SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2111.01007 · NeurIPS · 2021

Projected GANs Converge Faster

Axel Sauer, Andreas Geiger, Kashyap Chitta, Jens Müller

arXiv · PDF · Open in the Atlas

Code that ran

38 of 49 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (14 of them constructed an object rather than computing a result; 3 honoured the contract we drafted, 0 violated it and 17 have no contract verdict; 18 failures of our instrument; 11 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/2111.01007.json · badge: /api/ran/2111.01007.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
dome272/ProjectedGAN-pytorch reimplementation (classifier label)
classifier basis: classifier_2025
21 ran · 5 no recorded run
autonomousvision/projected_gan alias (classifier label)
classifier basis: classifier_2025
12 ran · 5 no recorded run
copy not recorded community
not in the paper's declared repositories
5 ran · 0 no recorded run
tsubota-kouga/ProjectedGAN community
Papers with Code archive
0 ran · 1 no recorded run
FunctionStatusWhere it lives
CSM Ran
ran on a synthesized input · constructed an object; no result was computed
dome272/ProjectedGAN-pytorch/projected_gan.py
code served (permissive licence) · get_code("aa08964cb4a41f68")
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 an instance of its own class, CSM: the run constructed the object and did not call it, so no result was computed. 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)

# Parameters for CSM.__init__
channels = 4
conv3_out_channels = 8

ARGS = {
    'channels': channels,
    'conv3_out_channels': conv3_out_channels,
}
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 CSM — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a CSM 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
DiffAugment Ran
ran on a synthesized input · constructed an object; no result was computed
dome272/ProjectedGAN-pytorch/projected_gan.py
code served (permissive licence) · get_code("fb73f5657d3d5dfd")
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 an instance of its own class, DiffAugment: the run constructed the object and did not call it, so no result was computed. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# The DiffAugment class takes two parameters:
# 1. policy: a string of comma-separated policy names
# 2. channels_first: a boolean indicating channel ordering

# Based on the code, valid policies are keys in AUGMENT_FNS:
# 'color', 'translation', 'cutout'

ARGS = {
    'policy': 'color,translation,cutout',
    'channels_first': True
}
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 DiffAugment — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a DiffAugment 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
DiffAugment Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
autonomousvision/projected_gan/pg_modules/discriminator.py
code served (permissive licence) · get_code("fb2ae62e229d8e49")
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 batch of image tensors
# Batch size: 2, Channels: 3, Height: 4, Width: 5
x = torch.randn(2, 3, 4, 5, dtype=torch.float32)

# Define ARGS with all parameters of DiffAugment
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, 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
DownBlock Ran
ran on a synthesized input · constructed an object; no result was computed
dome272/ProjectedGAN-pytorch/projected_gan.py
code served (permissive licence) · get_code("93f4a2947c32908b")
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 an instance of its own class, DownBlock: the run constructed the object and did not call it, so no result was computed. 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)

# DownBlock takes c_in and c_out as parameters
# c_in is the number of input channels
# c_out is the number of output channels
# The forward method expects a 4D tensor (batch, channels, height, width)

ARGS = {
    'c_in': 3,
    'c_out': 16
}
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 DownBlock — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a DownBlock 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
DownBlock Ran
ran on a synthesized input · constructed an object; no result was computed
autonomousvision/projected_gan/pg_modules/discriminator.py
code served (permissive licence) · get_code("be3ac00fad675c3d")
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 an instance of its own class, DownBlock: the run constructed the object and did not call it, so no result was computed. 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 torch.nn.utils import spectral_norm

torch.manual_seed(0)

def NormLayer(c, mode='batch'):
    if mode == 'group':
        return nn.GroupNorm(c // 2, c)
    elif mode == 'batch':
        return nn.BatchNorm2d(c)

class SeparableConv2d(nn.Module):

    def __init__(self, in_channels, out_channels, kernel_size, bias=False):
        super(SeparableConv2d, self).__init__()
        self.depthwise = conv2d(in_channels, in_channels, kernel_size=kernel_size, groups=in_channels, bias=bias, padding=1)
        self.pointwise = conv2d(in_channels, out_channels, kernel_size=1, bias=bias)

    def forward(self, x):
        out = self.depthwise(x)
        out = self.pointwise(out)
        return out

def conv2d(*args, **kwargs):
    return spectral_norm(nn.Conv2d(*args, **kwargs))

class DownBlock(nn.Module):

    def __init__(self, in_planes, out_planes, separable=False):
        super().__init__()
        if not separable:
            self.main = nn.Sequential(conv2d(in_planes, out_planes, 4, 2, 1), NormLayer(out_planes), nn.LeakyReLU(0.2, inplace=True))
        else:
            self.main = nn.Sequential(SeparableConv2d(in_planes, out_planes, 3), NormLayer(out_planes), nn.LeakyReLU(0.2, inplace=True), nn.AvgPool2d(2, 2))

    def forward(self, feat):
        return self.main(feat)

