SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2008.02401 · 2020

StyleFlow: Attribute-conditioned Exploration of StyleGAN-Generated Images using Conditional Continuous Normalizing Flows

arXiv · PDF · Open in the Atlas

Code that ran

9 of 14 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; 1 honoured the contract we drafted, 0 violated it and 7 have no contract verdict; 1 failure of our instrument; 5 no recorded run), measured 2026-11-21; 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/2008.02401.json · badge: /api/ran/2008.02401.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
xuduo35/StyleFlowPytorch community (classifier label)
classifier basis not recorded on the link
9 ran · 5 no recorded run
FunctionStatusWhere it lives
D_logistic Ran
ran on a synthesized input
xuduo35/StyleFlowPytorch/stylegan2/loss_fns.py
code served (permissive licence) · get_code("eb49d8acd5d18521")
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
from torch.nn import functional as F

torch.manual_seed(0)

# Define minimal Generator model
class Generator(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear = nn.Linear(4, 4)
    
    def forward(self, latents, labels=None):
        # Takes latents and optional labels, returns fake images
        return self.linear(latents)

# Define minimal Discriminator model
class Discriminator(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear = nn.Linear(4, 1)
    
    def forward(self, x, labels=None):
        # Takes images and optional labels, returns scores
        return self.linear(x)

# Define minimal augment function
class Augment(nn.Module):
    def __init__(self):
        super().__init__()
    
    def forward(self, x, p=0.6):
        # Returns augmented tensor and augmentation parameters
        return x, None

# Create instances
G = Generator()
D = Discriminator()
augment = Augment()

# Create batch tensors
batch_size = 2
latent_dim = 4
image_dim = 4

latents = torch.randn(batch_size, latent_dim)
reals = torch.randn(batch_size, image_dim)
latent_labels = torch.tensor([0, 1])
real_labels = torch.tensor([0, 1])

ARGS = {
    'G': G,
    'D': D,
    'latents': latents,
    'reals': reals,
    'latent_labels': latent_labels,
    'real_labels': real_labels,
    'augment': augment,
    'ada_augment': None,
    'ada_aug_p': 0.6,
    'ada_aug_step': 500 * 1000,
}
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
G_logistic Ran
ran on a synthesized input
xuduo35/StyleFlowPytorch/stylegan2/loss_fns.py
code served (permissive licence) · get_code("d660eee0510ffc84")
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
from torch.nn import functional as F

torch.manual_seed(0)

# Define a minimal Generator model
class SimpleGenerator(nn.Module):
    def __init__(self, latent_dim=4, output_dim=3):
        super().__init__()
        self.fc = nn.Linear(latent_dim, output_dim)
    
    def forward(self, latents, labels=None):
        # latents shape: (batch_size, latent_dim)
        # returns fake images of shape (batch_size, output_dim)
        return self.fc(latents)

# Define a minimal Discriminator model
class SimpleDiscriminator(nn.Module):
    def __init__(self, input_dim=3):
        super().__init__()
        self.fc = nn.Linear(input_dim, 1)
    
    def forward(self, x, labels=None):
        # x shape: (batch_size, input_dim)
        # returns logits of shape (batch_size, 1)
        return self.fc(x)

# Define a minimal Augment model
class SimpleAugment(nn.Module):
    def __init__(self):
        super().__init__()
    
    def forward(self, x, ada_aug_p):
        # x shape: (batch_size, input_dim)
        # returns augmented x and some auxiliary output
        return x, None

# Construct the arguments
G = SimpleGenerator(latent_dim=4, output_dim=3)
D = SimpleDiscriminator(input_dim=3)
latents = torch.randn(2, 4)  # batch_size=2, latent_dim=4
latent_labels = None
augment = SimpleAugment()
ada_augment = None
ada_aug_p = 0.6
ada_aug_step = 500 * 1000

ARGS = {
    'G': G,
    'D': D,
    'latents': latents,
    'latent_labels': latent_labels,
    'augment': augment,
    'ada_augment': ada_augment,
    'ada_aug_p': ada_aug_p,
    'ada_aug_step': ada_aug_step,
}
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
G_logistic_ns Ran
ran on a synthesized input
xuduo35/StyleFlowPytorch/stylegan2/loss_fns.py
code served (permissive licence) · get_code("e50ad75aca391fdc")
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
from torch.nn import functional as F

torch.manual_seed(0)

