalize takes an image parameter
# It performs arithmetic operations: (image + 1.0) / 2.0
# This suggests image should be a tensor or array-like object

# Create a small normalized image tensor (values in range [-1, 1])
# Using shape (2, 4) for batch of 2 images with 4 channels
image = torch.randn(2, 4)

ARGS = {
    "image": image
}
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 Tensor, 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
diff_augment Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
gaborvecsei/SLE-GAN/sle_gan/diff_augment.py
code served (permissive licence) · get_code("4e9a3e9a819d11b0")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1) and returned EagerTensor, ndim 4, 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 tensorflow as tf
import numpy as np

# Create a small batch of image tensors
# Shape: [batch_size, height, width, channels]
batch_size = 2
height = 4
width = 3
channels = 3

# Create random image data in range [0, 1]
x = tf.constant(np.random.rand(batch_size, height, width, channels).astype(np.float32))

# Define ARGS dictionary with all parameters for diff_augment
ARGS = {
    'x': x,
    'policy': 'color,translation,cutout',
    'channels_first': False
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
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 — no output shape is stated: observe_signature.py could not observe it (import:TypeError)
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
equal_lr Ran
ran; our contract draft was wrong, not the code
uzielroy/StyleGan_FewShot/model.py
code served (permissive licence) · get_code("5ef2eff36d9c9c0e")
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 SimpleModule. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn
from math import sqrt

torch.manual_seed(0)

# Create a simple module with a weight parameter
# The function accesses module._parameters[name] and uses getattr/setattr
# It also calls weight.data.size(1) and weight.data[0][0].numel()
# So weight needs to be at least 2D, and weight[0][0] needs to have numel()

class SimpleModule(nn.Module):
    def __init__(self):
        super().__init__()
        # Create a weight parameter with shape (out_features, in_features)
        # For a Conv2d-like weight, it would be (out_channels, in_channels, kernel_h, kernel_w)
        # Let's use a 4D weight to match typical conv layers: (2, 3, 2, 2)
        # This way: size(1) = 3 (in_channels), and [0][0] is a 2x2 tensor with numel() = 4
        self.weight = nn.Parameter(torch.randn(2, 3, 2, 2))
    
    def forward(self, x):
        return x

module = SimpleModule()

ARGS = {
    'module': module,
    'name': 'weight'
}
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 SimpleModule — a SimpleModule output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a SimpleModule 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_gaussian_kernel Ran
ran, honoured the contract we drafted
uzielroy/StyleGan_FewShot/metric/swd_score.py
code served (permissive licence) · get_code("c8ceba12b6020549")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 4, dtype family float, shape (1, 1, 5, 5) 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 torch

torch.manual_seed(0)

ARGS = {
    'device': 'cpu'
}
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 Tensor, ndim 4, dtype family float, shape (1, 1, 5, 5) — shape (1, 1, 5, 5) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
get_map Ran
ran on a synthesized input
milmor/ladagan/tensorflow/plot_utils.py
code served (permissive licence) · get_code("60afeee00005b62b")
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, 1) 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 math

# The function accesses maps[r][:, h] and maps[r][0][0].shape[0]
# So maps needs to be a dict where maps[r] is a 3D tensor
# maps[r][0][0] should have a shape where shape[0] is a perfect square
# maps[r][:, h] should be indexable by batch and head dimensions

# Create a simple dict structure for maps
# maps[resolution][batch, head, features]
# We need maps[r][0][0].shape[0] to be a perfect square
# Let's make it 16 (4x4), so map_size = 4

batch_size = 2
num_heads = 3
feature_dim = 16  # Perfect square: sqrt(16) = 4
resolution = "high"

# Create the maps structure
# maps[r] should be shape [batch, num_heads, feature_dim]
maps = {
    resolution: tf.constant([
        [[1.0]*feature_dim, [2.0]*feature_dim, [3.0]*feature_dim],
        [[4.0]*feature_dim, [5.0]*feature_dim, [6.0]*feature_dim]
    ], dtype=tf.float32)
}

head = 1
resolution_param = resolution

ARGS = {
    "maps": maps,
    "resolution": resolution_param,
    "head": head
}
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 EagerTensor, ndim 4, dtype family float, shape (2, 4, 4, 1) — shape (2, 4, 4, 1) 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
graipher Ran
ran on a synthesized input
uzielroy/StyleGan_FewShot/save_images_monet.py
code served (permissive licence) · get_code("e176b3170e9eb7f9")
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 set of 2D points (pts parameter)
pts = np.array([
    [0.0, 0.0],
    [1.0, 1.0],
    [2.0, 0.5],
    [0.5, 2.0],
    [3.0, 3.0],
    [1.5, 1.5]
], dtype=np.float32)