ARGS = {
    'in_planes': 3,
    'out_planes': 16,
    'separable': False
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type DownBlock — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a DownBlock 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
DownBlockPatch Ran
ran on a synthesized input · constructed an object; no result was computed
autonomousvision/projected_gan/pg_modules/discriminator.py
code served (permissive licence) · get_code("2698386cbe6cddd4")
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 an instance of its own class, DownBlockPatch: the run constructed the object and did not call it, so no result was computed. 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 torch.nn.utils import spectral_norm

torch.manual_seed(0)

def NormLayer(c, mode='batch'):
    if mode == 'group':
        return nn.GroupNorm(c // 2, c)
    elif mode == 'batch':
        return nn.BatchNorm2d(c)

def conv2d(*args, **kwargs):
    return spectral_norm(nn.Conv2d(*args, **kwargs))

class SeparableConv2d(nn.Module):
    def __init__(self, in_channels, out_channels, kernel_size, bias=False):
        super(SeparableConv2d, self).__init__()
        self.depthwise = conv2d(in_channels, in_channels, kernel_size=kernel_size, groups=in_channels, bias=bias, padding=1)
        self.pointwise = conv2d(in_channels, out_channels, kernel_size=1, bias=bias)

    def forward(self, x):
        out = self.depthwise(x)
        out = self.pointwise(out)
        return out

class DownBlock(nn.Module):
    def __init__(self, in_planes, out_planes, separable=False):
        super().__init__()
        if not separable:
            self.main = nn.Sequential(conv2d(in_planes, out_planes, 4, 2, 1), NormLayer(out_planes), nn.LeakyReLU(0.2, inplace=True))
        else:
            self.main = nn.Sequential(SeparableConv2d(in_planes, out_planes, 3), NormLayer(out_planes), nn.LeakyReLU(0.2, inplace=True), nn.AvgPool2d(2, 2))

    def forward(self, feat):
        return self.main(feat)

class DownBlockPatch(nn.Module):
    def __init__(self, in_planes, out_planes, separable=False):
        super().__init__()
        self.main = nn.Sequential(DownBlock(in_planes, out_planes, separable), conv2d(out_planes, out_planes, 1, 1, 0, bias=False), NormLayer(out_planes), nn.LeakyReLU(0.2, inplace=True))

    def forward(self, feat):
        return self.main(feat)

ARGS = {
    'in_planes': 3,
    'out_planes': 8,
    'separable': False,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type DownBlockPatch — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a DownBlockPatch 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
EfficientNetLite Ran
ran on a synthesized input · constructed an object; no result was computed
dome272/ProjectedGAN-pytorch/projected_gan.py
code served (permissive licence) · get_code("a7241f6b2a584c67")
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 an instance of its own class, EfficientNetLite: the run constructed the object and did not call it, so no result was computed. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn
import math

torch.manual_seed(0)

ARGS = {
    "widthi_multiplier": 1.0,
    "depth_multiplier": 1.0,
    "num_classes": 10,
    "drop_connect_rate": 0.2,
    "dropout_rate": 0.3
}
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 EfficientNetLite — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a EfficientNetLite 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
FeatureFusionBlock Ran
ran on a synthesized input · constructed an object; no result was computed
autonomousvision/projected_gan/pg_modules/discriminator.py
code served (permissive licence) · get_code("d594247fe0a8ac63")
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 an instance of its own class, FeatureFusionBlock: the run constructed the object and did not call it, so no result was computed. 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)

# Define the arguments for FeatureFusionBlock
ARGS = {
    "features": 64,
    "activation": nn.ReLU(),
    "deconv": False,
    "bn": False,
    "expand": False,
    "align_corners": True,
    "lowest": False,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type FeatureFusionBlock — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a FeatureFusionBlock 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
GLU Ran
ran on a synthesized input
dome272/ProjectedGAN-pytorch/projected_gan.py
code served (permissive licence) · get_code("daa6b7ac4d6d0dfe")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, class construct-then-call prompt, model not recorded) and returned Tensor, ndim 2, dtype family float, shape (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 torch.nn as nn

class GLU(nn.Module):

    def forward(self, x):
        nc = x.size(1)
        assert nc % 2 == 0, 'channels dont divide 2!'
        nc = int(nc / 2)
        return x[:, :nc] * torch.sigmoid(x[:, nc:])


INIT_ARGS = {}

ARGS = {
    'x': torch.randn(2, 4)
}

if __name__ == '__main__':
    model = GLU(**INIT_ARGS)
    output = model(**ARGS)
    print(output)
    print(output.shape)
Setup block drafted by
a language model, class construct-then-call prompt, model not recorded
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 2, dtype family float, shape (2, 2) — shape (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: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
Generator Ran
ran on a synthesized input · constructed an object; no result was computed
dome272/ProjectedGAN-pytorch/projected_gan.py
code served (permissive licence) · get_code("860a49493f25946b")
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 an instance of its own class, Generator: the run constructed the object and did not call it, so no result was computed. 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 = {
    'ngf': 64,
    'nz': 100,
    'nc': 3,
    'im_size': 1024,
}
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 Generator — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a Generator 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
InitLayer Ran
ran on a synthesized input · constructed an object; no result was computed
dome272/ProjectedGAN-pytorch/projected_gan.py
code served (permissive licence) · get_code("c020e77bd5b82cc6")
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 an instance of its own class, InitLayer: the run constructed the object and did not call it, so no result was computed. 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 torch.nn.utils import spectral_norm