# Define a minimal Generator model
class MinimalGenerator(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear = nn.Linear(4, 3)
    
    def forward(self, latents, labels=None):
        # Returns fake images with shape (batch, channels, height, width)
        batch_size = latents.shape[0]
        out = self.linear(latents)
        # Reshape to (batch, 3, 1, 1) to simulate image output
        return out.view(batch_size, 3, 1, 1)

# Define a minimal Discriminator model
class MinimalDiscriminator(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear = nn.Linear(3, 1)
    
    def forward(self, fakes, labels=None):
        # Takes fake images and returns scores
        batch_size = fakes.shape[0]
        # Flatten the images
        flattened = fakes.view(batch_size, -1)
        scores = self.linear(flattened)
        return scores

# Define a minimal Augmentation model
class MinimalAugment(nn.Module):
    def __init__(self):
        super().__init__()
    
    def forward(self, fakes, ada_aug_p):
        # Returns augmented fakes and augmentation metadata
        return fakes, None

# Create instances
G = MinimalGenerator()
D = MinimalDiscriminator()
augment = MinimalAugment()

# Create latents and labels
latents = torch.randn(2, 4)
latent_labels = torch.randint(0, 2, (2,))

ARGS = {
    'G': G,
    'D': D,
    'latents': latents,
    'latent_labels': latent_labels,
    'augment': augment,
    'ada_augment': None,
    'ada_aug_p': 0.6,
    'ada_aug_step': 500 * 1000,
}
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
bernoulli_log_likelihood Ran
ran on a synthesized input
xuduo35/StyleFlowPytorch/module/utils.py
code served (permissive licence) · get_code("19c5e2db00d325b8")
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

torch.manual_seed(0)

# Create small tensors for bernoulli log likelihood calculation
# x should be binary values (0 or 1)
x = torch.tensor([[1.0, 0.0, 1.0, 0.0],
                   [0.0, 1.0, 1.0, 0.0]], dtype=torch.float32)

# p should be probabilities between 0 and 1
p = torch.tensor([[0.8, 0.3, 0.9, 0.2],
                   [0.4, 0.7, 0.6, 0.1]], dtype=torch.float32)

# clip parameter is a boolean
clip = True

# eps is a small epsilon value to avoid log(0)
eps = 1e-06

ARGS = {
    'x': x,
    'p': p,
    'clip': clip,
    'eps': eps
}
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
count_parameters Ran
ran, honoured the contract we drafted
xuduo35/StyleFlowPytorch/module/flow.py
code served (permissive licence) · get_code("f6b944f50d3f15ae")
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 int, ndim 0, dtype family int, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn

torch.manual_seed(0)

class SimpleModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear1 = nn.Linear(4, 3)
        self.linear2 = nn.Linear(3, 2)
    
    def forward(self, x):
        x = self.linear1(x)
        x = self.linear2(x)
        return x

model = SimpleModel()

ARGS = {
    "model": model
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type int, ndim 0, dtype family int, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
gaussian_log_likelihood Ran
ran on a synthesized input
xuduo35/StyleFlowPytorch/module/utils.py
code served (permissive licence) · get_code("6306abd15a0062d5")
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 math import log, pi

torch.manual_seed(0)

# Create small tensors for the function
# x: observed data, shape (2, 4)
x = torch.randn(2, 4)

# mean: mean of the distribution, shape (2, 4)
mean = torch.randn(2, 4)

# logvar: log variance, shape (2, 4)
logvar = torch.randn(2, 4)

# clip: boolean flag
clip = True

ARGS = {
    "x": x,
    "mean": mean,
    "logvar": logvar,
    "clip": clip
}
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
get_activation Ran
ran on a synthesized input
xuduo35/StyleFlowPytorch/stylegan2/modules.py
code served (permissive licence) · get_code("1cb9925ee5079d88")
How it ran