# Set K to a small number less than the number of points
K = 3

ARGS = {
    "pts": pts,
    "K": K
}
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
l2_reg Ran
ran on a synthesized input
uzielroy/StyleGan_FewShot/finetune.py
code served (permissive licence) · get_code("3d111932e00fa877")
How it ran

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

import torch
import torch.nn as nn

torch.manual_seed(0)

class SimpleNet(nn.Module):
    def __init__(self, in_features, out_features):
        super().__init__()
        self.linear1 = nn.Linear(in_features, out_features)
        self.linear2 = nn.Linear(out_features, out_features)
    
    def forward(self, x):
        x = self.linear1(x)
        x = self.linear2(x)
        return x

net_src = SimpleNet(4, 3)
net_tgt = SimpleNet(4, 3)

ARGS = {
    "net_src": net_src,
    "net_tgt": net_tgt,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 0, dtype family float, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
list_image_files Ran
ran on a synthesized input
milmor/LadaGAN-pytorch/image_datasets.py
code served (permissive licence) · get_code("9b9d7e8b8644417b")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned list, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import os
import tempfile
import shutil

# Create a temporary directory with some test image files
temp_dir = tempfile.mkdtemp()

# Create some test files with image extensions
test_files = [
    'image1.jpg',
    'image2.jpeg',
    'image3.png',
    'image4.gif',
    'document.pdf',
    'readme.txt',
    'image5.JPG',  # uppercase extension
    'photo.PNG',
]

for filename in test_files:
    filepath = os.path.join(temp_dir, filename)
    with open(filepath, 'w') as f:
        f.write('dummy content')

# Define ARGS with the temporary directory
ARGS = {
    'data_dir': temp_dir
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 1, dtype family float — a list output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
pixel_upsample Ran
ran on a synthesized input
milmor/LadaGAN-pytorch/model.py
code served (permissive licence) · get_code("79d4e2efdb8db1f5")
How it ran

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

import torch
import torch.nn as nn

torch.manual_seed(0)

# The function expects:
# x: a tensor of shape (B, N, C) where N = H * W
# H: height dimension
# W: width dimension

B = 2
H = 4
W = 3
C = 8
N = H * W

# Create input tensor of shape (B, N, C)
x = torch.randn(B, N, C)

ARGS = {
    "x": x,
    "H": H,
    "W": W,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple — a tuple output has no array shape of its own; none is stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
polynomial_mmd Ran
ran on a synthesized input
uzielroy/StyleGan_FewShot/metric/kid_score.py
code served (permissive licence) · get_code("8fdb950a259799fe")
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

# Create small sample data for codes_r and codes_g
# Using distinct dimensions: 5 samples, 3 features
np.random.seed(0)
codes_r = np.random.randn(5, 3).astype(np.float32)
codes_g = np.random.randn(5, 3).astype(np.float32)

ARGS = {
    'codes_r': codes_r,
    'codes_g': codes_g,
    'degree': 3,
    'gamma': None,
    'coef0': 1,
    'var_at_m': None,
    'ret_var': True
}
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
polynomial_mmd_averages Ran
ran on a synthesized input
uzielroy/StyleGan_FewShot/metric/kid_score.py
code served (permissive licence) · get_code("afe497bbaa6085f4")
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

np.random.seed(0)

# Create small sample arrays for codes_r and codes_g
# These should be 2D arrays where each row is a code vector
codes_r = np.random.randn(50, 8).astype(np.float32)
codes_g = np.random.randn(50, 8).astype(np.float32)

ARGS = {
    'codes_r': codes_r,
    'codes_g': codes_g,
    'n_subsets': 10,
    'subset_size': 20,
    'ret_var': True,
    'replace': False,
}
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 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
pyramid_down Ran
ran; our contract draft was wrong, not the code
uzielroy/StyleGan_FewShot/metric/swd_score.py
code served (permissive licence) · get_code("4a3a50d5b90d9dac")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 4, dtype family float, shape (2, 3, 2, 2) 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

torch.manual_seed(0)