torch.manual_seed(0)

class GLU(nn.Module):
    def forward(self, x):
        nc = x.size(1)
        assert nc % 2 == 0, 'channels dont divide 2!'
        nc = int(nc / 2)
        return x[:, :nc] * torch.sigmoid(x[:, nc:])

def batchNorm2d(*args, **kwargs):
    return nn.BatchNorm2d(*args, **kwargs)

def convTranspose2d(*args, **kwargs):
    return spectral_norm(nn.ConvTranspose2d(*args, **kwargs))

class InitLayer(nn.Module):
    def __init__(self, nz, channel):
        super().__init__()
        self.init = nn.Sequential(convTranspose2d(nz, channel * 2, 4, 1, 0, bias=False), batchNorm2d(channel * 2), GLU())

    def forward(self, noise):
        noise = noise.view(noise.shape[0], -1, 1, 1)
        return self.init(noise)

ARGS = {
    'nz': 4,
    'channel': 3
}
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 InitLayer — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a InitLayer 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
MBConvBlock Ran
ran on a synthesized input · constructed an object; no result was computed
dome272/ProjectedGAN-pytorch/projected_gan.py
code served (permissive licence) · get_code("d65e8029380ee87d")
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 an instance of its own class, MBConvBlock: the run constructed the object and did not call it, so no result was computed. 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)

# Create input tensor: batch_size=2, channels=4, height=8, width=8
input_tensor = torch.randn(2, 4, 8, 8)

ARGS = {
    'inp': 4,                    # input channels
    'final_oup': 6,              # output channels (different from input)
    'k': 3,                       # kernel size for depthwise conv
    's': 1,                       # stride
    'expand_ratio': 2,            # expansion ratio
    'se_ratio': 0.25,             # squeeze-excitation ratio
    'has_se': True                # enable SE module
}
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 MBConvBlock — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a MBConvBlock 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
MultiScaleDiscriminator Ran
ran on a synthesized input · constructed an object; no result was computed
dome272/ProjectedGAN-pytorch/projected_gan.py
code served (permissive licence) · get_code("c102280a98c32877")
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 an instance of its own class, MultiScaleDiscriminator: the run constructed the object and did not call it, so no result was computed. 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 torch.optim import Adam

torch.manual_seed(0)

ARGS = {
    "channels": 3,
    "l": 2,
}
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 MultiScaleDiscriminator — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a MultiScaleDiscriminator 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
NoiseInjection Ran
ran on a synthesized input
dome272/ProjectedGAN-pytorch/projected_gan.py
code served (permissive licence) · get_code("9a620288c91dd5f1")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, class construct-then-call prompt, model not recorded) and returned Tensor, 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 torch
import torch.nn as nn

class NoiseInjection(nn.Module):

    def __init__(self):
        super().__init__()
        self.weight = nn.Parameter(torch.zeros(1), requires_grad=True)

    def forward(self, feat, noise=None):
        if noise is None:
            batch, _, height, width = feat.shape
            noise = torch.randn(batch, 1, height, width).to(feat.device)
        return feat + self.weight * noise


INIT_ARGS = {}

ARGS = {
    'feat': torch.randn(2, 4, 3, 5),
    'noise': None
}

if __name__ == '__main__':
    model = NoiseInjection(**INIT_ARGS)
    output = model(**ARGS)
    print(f"Output shape: {output.shape}")
    print(f"Output: {output}")
Setup block drafted by
a language model, class construct-then-call prompt, model not recorded
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 — no output shape is stated: observe_signature.py observed this class with its constructor sizes canonicalised where it could, which the node does not record, so its shape is not stated as this setup block's
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: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
NormLayer Ran
ran; our contract draft was wrong, not the code
autonomousvision/projected_gan/pg_modules/discriminator.py
code served (permissive licence) · get_code("ef1e7dffb5d433d4")
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 BatchNorm2d. 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 = {
    'c': 16,
    'mode': 'batch'
}
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 BatchNorm2d — a BatchNorm2d output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a BatchNorm2d 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
SEBlock Ran
ran on a synthesized input · constructed an object; no result was computed
dome272/ProjectedGAN-pytorch/projected_gan.py
code served (permissive licence) · get_code("3ee6dbc09a4b76ac")
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 an instance of its own class, SEBlock: the run constructed the object and did not call it, so no result was computed. 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 torch.nn.utils import spectral_norm

torch.manual_seed(0)

class Swish(nn.Module):
    def forward(self, feat):
        return feat * torch.sigmoid(feat)

def conv2d(*args, **kwargs):
    return spectral_norm(nn.Conv2d(*args, **kwargs))

class SEBlock(nn.Module):
    def __init__(self, ch_in, ch_out):
        super().__init__()
        self.main = nn.Sequential(
            nn.AdaptiveAvgPool2d(4),
            conv2d(ch_in, ch_out, 4, 1, 0, bias=False),
            Swish(),
            conv2d(ch_out, ch_out, 1, 1, 0, bias=False),
            nn.Sigmoid()
        )

    def forward(self, feat_small, feat_big):
        return feat_big * self.main(feat_small)