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

import numpy as np
import torch
from torch import nn

class Swish(nn.Module):
    """
    Performs the 'Swish' non-linear activation function.
    https://arxiv.org/pdf/1710.05941.pdf
    Arguments:
        affine (bool): Multiply the input to sigmoid
            with a learnable scale. Default value is False.
    """

    def __init__(self, affine=False):
        super(Swish, self).__init__()
        if affine:
            self.beta = nn.Parameter(torch.tensor([1.0]))
        self.affine = affine

    def forward(self, input, *args, **kwargs):
        """
        Apply the swish non-linear activation function
        and return the results.
        Arguments:
            input (torch.Tensor)
        Returns:
            output (torch.Tensor)
        """
        x = input
        if self.affine:
            x *= self.beta
        return x * torch.sigmoid(x)

def get_activation(activation):
    """
    Get the module for a specific activation function and its gain if
    it can be calculated.
    Arguments:
        activation (str, callable, nn.Module): String representing the activation.
    Returns:
        activation_module (torch.nn.Module): The module representing
            the activation function.
        gain (float): The gain value. Defaults to 1 if it can not be calculated.
    """
    if isinstance(activation, nn.Module) or callable(activation):
        return (activation, 1.0)
    if isinstance(activation, str):
        activation = activation.lower()
    if activation in [None, 'linear']:
        return (nn.Identity(), 1.0)
    lrelu_strings = ('leaky', 'leakyrely', 'leaky_relu', 'leaky relu', 'lrelu')
    if activation.startswith(lrelu_strings):
        for l_s in lrelu_strings:
            activation = activation.replace(l_s, '')
        slope = ''.join((char for char in activation if char.isdigit() or char == '.'))
        slope = float(slope) if slope else 0.01
        return (nn.LeakyReLU(slope), np.sqrt(2))
    elif activation.startswith('swish'):
        return (Swish(affine=activation != 'swish'), np.sqrt(2))
    elif activation in ['relu']:
        return (nn.ReLU(), np.sqrt(2))
    elif activation in ['elu']:
        return (nn.ELU(), 1.0)
    elif activation in ['prelu']:
        return (nn.PReLU(), np.sqrt(2))
    elif activation in ['rrelu', 'randomrelu']:
        return (nn.RReLU(), np.sqrt(2))
    elif activation in ['selu']:
        return (nn.SELU(), 1.0)
    elif activation in ['softplus']:
        return (nn.Softplus(), 1)
    elif activation in ['softsign']:
        return (nn.Softsign(), 1)
    elif activation in ['sigmoid', 'logistic']:
        return (nn.Sigmoid(), 1.0)
    elif activation in ['tanh']:
        return (nn.Tanh(), 1.0)
    else:
        raise ValueError('Activation "{}" not available.'.format(activation))

torch.manual_seed(0)

ARGS = {
    'activation': 'relu'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 1, dtype family float — a tuple output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
kl_diagnormal_stdnormal Ran
ran on a synthesized input
xuduo35/StyleFlowPytorch/module/utils.py
code served (permissive licence) · get_code("ad16d0961b8183d2")
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

torch.manual_seed(0)

# Construct arguments for kl_diagnormal_stdnormal
# The function expects:
# - mean: a tensor of means
# - logvar: a tensor of log variances

mean = torch.randn(2, 4)
logvar = torch.randn(2, 4)

ARGS = {
    "mean": mean,
    "logvar": logvar,
}
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
stable_var Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
xuduo35/StyleFlowPytorch/module/normalization.py
pointer only (licence: MIT); the stored copy get_code serves, from deepgraphlearning/cgcf-confgen, is pointer only (licence: NONE) · get_code("27ca363b776a43f8")
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 deepgraphlearning/cgcf-confgen, 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.

count_nfe Not yet run
unverified (harvested, no recorded run)
xuduo35/StyleFlowPytorch/module/flow.py
code served (permissive licence) · get_code("6e88b4b7a2d8cc87")
How it ran

no invocation recorded.

count_total_time Not yet run
unverified (harvested, no recorded run)
xuduo35/StyleFlowPytorch/module/flow.py
code served (permissive licence) · get_code("ce228183dec2a0fc")
How it ran

no invocation recorded.

divergence_approx Not yet run
unverified (harvested, no recorded run)
xuduo35/StyleFlowPytorch/module/odefunc.py
pointer only (licence: MIT); the stored copy get_code serves, from luost26/CGCF-ConfGen, is pointer only (licence: NONE) · get_code("58625622277c98f5")
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 luost26/CGCF-ConfGen, 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. No invocation recorded.

image_align Not yet run
unverified (harvested, no recorded run)
xuduo35/StyleFlowPytorch/ffhq_dataset/face_alignment.py
code served (permissive licence) · get_code("c68ee58b2a736123")
How it ran

no invocation recorded.

load Not yet run
unverified (harvested, no recorded run)
xuduo35/StyleFlowPytorch/stylegan2/models.py
code served (permissive licence) · get_code("24150bba4666d3f0")
How it ran

no invocation recorded.