# Create a small 3-channel image tensor (batch=2, channels=3, height=4, width=4)
image = torch.randn(2, 3, 4, 4, dtype=torch.float32)

ARGS = {
    'image': image,
    'device': 'cpu'
}
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 Tensor, ndim 4, dtype family float, shape (2, 3, 2, 2) — shape (2, 3, 2, 2) 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
pyramid_up Ran
ran; our contract draft was wrong, not the code
uzielroy/StyleGan_FewShot/metric/swd_score.py
code served (permissive licence) · get_code("d6f847810616ed82")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 4, dtype family float, shape (2, 3, 8, 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 torch
import torch.nn.functional as F
import numpy as np

torch.manual_seed(0)

# Create a small 3-channel image tensor: batch_size=2, channels=3, height=4, width=4
image = torch.randn(2, 3, 4, 4, dtype=torch.float32)

ARGS = {
    'image': image,
    'device': 'cpu'
}
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 Tensor, ndim 4, dtype family float, shape (2, 3, 8, 8) — shape (2, 3, 8, 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
rand_brightness Ran
ran; our contract draft was wrong, not the code
mit-han-lab/data-efficient-gans/DiffAugment_pytorch.py
code served (permissive licence) · get_code("fd66a45971195f9a")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1) and returned Tensor, ndim 4, dtype family float, shape (2, 4, 3, 5) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# Create a small tensor with batch size 2, channels 4, height 3, width 5
x = torch.randn(2, 4, 3, 5, dtype=torch.float32)

ARGS = {
    "x": x
}
Setup block drafted by
a language model, Claude Haiku 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 Tensor, ndim 4, dtype family float, shape (2, 4, 3, 5) — shape (2, 4, 3, 5) 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
rand_brightness Ran
ran; our contract draft was wrong, not the code
milmor/LadaGAN-pytorch/diffaug.py
pointer only (licence: MIT); the stored copy get_code serves, from postech-cvlab/pytorch-studiogan, is pointer only (licence: NOASSERTION) · get_code("011230b2b9b8fb6f")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from postech-cvlab/pytorch-studiogan, and GitHub could not classify the licence of that repository (NOASSERTION). 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.

rand_brightness Ran
ran, honoured the contract we drafted
gaborvecsei/SLE-GAN/sle_gan/diff_augment.py
code served (permissive licence) · get_code("c0fd34aac0b4d83d")
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; 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 tensorflow as tf
import numpy as np

# Create a small test image batch
# Shape: [batch_size, height, width, channels]
# Using batch=2, height=4, width=3, channels=1 for distinct dimensions
x = tf.constant(np.random.rand(2, 4, 3, 1).astype(np.float32))

ARGS = {
    "x": x
}
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 — no output shape is stated: observe_signature.py could not observe it (import:TypeError)
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
rand_brightness Ran
ran, honoured the contract we drafted
eps696/stylegan2/src/training/DiffAugment_tf.py
pointer only (licence: NOASSERTION) · get_code("23b9e86dc02652cc")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, eps696/stylegan2 (NOASSERTION). 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.

rand_brightness Ran
ran; our contract draft was wrong, not the code
claim-berlin/3d_stylegan_circle_of_willis/diff_augment.py
pointer only (licence: NOASSERTION) · get_code("0054342024c45e46")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, claim-berlin/3d_stylegan_circle_of_willis (NOASSERTION). 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.

rand_contrast Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
mit-han-lab/data-efficient-gans/DiffAugment_pytorch.py
code served (permissive licence) · get_code("cf3f99cfbe8f451a")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1) and returned Tensor, ndim 4, dtype family float, shape (2, 3, 4, 5) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# The function rand_contrast takes a tensor x and applies random contrast adjustment
# It expects x to be a 4D tensor (batch, channels, height, width)
# The function:
# 1. Computes mean across spatial dimensions (dims 1,2,3)
# 2. Normalizes by subtracting mean
# 3. Scales by random factor (0.5 to 1.5)
# 4. Adds mean back

# Create a small 4D tensor: (batch=2, channels=3, height=4, width=5)
x = torch.randn(2, 3, 4, 5, dtype=torch.float32)

ARGS = {
    "x": x
}
Setup block drafted by
a language model, Claude Haiku 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 Tensor, ndim 4, dtype family float, shape (2, 3, 4, 5) — shape (2, 3, 4, 5) 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
rand_contrast Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
gaborvecsei/SLE-GAN/sle_gan/diff_augment.py
code served (permissive licence) · get_code("1ea606a1b578d9fd")
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; 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 tensorflow as tf
import numpy as np