# Create SEBlock instance with ch_in=4, ch_out=3
se_block = SEBlock(ch_in=4, ch_out=3)

# Create test tensors
# feat_small: batch=2, channels=4, height=8, width=8
feat_small = torch.randn(2, 4, 8, 8)
# feat_big: batch=2, channels=3, height=8, width=8
feat_big = torch.randn(2, 3, 8, 8)

ARGS = {
    'ch_in': 4,
    'ch_out': 3
}
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 SEBlock — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a SEBlock 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
SeparableConv2d Ran
ran on a synthesized input · constructed an object; no result was computed
autonomousvision/projected_gan/pg_modules/discriminator.py
code served (permissive licence) · get_code("537c69fac8de5c08")
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 an instance of its own class, SeparableConv2d: the run constructed the object and did not call it, so no result was computed. 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 torch.nn.utils import spectral_norm

torch.manual_seed(0)

def conv2d(*args, **kwargs):
    return spectral_norm(nn.Conv2d(*args, **kwargs))

class SeparableConv2d(nn.Module):

    def __init__(self, in_channels, out_channels, kernel_size, bias=False):
        super(SeparableConv2d, self).__init__()
        self.depthwise = conv2d(in_channels, in_channels, kernel_size=kernel_size, groups=in_channels, bias=bias, padding=1)
        self.pointwise = conv2d(in_channels, out_channels, kernel_size=1, bias=bias)

    def forward(self, x):
        out = self.depthwise(x)
        out = self.pointwise(out)
        return out


ARGS = {
    'in_channels': 3,
    'out_channels': 16,
    'kernel_size': 3,
    'bias': False
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type SeparableConv2d — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a SeparableConv2d 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
SingleDisc Ran
ran on a synthesized input · constructed an object; no result was computed
autonomousvision/projected_gan/pg_modules/discriminator.py
code served (permissive licence) · get_code("012485a042fbe914")
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 an instance of its own class, SingleDisc: the run constructed the object and did not call it, so no result was computed. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn
from functools import partial

torch.manual_seed(0)

# Define ARGS dictionary with all parameters for SingleDisc
ARGS = {
    'nc': 3,
    'ndf': 64,
    'start_sz': 256,
    'end_sz': 8,
    'head': True,
    'separable': False,
    'patch': False
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type SingleDisc — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a SingleDisc 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
Swish Ran
ran on a synthesized input
dome272/ProjectedGAN-pytorch/projected_gan.py
code served (permissive licence) · get_code("396c57f5f6d8edd7")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, class construct-then-call prompt, model not recorded) 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 torch.nn as nn

class Swish(nn.Module):

    def forward(self, feat):
        return feat * torch.sigmoid(feat)


INIT_ARGS = {}

ARGS = {
    'feat': torch.randn(2, 4)
}

if __name__ == '__main__':
    module = Swish(**INIT_ARGS)
    output = module(**ARGS)
    print(output)
Setup block drafted by
a language model, class construct-then-call prompt, model not recorded
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 2, dtype family float, shape (2, 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: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
UpBlock Ran
ran; our contract draft was wrong, not the code
dome272/ProjectedGAN-pytorch/projected_gan.py
code served (permissive licence) · get_code("32d6c5741cfd6028")
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 torch
import torch.nn as nn
from torch.nn.utils import spectral_norm

torch.manual_seed(0)

class GLU(nn.Module):
    def forward(self, x):
        nc = x.size(1)
        assert nc % 2 == 0, 'channels dont divide 2!'
        nc = int(nc / 2)
        return x[:, :nc] * torch.sigmoid(x[:, nc:])

def batchNorm2d(*args, **kwargs):
    return nn.BatchNorm2d(*args, **kwargs)

def conv2d(*args, **kwargs):
    return spectral_norm(nn.Conv2d(*args, **kwargs))

def UpBlock(in_planes, out_planes):
    block = nn.Sequential(nn.Upsample(scale_factor=2, mode='nearest'), conv2d(in_planes, out_planes * 2, 3, 1, 1, bias=False), batchNorm2d(out_planes * 2), GLU())
    return block

ARGS = {
    'in_planes': 4,
    'out_planes': 8
}
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 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 1 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
UpBlockComp Ran
ran; our contract draft was wrong, not the code
dome272/ProjectedGAN-pytorch/projected_gan.py
code served (permissive licence) · get_code("d8f6aa2842cb7bc6")
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 torch
import torch.nn as nn
from torch.nn.utils import spectral_norm

torch.manual_seed(0)

class GLU(nn.Module):
    def forward(self, x):
        nc = x.size(1)
        assert nc % 2 == 0, 'channels dont divide 2!'
        nc = int(nc / 2)
        return x[:, :nc] * torch.sigmoid(x[:, nc:])

class NoiseInjection(nn.Module):
    def __init__(self):
        super().__init__()
        self.weight = nn.Parameter(torch.zeros(1), requires_grad=True)