Also present in other papers' code

Of 14 samples lifted from this paper's linked repositories, 14 carry a function identity and 0 do not. Of the 14 distinct functions with an identity, 8 are also present in other papers whose code we hold (8 with at least one paper dated earlier, 3 the same month, 3 later) and 6 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
count_parameters 180 other papers
22 dated earlier, 2 the same month, 156 later · 8 only as an assembly brick
1705.08415 (2017-05)
Supervised Community Detection with Line Graph Neural Networks
1707.01926 (2017-07)
Diffusion Convolutional Recurrent Neural Network: Data-Driven Traffic Forecasting
1808.09367 (2018-08)
Deriving Machine Attention from Human Rationales
and 177 more not named here
divergence_approx 3 other papers
1 dated earlier, 1 the same month, 1 later
1906.12320 (2019-06)
PointFlow: 3D Point Cloud Generation with Continuous Normalizing Flows
2008.02792 (2020-08)
CaSPR: Learning Canonical Spatiotemporal Point Cloud Representations
2102.10240 (2021-02)
Learning Neural Generative Dynamics for Molecular Conformation Generation
count_total_time 2 other papers
1 dated earlier, 1 the same month
1906.12320 (2019-06)
PointFlow: 3D Point Cloud Generation with Continuous Normalizing Flows
2008.02792 (2020-08)
CaSPR: Learning Canonical Spatiotemporal Point Cloud Representations
stable_var 2 other papers
1 dated earlier, 1 later
1906.12320 (2019-06)
PointFlow: 3D Point Cloud Generation with Continuous Normalizing Flows
2102.10240 (2021-02)
Learning Neural Generative Dynamics for Molecular Conformation Generation
bernoulli_log_likelihood 1 other paper
1 dated earlier
1906.12320 (2019-06)
PointFlow: 3D Point Cloud Generation with Continuous Normalizing Flows
count_nfe 1 other paper
1 dated earlier
1906.12320 (2019-06)
PointFlow: 3D Point Cloud Generation with Continuous Normalizing Flows
gaussian_log_likelihood 1 other paper
1 dated earlier
1906.12320 (2019-06)
PointFlow: 3D Point Cloud Generation with Continuous Normalizing Flows
kl_diagnormal_stdnormal 1 other paper
1 dated earlier
1906.12320 (2019-06)
PointFlow: 3D Point Cloud Generation with Continuous Normalizing Flows

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 2026-11-21.

Repositories linked to this paper

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

Abstract

High-quality, diverse, and photorealistic images can now be generated by unconditional GANs (e.g., StyleGAN). However, limited options exist to control the generation process using (semantic) attributes, while still preserving the quality of the output. Further, due to the entangled nature of the GAN latent space, performing edits along one attribute can easily result in unwanted changes along other attributes. In this paper, in the context of conditional exploration of entangled latent spaces, we investigate the two sub-problems of attribute-conditioned sampling and attribute-controlled editing. We present StyleFlow as a simple, effective, and robust solution to both the sub-problems by formulating conditional exploration as an instance of conditional continuous normalizing flows in the GAN latent space conditioned by attribute features. We evaluate our method using the face and the car latent space of StyleGAN, and demonstrate fine-grained disentangled edits along various attributes on both real photographs and StyleGAN generated images. For example, for faces, we vary camera pose, illumination variation, expression, facial hair, gender, and age. Finally, via extensive qualitative and quantitative comparisons, we demonstrate the superiority of StyleFlow to other concurrent works.

For agents

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

get_harvested_code_for_paper("2008.02401")
get_code_for_paper("2008.02401")
have("2008.02401")

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

curl https://syntology.ai/api/ran/2008.02401.json

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

[![Syntology run record](https://syntology.ai/api/ran/2008.02401.svg)](https://syntology.ai/paper/2008.02401)

Connect an agent — have() is free.