# Create a small test tensor with shape (batch=2, height=4, width=3, channels=1)
# The function expects a 4D tensor (batch, height, width, channels)
x = tf.constant(np.random.randn(2, 4, 3, 1).astype(np.float32))

ARGS = {
    "x": x
}
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 — no output shape is stated: observe_signature.py could not observe it (import:TypeError)
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
rand_contrast Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
eps696/stylegan2/src/training/DiffAugment_tf.py
pointer only (licence: NOASSERTION) · get_code("7668b11c66661278")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, eps696/stylegan2 (NOASSERTION). 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.

rand_contrast Ran
ran; our contract draft was wrong, not the code
claim-berlin/3d_stylegan_circle_of_willis/diff_augment.py
pointer only (licence: NOASSERTION) · get_code("93a3d74d32b20296")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, claim-berlin/3d_stylegan_circle_of_willis (NOASSERTION). 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.

rand_cutout Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
mit-han-lab/data-efficient-gans/DiffAugment_pytorch.py
code served (permissive licence) · get_code("8a8f1bb31f56eb41")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1) and returned Tensor, ndim 4, dtype family float, shape (2, 3, 4, 5) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# Create a small input tensor: (batch_size=2, channels=3, height=4, width=5)
x = torch.randn(2, 3, 4, 5)

ARGS = {
    "x": x,
    "ratio": 0.5
}
Setup block drafted by
a language model, Claude Haiku 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 Tensor, ndim 4, dtype family float, shape (2, 3, 4, 5) — shape (2, 3, 4, 5) 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
rand_cutout Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
gaborvecsei/SLE-GAN/sle_gan/diff_augment.py
code served (permissive licence) · get_code("7eedee91371852fd")
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; 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 tensorflow as tf

# Create a small batch of images
# Shape: (batch_size, height, width, channels)
batch_size = 2
height = 4
width = 3
channels = 3

x = tf.random.normal([batch_size, height, width, channels])

# Define arguments dict
ARGS = {
    'x': x,
    'ratio': 0.5
}
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 EagerTensor, ndim 4, dtype family float — no output shape is stated: observe_signature.py could not observe it (import:TypeError)
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
rand_cutout Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
eps696/stylegan2/src/training/DiffAugment_tf.py
pointer only (licence: NOASSERTION) · get_code("d3f6dbbf7ed4b605")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, eps696/stylegan2 (NOASSERTION). 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.

rand_cutout Ran
ran, violated the contract we drafted
milmor/TransGAN/diffaug.py
code served (permissive licence) · get_code("d2d7ad30f29efa00")
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, 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 tensorflow as tf
import numpy as np

# Create a small batch of images: batch_size=2, height=4, width=3, channels=3
# Using float32 to match TensorFlow operations
x = tf.constant(np.random.rand(2, 4, 3, 3).astype(np.float32))

ARGS = {
    "x": x,
    "ratio": 0.5
}
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, 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
rand_saturation Ran
ran; our contract draft was wrong, not the code
mit-han-lab/data-efficient-gans/DiffAugment_pytorch.py
code served (permissive licence) · get_code("abafb0afcfb5d028")
How it ran

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

torch.manual_seed(0)

# Construct a small tensor for the function
# The function expects x to have at least 2 dimensions (uses dim=1 for mean)
# and creates random noise with shape (batch_size, 1, 1, 1)
# Let's create a 4D tensor: (batch=2, channels=4, height=3, width=3)
x = torch.randn(2, 4, 3, 3, dtype=torch.float32)

ARGS = {
    "x": x
}
Setup block drafted by
a language model, Claude Haiku 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 Tensor, 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
rand_saturation Ran
ran; our contract draft was wrong, not the code
milmor/LadaGAN-pytorch/diffaug.py
pointer only (licence: MIT); the stored copy get_code serves, from postech-cvlab/pytorch-studiogan, is pointer only (licence: NOASSERTION) · get_code("5b0d8787e670fc63")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from postech-cvlab/pytorch-studiogan, and GitHub could not classify the licence of that repository (NOASSERTION). 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.

rand_saturation Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
gaborvecsei/SLE-GAN/sle_gan/diff_augment.py
code served (permissive licence) · get_code("fe6b745a5727de14")
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; 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 tensorflow as tf
import numpy as np