    def forward(self, feat, noise=None):
        if noise is None:
            batch, _, height, width = feat.shape
            noise = torch.randn(batch, 1, height, width).to(feat.device)
        return feat + self.weight * noise

def batchNorm2d(*args, **kwargs):
    return nn.BatchNorm2d(*args, **kwargs)

def conv2d(*args, **kwargs):
    return spectral_norm(nn.Conv2d(*args, **kwargs))

def UpBlockComp(in_planes, out_planes):
    block = nn.Sequential(
        nn.Upsample(scale_factor=2, mode='nearest'),
        conv2d(in_planes, out_planes * 2, 3, 1, 1, bias=False),
        NoiseInjection(),
        batchNorm2d(out_planes * 2),
        GLU(),
        conv2d(out_planes, out_planes * 2, 3, 1, 1, bias=False),
        NoiseInjection(),
        batchNorm2d(out_planes * 2),
        GLU()
    )
    return block

ARGS = {
    "in_planes": 4,
    "out_planes": 8
}
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 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 1 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_efficientnet Ran
ran; our contract draft was wrong, not the code
autonomousvision/projected_gan/pg_modules/discriminator.py
code served (permissive licence) · get_code("6e9fa10927d234c5")
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 Module. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn

torch.manual_seed(0)

class MockEfficientNet(nn.Module):
    def __init__(self):
        super().__init__()
        self.conv_stem = nn.Conv2d(3, 8, kernel_size=3, padding=1)
        self.bn1 = nn.BatchNorm2d(8)
        self.act1 = nn.ReLU()
        
        self.blocks = nn.ModuleList([
            nn.Conv2d(8, 16, kernel_size=3, padding=1),
            nn.Conv2d(16, 16, kernel_size=3, padding=1),
            nn.Conv2d(16, 32, kernel_size=3, padding=1),
            nn.Conv2d(32, 32, kernel_size=3, padding=1),
            nn.Conv2d(32, 32, kernel_size=3, padding=1),
            nn.Conv2d(32, 64, kernel_size=3, padding=1),
            nn.Conv2d(64, 64, kernel_size=3, padding=1),
            nn.Conv2d(64, 64, kernel_size=3, padding=1),
            nn.Conv2d(64, 64, kernel_size=3, padding=1),
        ])

model = MockEfficientNet()

ARGS = {
    "model": model
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Module — a Module output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Module 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_scratch_ccm Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
autonomousvision/projected_gan/pg_modules/discriminator.py
code served (permissive licence) · get_code("34b8baa98458240a")
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 Scratch. 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)

# Create a simple scratch object to hold attributes
class Scratch:
    pass

scratch = Scratch()

# Define input channels for each layer (4 layers with distinct sizes)
in_channels = [64, 128, 256, 512]

# Define output channel count
cout = 32

# Define expand flag
expand = False

ARGS = {
    "scratch": scratch,
    "in_channels": in_channels,
    "cout": cout,
    "expand": expand
}
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 Scratch — a Scratch output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Scratch 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_scratch_csm Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
autonomousvision/projected_gan/pg_modules/discriminator.py
code served (permissive licence) · get_code("022a2fd988863842")
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 ScratchStandIn. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn

torch.manual_seed(0)

class ScratchStandIn:
    """Stand-in class for scratch object with layer attributes"""
    pass

# Create a scratch object
scratch = ScratchStandIn()

# Define input channels for the different layers
in_channels = [32, 64, 128, 256]

# Output channel size
cout = 16

# Expand flag
expand = True

ARGS = {
    "scratch": scratch,
    "in_channels": in_channels,
    "cout": cout,
    "expand": expand,
}
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 ScratchStandIn — a ScratchStandIn output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a ScratchStandIn 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
batchNorm2d Ran
ran; our contract draft was wrong, not the code
dome272/ProjectedGAN-pytorch/projected_gan.py
code served (permissive licence) · get_code("83ae32deb9c6d5eb")
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 BatchNorm2d. 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 = {
    'num_features': 4,
}
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 BatchNorm2d — a BatchNorm2d output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a BatchNorm2d 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
calc_channels Ran
ran, honoured the contract we drafted
autonomousvision/projected_gan/pg_modules/discriminator.py
code served (permissive licence) · get_code("d1223b5ae8d2262a")
How it ran

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

import torch
import torch.nn as nn

torch.manual_seed(0)

class StandInModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.layer0 = nn.Conv2d(3, 4, kernel_size=3, padding=1)
        self.layer1 = nn.Conv2d(4, 5, kernel_size=3, padding=1)
        self.layer2 = nn.Conv2d(5, 6, kernel_size=3, padding=1)
        self.layer3 = nn.Conv2d(6, 7, kernel_size=3, padding=1)
    
    def forward(self, x):
        return x

pretrained_model = StandInModel()

ARGS = {
    'pretrained': pretrained_model,
    'inp_res': 224
}
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 int — a list output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
conv2d Ran
ran; our contract draft was wrong, not the code
dome272/ProjectedGAN-pytorch/projected_gan.py
code served (permissive licence) · get_code("d840598eafed7a9d")
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 Conv2d. 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 torch.nn.utils import spectral_norm

torch.manual_seed(0)