# Create a small test tensor
# Shape: (batch_size=2, height=4, width=3, channels=1)
# The function expects a 4D tensor and performs operations on axis 3 (channels)
x = tf.constant(np.random.randn(2, 4, 3, 1).astype(np.float32))

ARGS = {
    "x": x
}
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 — no output shape is stated: observe_signature.py could not observe it (import:TypeError)
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
rand_saturation Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
eps696/stylegan2/src/training/DiffAugment_tf.py
pointer only (licence: NOASSERTION) · get_code("c8f9bf34313b879d")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, eps696/stylegan2 (NOASSERTION). 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.

rand_translation Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
mit-han-lab/data-efficient-gans/DiffAugment_pytorch.py
code served (permissive licence) · get_code("64f17a01e2d642b0")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1) and returned Tensor, ndim 4, dtype family float, shape (2, 4, 8, 6) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn.functional as F

torch.manual_seed(0)

# The function expects a tensor x with shape [batch_size, channels, height, width]
# Based on the function body:
# - x.size(0) is batch size
# - x.size(2) is height
# - x.size(3) is width
# - The function performs random translation on spatial dimensions

x = torch.randn(2, 4, 8, 6)  # batch=2, channels=4, height=8, width=6

ratio = 0.125

ARGS = {
    "x": x,
    "ratio": ratio,
}
Setup block drafted by
a language model, Claude Haiku 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 Tensor, ndim 4, dtype family float, shape (2, 4, 8, 6) — shape (2, 4, 8, 6) 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
rand_translation Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
VITA-Group/Ultra-Data-Efficient-GAN-Training/BigGAN and DiffAugGAN/utils/diff_aug.py
code served (permissive licence) · get_code("4cfe42eb20e188c9")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 4, dtype family float, shape (2, 3, 4, 5) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn.functional as F

torch.manual_seed(0)

# Create a small tensor with shape (batch, channels, height, width)
# Using distinct dimensions: batch=2, channels=3, height=4, width=5
x = torch.randn(2, 3, 4, 5)

# ratio parameter - controls the translation shift as a fraction of image dimensions
ratio = 0.125

ARGS = {
    "x": x,
    "ratio": ratio,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 4, dtype family float, shape (2, 3, 4, 5) — shape (2, 3, 4, 5) 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
rand_translation Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
gaborvecsei/SLE-GAN/sle_gan/diff_augment.py
code served (permissive licence) · get_code("3960a5e348576c26")
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; 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 tensorflow as tf
import numpy as np

# Create a small batch of test images
# Shape: (batch_size, height, width, channels)
batch_size = 2
height = 8
width = 6
channels = 3

x = tf.random.normal([batch_size, height, width, channels], dtype=tf.float32)

ratio = 0.125

ARGS = {
    "x": x,
    "ratio": ratio,
}
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 — no output shape is stated: observe_signature.py could not observe it (import:TypeError)
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
rand_translation Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
eps696/stylegan2/src/training/DiffAugment_tf.py
pointer only (licence: NOASSERTION) · get_code("5895184332d50243")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, eps696/stylegan2 (NOASSERTION). 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.

rand_translation Ran
ran; our contract draft was wrong, not the code
claim-berlin/3d_stylegan_circle_of_willis/diff_augment.py
pointer only (licence: NOASSERTION) · get_code("164a3d8cb51836fd")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, claim-berlin/3d_stylegan_circle_of_willis (NOASSERTION). 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.

FM_reg Not yet run
unverified (harvested, no recorded run)
uzielroy/StyleGan_FewShot/finetune.py
code served (permissive licence) · get_code("a76fa2aebbbe157d")
How it ran

no invocation recorded.

calculate_activation_statistics Not yet run
unverified (harvested, no recorded run)
milmor/LadaGAN-pytorch/fid.py
code served (permissive licence) · get_code("c5a04d1cab5f16a8")
How it ran

no invocation recorded.

calculate_activation_statistics Not yet run
unverified (harvested, no recorded run)
uzielroy/StyleGan_FewShot/metric/fid_score.py
code served (permissive licence) · get_code("9128352cd567690a")
How it ran

no invocation recorded.

get_activations Not yet run
unverified (harvested, no recorded run)
milmor/LadaGAN-pytorch/fid.py
code served (permissive licence) · get_code("63749a9f8d05d87c")
How it ran

no invocation recorded.