ARGS = {
    'in_channels': 3,
    'out_channels': 16,
    'kernel_size': 3,
    'stride': 1,
    'padding': 1,
    'dilation': 1,
    'groups': 1,
    'bias': True,
    'padding_mode': 'zeros',
}
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 Conv2d — a Conv2d output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Conv2d 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
convTranspose2d Ran
ran; our contract draft was wrong, not the code
dome272/ProjectedGAN-pytorch/projected_gan.py
code served (permissive licence) · get_code("b27240435837ab3d")
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 ConvTranspose2d. 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 torch.nn.utils import spectral_norm

torch.manual_seed(0)

ARGS = {
    'in_channels': 3,
    'out_channels': 4,
    'kernel_size': 3,
    'stride': 1,
    'padding': 0,
    'output_padding': 0,
    'groups': 1,
    'bias': True,
    'dilation': 1,
    'padding_mode': 'zeros'
}
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 ConvTranspose2d — a ConvTranspose2d output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a ConvTranspose2d 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
drop_connect Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
dome272/ProjectedGAN-pytorch/projected_gan.py
code served (permissive licence) · get_code("ed953cd5fd3a44c8")
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, 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)

# Create a small input tensor with shape [batch_size=2, channels=4, height=3, width=3]
x = torch.randn(2, 4, 3, 3, dtype=torch.float32)

# Drop connect rate (probability of dropping a connection)
drop_connect_rate = 0.2

# Training flag
training = True

ARGS = {
    "x": x,
    "drop_connect_rate": drop_connect_rate,
    "training": training,
}
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, 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
load_data Ran
ran; our contract draft was wrong, not the code
dome272/ProjectedGAN-pytorch/projected_gan.py
code served (permissive licence) · get_code("243747f15fce5ac8")
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 DataLoader. 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 tempfile
import os
from PIL import Image
import numpy as np

# Create a temporary directory with dummy image data
temp_dir = tempfile.mkdtemp()
train_dir = os.path.join(temp_dir, "train")
class_dir = os.path.join(train_dir, "class_0")
os.makedirs(class_dir, exist_ok=True)

# Create a few dummy images
for i in range(3):
    img_array = np.random.randint(0, 256, (64, 64, 3), dtype=np.uint8)
    img = Image.fromarray(img_array)
    img.save(os.path.join(class_dir, f"image_{i}.jpg"))

ARGS = {
    "data_path": train_dir,
    "batch_size": 2
}
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 DataLoader — a DataLoader output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a DataLoader 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
rand_brightness Ran
ran; our contract draft was wrong, not the code
this paper's copy was not recorded; identical code first harvested from postech-cvlab/pytorch-studiogan
pointer only; 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 copy of the code was not recorded (its link names no repository or source path), so no licence was read for it. 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)
this paper's copy was not recorded; identical code first harvested from mit-han-lab/data-efficient-gans
pointer only · get_code("cf3f99cfbe8f451a")
How it ran

The body of this sample is withheld: this paper's copy of the code was not recorded (its link names no repository or source path), so no licence was read for it. 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)
this paper's copy was not recorded; identical code first harvested from mit-han-lab/data-efficient-gans
pointer only · get_code("8a8f1bb31f56eb41")
How it ran

The body of this sample is withheld: this paper's copy of the code was not recorded (its link names no repository or source path), so no licence was read for it. 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; our contract draft was wrong, not the code
autonomousvision/projected_gan/pg_modules/discriminator.py
code served (permissive licence) · get_code("ad4806edfa54ef28")
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, 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

torch.manual_seed(0)

# x is a 4D tensor with shape (batch, channels, height, width)
# The function uses:
# - x.size(0), x.size(2), x.size(3) for dimensions
# - x.device for device placement
# - x.dtype for dtype
# - Indexing and multiplication operations
x = torch.randn(2, 3, 8, 6)  # batch=2, channels=3, height=8, width=6

# ratio parameter - a float controlling cutout size
ratio = 0.2

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, 8, 6) — shape (2, 3, 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_saturation Ran
ran; our contract draft was wrong, not the code
this paper's copy was not recorded; identical code first harvested from postech-cvlab/pytorch-studiogan
pointer only; 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 copy of the code was not recorded (its link names no repository or source path), so no licence was read for it. 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)
this paper's copy was not recorded; identical code first harvested from mit-han-lab/data-efficient-gans
pointer only · get_code("64f17a01e2d642b0")
How it ran

The body of this sample is withheld: this paper's copy of the code was not recorded (its link names no repository or source path), so no licence was read for it. 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.

round_filters Ran
ran, honoured the contract we drafted
dome272/ProjectedGAN-pytorch/projected_gan.py
code served (permissive licence) · get_code("1522c02775ece4c4")
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

torch.manual_seed(0)

ARGS = {
    "filters": 32,
    "multiplier": 1.5,
    "divisor": 8,
    "min_width": 4
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type 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
round_repeats Ran
ran, honoured the contract we drafted
dome272/ProjectedGAN-pytorch/projected_gan.py
code served (permissive licence) · get_code("b5099ac2f4a497ef")
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 math

def round_repeats(repeats, multiplier):
    """Round number of filters based on depth multiplier."""
    if not multiplier:
        return repeats
    return int(math.ceil(multiplier * repeats))

# Setup code
ARGS = {
    "repeats": 4,
    "multiplier": 1.5
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type int, ndim 0, dtype family int, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
F_RandomProj Not yet run
unverified (harvested, no recorded run)
autonomousvision/projected_gan/pg_modules/discriminator.py
code served (permissive licence) · get_code("d50f654da1e833ff")
How it ran

no invocation recorded.

MultiScaleD Not yet run
unverified (harvested, no recorded run)
autonomousvision/projected_gan/pg_modules/discriminator.py
code served (permissive licence) · get_code("c5ec5be19a01b214")
How it ran

no invocation recorded.

ProjectedDiscriminator Not yet run
unverified (harvested, no recorded run)
autonomousvision/projected_gan/pg_modules/discriminator.py
code served (permissive licence) · get_code("c3ff68174f3a9086")
How it ran

no invocation recorded.

ProjectedGAN Not yet run
unverified (harvested, no recorded run)
dome272/ProjectedGAN-pytorch/projected_gan.py
code served (permissive licence) · get_code("d133245b208a1c3b")
How it ran

no invocation recorded.

Projector Not yet run
unverified (harvested, no recorded run)
tsubota-kouga/ProjectedGAN/network.py
pointer only (licence: NONE) · get_code("41472db28a1d6bb6")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, tsubota-kouga/ProjectedGAN 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.

SingleDiscCond Not yet run
unverified (harvested, no recorded run)
autonomousvision/projected_gan/pg_modules/discriminator.py
code served (permissive licence) · get_code("6b8d529b6b9a1c09")
How it ran

no invocation recorded.

_make_projector Not yet run
unverified (harvested, no recorded run)
autonomousvision/projected_gan/pg_modules/discriminator.py
code served (permissive licence) · get_code("15479f3c367a9af9")
How it ran

no invocation recorded.

build_efficientnet_lite Not yet run
unverified (harvested, no recorded run)
dome272/ProjectedGAN-pytorch/projected_gan.py
code served (permissive licence) · get_code("e1b3d8c3301b619a")
How it ran

no invocation recorded.

kaiming_init Not yet run
unverified (harvested, no recorded run)
dome272/ProjectedGAN-pytorch/projected_gan.py
code served (permissive licence) · get_code("b1c1a8a6e7a54d45")
How it ran

no invocation recorded.

load_checkpoint Not yet run
unverified (harvested, no recorded run)
dome272/ProjectedGAN-pytorch/projected_gan.py
code served (permissive licence) · get_code("0754baca942753a6")
How it ran

no invocation recorded.

weights_init Not yet run
unverified (harvested, no recorded run)
dome272/ProjectedGAN-pytorch/projected_gan.py
code served (permissive licence) · get_code("0fb299a87938eec8")
How it ran

no invocation recorded.