get_activations Not yet run
unverified (harvested, no recorded run)
uzielroy/StyleGan_FewShot/metric/fid_score.py
code served (permissive licence) · get_code("a3ef5898d81d8dd1")
How it ran

no invocation recorded.

images_pca Not yet run
unverified (harvested, no recorded run)
uzielroy/StyleGan_FewShot/save_images_monet.py
code served (permissive licence) · get_code("e997ac6a2c12732a")
How it ran

no invocation recorded.

l2_loss Not yet run
unverified (harvested, no recorded run)
milmor/ladagan/tensorflow/trainer.py
code served (permissive licence) · get_code("76bc532052b7504c")
How it ran

no invocation recorded.

pixel_upsample Not yet run
unverified (harvested, no recorded run)
milmor/ladagan/tensorflow/model.py
code served (permissive licence) · get_code("d4eb4c7282c41f98")
How it ran

no invocation recorded.

rand_cutout Not yet run
unverified (harvested, no recorded run)
claim-berlin/3d_stylegan_circle_of_willis/diff_augment.py
pointer only (licence: NOASSERTION) · get_code("390fc5d117993c38")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, claim-berlin/3d_stylegan_circle_of_willis (NOASSERTION). 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.

Also present in other papers' code

Of 58 samples lifted from this paper's linked repositories, 58 carry a function identity and 0 do not. Of the 47 distinct functions with an identity, 21 are also present in other papers whose code we hold (13 with at least one paper dated earlier, 2 the same month, 18 later) and 26 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
calculate_frechet_distance 74 other papers
13 dated earlier, 3 the same month, 58 later
1610.09585 (2016-10)
Conditional Image Synthesis With Auxiliary Classifier GANs
1706.08500 (2017-06)
GANs Trained by a Two Time-Scale Update Rule Converge to a Local Nash Equilibrium
1807.11458 (2018-07)
To learn image super-resolution, use a GAN to learn how to do image degradation first
and 71 more not named here
rand_brightness
brick of this paper's assembly: data_transform, helper
17 other papers
1 dated earlier, 16 later
1912.04958 (2019-12)
Analyzing and Improving the Image Quality of StyleGAN
2007.15646 (2020-07)
Rewriting a Deep Generative Model
2104.03310 (2021-04)
Regularizing Generative Adversarial Networks under Limited Data
and 14 more not named here
rand_saturation
brick of this paper's assembly: data_transform, helper
16 other papers
1 dated earlier, 15 later
1912.04958 (2019-12)
Analyzing and Improving the Image Quality of StyleGAN
2007.15646 (2020-07)
Rewriting a Deep Generative Model
2104.03310 (2021-04)
Regularizing Generative Adversarial Networks under Limited Data
and 13 more not named here
equal_lr 10 other papers
4 dated earlier, 6 later · 1 only as an assembly brick
1710.10196 (2017-10)
Progressive Growing of GANs for Improved Quality, Stability, and Variation
1804.07723 (2018-04)
Image Inpainting for Irregular Holes Using Partial Convolutions
1812.04948 (2018-12)
A Style-Based Generator Architecture for Generative Adversarial Networks
and 7 more not named here
DiffAugment
brick of this paper's assembly: entry
9 other papers
9 later
2104.03310 (2021-04)
Regularizing Generative Adversarial Networks under Limited Data
2111.01007 (2021-11)
Projected GANs Converge Faster
2112.09130 (2021-12)
Ensembling Off-the-shelf Models for GAN Training
and 6 more not named here
calculate_activation_statistics 6 other papers
2 dated earlier, 1 the same month, 3 later
1912.07414 (2019-12)
Learning Canonical Representations for Scene Graph to Image Generation
2003.02567 (2020-03)
GANwriting: Content-Conditioned Generation of Styled Handwritten Word Images
2006.08198 (2020-06)
AutoGAN-Distiller: Searching to Compress Generative Adversarial Networks
and 3 more not named here
d_r1_loss 5 other papers
1 dated earlier, 4 later
1910.05852 (2019-10)
Implicit competitive regularization in GANs
2112.10762 (2021-12)
StyleSwin: Transformer-based GAN for High-resolution Image Generation
2203.04121 (2022-03)
Few Shot Generative Model Adaption via Relaxed Spatial Structural Alignment
and 2 more not named here
get_gaussian_kernel 3 other papers
1 dated earlier, 2 later
1710.10196 (2017-10)
Progressive Growing of GANs for Improved Quality, Stability, and Variation
2007.15651 (2020-07)
Contrastive Learning for Unpaired Image-to-Image Translation
2605.00578 (2026-05)
Federated Distillation for Whole Slide Image via Gaussian-Mixture Feature Alignment and Curriculum Integration
pyramid_down 3 other papers
1 dated earlier, 2 later
1710.10196 (2017-10)
Progressive Growing of GANs for Improved Quality, Stability, and Variation
2007.15651 (2020-07)
Contrastive Learning for Unpaired Image-to-Image Translation
2605.00578 (2026-05)
Federated Distillation for Whole Slide Image via Gaussian-Mixture Feature Alignment and Curriculum Integration
pyramid_up 3 other papers
1 dated earlier, 2 later
1710.10196 (2017-10)
Progressive Growing of GANs for Improved Quality, Stability, and Variation
2007.15651 (2020-07)
Contrastive Learning for Unpaired Image-to-Image Translation
2605.00578 (2026-05)
Federated Distillation for Whole Slide Image via Gaussian-Mixture Feature Alignment and Curriculum Integration
DiffAugment
brick of this paper's assembly: entry
2 other papers
1 dated earlier, 1 later
1912.04958 (2019-12)
Analyzing and Improving the Image Quality of StyleGAN
136830137 (2022 (year only))
Neural Scene Decoration from a Single Photograph
rand_contrast
brick of this paper's assembly: data_transform, helper
2 other papers
2 later
2007.15646 (2020-07)
Rewriting a Deep Generative Model
2111.01007 (2021-11)
Projected GANs Converge Faster
FM_reg 1 other paper
1 dated earlier
1912.05270 (2019-12)
MineGAN: effective knowledge transfer from GANs to target domains with few images
apply_diffaug
brick of this paper's assembly: entry
1 other paper
1 later
2111.01118 (2021-11)
Rebooting ACGAN: Auxiliary Classifier GANs with Stable Training
calculate_activation_statistics 1 other paper
1 later
2212.09748 (2022-12)
Scalable Diffusion Models with Transformers
get_activations 1 other paper
1 dated earlier
2003.08932 (2020-03)
GIQA: Generated Image Quality Assessment
get_activations 1 other paper
1 later
2212.09748 (2022-12)
Scalable Diffusion Models with Transformers
l2_reg 1 other paper
1 dated earlier
1912.05270 (2019-12)
MineGAN: effective knowledge transfer from GANs to target domains with few images
rand_cutout
brick of this paper's assembly: data_transform, helper
1 other paper
1 later
2111.01007 (2021-11)
Projected GANs Converge Faster
rand_translation 1 other paper
1 later
2111.01007 (2021-11)
Projected GANs Converge Faster