Also present in other papers' code

Of 49 samples lifted from this paper's linked repositories, 49 carry a function identity and 0 do not. Of the 49 distinct functions with an identity, 18 are also present in other papers whose code we hold (17 with at least one paper dated earlier, 2 the same month, 6 later) and 31 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
rand_brightness
brick of this paper's assembly: helper
17 other papers
4 dated earlier, 1 the same month, 12 later
1912.04958 (2019-12)
Analyzing and Improving the Image Quality of StyleGAN
2006.10738 (2020-06)
Differentiable Augmentation for Data-Efficient GAN Training
2007.15646 (2020-07)
Rewriting a Deep Generative Model
and 14 more not named here
rand_saturation
brick of this paper's assembly: helper
16 other papers
4 dated earlier, 1 the same month, 11 later
1912.04958 (2019-12)
Analyzing and Improving the Image Quality of StyleGAN
2006.10738 (2020-06)
Differentiable Augmentation for Data-Efficient GAN Training
2007.15646 (2020-07)
Rewriting a Deep Generative Model
and 13 more not named here
DiffAugment
brick of this paper's assembly: helper
9 other papers
2 dated earlier, 7 later
2006.10738 (2020-06)
Differentiable Augmentation for Data-Efficient GAN Training
2104.03310 (2021-04)
Regularizing Generative Adversarial Networks under Limited Data
2112.09130 (2021-12)
Ensembling Off-the-shelf Models for GAN Training
and 6 more not named here
NormLayer
brick of this paper's assembly: helper
6 other papers
6 later
2202.00273 (2022-02)
StyleGAN-XL: Scaling StyleGAN to Large Diverse Datasets
2306.04675 (2023-06)
Exposing flaws of generative model evaluation metrics and their unfair treatment of diffusion models
2310.02279 (2023-10)
Consistency Trajectory Models: Learning Probability Flow ODE Trajectory of Diffusion
and 3 more not named here
UpBlock
brick of this paper's assembly: helper
4 other papers
1 dated earlier, 3 later · 1 only as an assembly brick
2101.04775 (2021-01)
Towards Faster and Stabilized GAN Training for High-fidelity Few-shot Image Synthesis
2310.18894 (2023-10)
Emergence of Shape Bias in Convolutional Neural Networks through Activation Sparsity
Ni_CHAIN_Enhancing_Generalization_in_Data-Efficient_GANs_via_lipsCHitz_continuity_constrAIned_CVPR_2024_paper (2024 (year only))
⃝ ⃝ CHAIN: Enhancing Generalization in Data-Efficient GANs via lipsCHitz continuity constrAIned Normalization
and 1 more not named here
UpBlockComp
brick of this paper's assembly: helper
2 other papers
1 dated earlier, 1 later · 1 only as an assembly brick
2101.04775 (2021-01)
Towards Faster and Stabilized GAN Training for High-fidelity Few-shot Image Synthesis
Ni_CHAIN_Enhancing_Generalization_in_Data-Efficient_GANs_via_lipsCHitz_continuity_constrAIned_CVPR_2024_paper (2024 (year only))
⃝ ⃝ CHAIN: Enhancing Generalization in Data-Efficient GANs via lipsCHitz continuity constrAIned Normalization
rand_contrast
brick of this paper's assembly: helper
2 other papers
2 dated earlier
2006.10738 (2020-06)
Differentiable Augmentation for Data-Efficient GAN Training
2007.15646 (2020-07)
Rewriting a Deep Generative Model
GLU
brick of this paper's assembly: helper
1 other paper
1 dated earlier · 1 only as an assembly brick
2101.04775 (2021-01)
Towards Faster and Stabilized GAN Training for High-fidelity Few-shot Image Synthesis
InitLayer
brick of this paper's assembly: helper
1 other paper
1 dated earlier · 1 only as an assembly brick
2101.04775 (2021-01)
Towards Faster and Stabilized GAN Training for High-fidelity Few-shot Image Synthesis
NoiseInjection
brick of this paper's assembly: helper
1 other paper
1 dated earlier · 1 only as an assembly brick
2101.04775 (2021-01)
Towards Faster and Stabilized GAN Training for High-fidelity Few-shot Image Synthesis
SEBlock
brick of this paper's assembly: helper
1 other paper
1 dated earlier · 1 only as an assembly brick
2101.04775 (2021-01)
Towards Faster and Stabilized GAN Training for High-fidelity Few-shot Image Synthesis
Swish
brick of this paper's assembly: helper
1 other paper
1 dated earlier · 1 only as an assembly brick
2101.04775 (2021-01)
Towards Faster and Stabilized GAN Training for High-fidelity Few-shot Image Synthesis
batchNorm2d
brick of this paper's assembly: helper
1 other paper
1 dated earlier · 1 only as an assembly brick
2101.04775 (2021-01)
Towards Faster and Stabilized GAN Training for High-fidelity Few-shot Image Synthesis
conv2d
brick of this paper's assembly: helper
1 other paper
1 dated earlier · 1 only as an assembly brick
2101.04775 (2021-01)
Towards Faster and Stabilized GAN Training for High-fidelity Few-shot Image Synthesis
convTranspose2d
brick of this paper's assembly: helper
1 other paper
1 dated earlier · 1 only as an assembly brick
2101.04775 (2021-01)
Towards Faster and Stabilized GAN Training for High-fidelity Few-shot Image Synthesis
rand_cutout
brick of this paper's assembly: helper
1 other paper
1 dated earlier
2006.10738 (2020-06)
Differentiable Augmentation for Data-Efficient GAN Training
rand_translation
brick of this paper's assembly: helper
1 other paper
1 dated earlier
2006.10738 (2020-06)
Differentiable Augmentation for Data-Efficient GAN Training
round_repeats
brick of this paper's assembly: helper
1 other paper
1 dated earlier
1508.04306 (2015-08)
Deep clustering: Discriminative embeddings for segmentation and separation

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

Generative Adversarial Networks (GANs) produce high-quality images but are challenging to train. They need careful regularization, vast amounts of compute, and expensive hyper-parameter sweeps. We make significant headway on these issues by projecting generated and real samples into a fixed, pretrained feature space. Motivated by the finding that the discriminator cannot fully exploit features from deeper layers of the pretrained model, we propose a more effective strategy that mixes features across channels and resolutions. Our Projected GAN improves image quality, sample efficiency, and convergence speed. It is further compatible with resolutions of up to one Megapixel and advances the state-of-the-art Fréchet Inception Distance (FID) on twenty-two benchmark datasets. Importantly, Projected GANs match the previously lowest FIDs up to 40 times faster, cutting the wall-clock time from 5 days to less than 3 hours given the same computational resources.

For agents

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

get_harvested_code_for_paper("2111.01007")
get_code_for_paper("2111.01007")
have("2111.01007")

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

curl https://syntology.ai/api/ran/2111.01007.json

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

[![Syntology run record](https://syntology.ai/api/ran/2111.01007.svg)](https://syntology.ai/paper/2111.01007)

Connect an agent — have() is free.