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

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

Abstract

The performance of generative adversarial networks (GANs) heavily deteriorates given a limited amount of training data. This is mainly because the discriminator is memorizing the exact training set. To combat it, we propose Differentiable Augmentation (DiffAugment), a simple method that improves the data efficiency of GANs by imposing various types of differentiable augmentations on both real and fake samples. Previous attempts to directly augment the training data manipulate the distribution of real images, yielding little benefit; DiffAugment enables us to adopt the differentiable augmentation for the generated samples, effectively stabilizes training, and leads to better convergence. Experiments demonstrate consistent gains of our method over a variety of GAN architectures and loss functions for both unconditional and class-conditional generation. With DiffAugment, we achieve a state-of-the-art FID of 6.80 with an IS of 100.8 on ImageNet 128×128 and 2-4× reductions of FID given 1,000 images on FFHQ and LSUN. Furthermore, with only 20% training data, we can match the top performance on CIFAR-10 and CIFAR-100. Finally, our method can generate high-fidelity images using only 100 images without pre-training, while being on par with existing transfer learning algorithms. Code is available at https://github.com/mit-han-lab/data-efficient-gans. StyleGAN2 + DiffAugment (ours)

For agents

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

get_harvested_code_for_paper("2006.10738")
get_code_for_paper("2006.10738")
have("2006.10738")

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

curl https://syntology.ai/api/ran/2006.10738.json

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

[![Syntology run record](https://syntology.ai/api/ran/2006.10738.svg)](https://syntology.ai/paper/2006.10738)

Connect an agent — have() is free.