SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1606.05328 · 2016

Conditional Image Generation with PixelCNN Decoders

arXiv · PDF · Open in the Atlas

Code that ran

8 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; 2 honoured the contract we drafted, 0 violated it and 4 have no contract verdict; 2 failures of our instrument; 6 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/1606.05328.json · badge: /api/ran/1606.05328.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
eyalbetzalel/pytorch-generative community (classifier label)
classifier basis not recorded on the link
4 ran · 0 no recorded run
kamenbliznashki/pixel_models reimplementation (classifier label)
classifier basis: classifier_2025
2 ran · 1 no recorded run
kkleidal/gatedpixelcnnpytorch reimplementation (classifier label)
classifier basis: classifier_2025
1 ran · 3 no recorded run
chandlersupple/Tensorflow-PixelCNN reimplementation (classifier label)
classifier basis: classifier_2025
1 ran · 0 no recorded run
EugenHotaj/pytorch-generative reimplementation (classifier label)
classifier basis: classifier_2025
0 ran · 2 no recorded run
FunctionStatusWhere it lives
animate Ran
ran on a synthesized input
eyalbetzalel/pytorch-generative/pytorch_generative/colab_utils.py
code served (permissive licence) · get_code("7486d39d659887c7")
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 FuncAnimation. 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 matplotlib.pyplot as plt
from torchvision import transforms

torch.manual_seed(0)

# Create a list of small image tensors
# Images should be in format (C, H, W) for transforms.ToPILImage() to work
# Each tensor needs to be 3D: (channels, height, width)
frames = [
    torch.rand(3, 4, 5),  # 3 channels, 4 height, 5 width
    torch.rand(3, 4, 5),
    torch.rand(3, 4, 5),
]

ARGS = {
    "frames": frames,
    "figsize": (6, 4),
    "fps": 24,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
matplotlib, torch, torchvision — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type FuncAnimation — a FuncAnimation output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a FuncAnimation 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
compute_receptive_field Ran
ran on a synthesized input
eyalbetzalel/pytorch-generative/pytorch_generative/debug.py
code served (permissive licence) · get_code("54d26252e3d6b564")
How it ran

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

import torch
import torch.nn as nn

torch.manual_seed(0)

# Build a minimal model that satisfies the function's usage
class SimpleConvModel(nn.Module):
    def __init__(self):
        super().__init__()
        # Use a simple conv layer that preserves spatial dimensions
        self.conv = nn.Conv2d(3, 1, kernel_size=3, padding=1)
    
    def forward(self, x):
        # x shape: (batch, channels, height, width)
        # Output shape should be (batch, channels, height, width)
        return self.conv(x)

model = SimpleConvModel()

ARGS = {
    "model": model,
    "img_size": (3, 5, 5),
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 2, dtype family float, shape (5, 5) — shape (5, 5) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
image_positional_encoding Ran
ran on a synthesized input
eyalbetzalel/pytorch-generative/pytorch_generative/nn.py
code served (permissive licence) · get_code("97ff1e90a6a32884")
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, 8, 3, 5) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import functools

torch.manual_seed(0)

# The function takes a shape parameter which is expected to be a tuple of (N, C, H, W)
# Based on the function body:
# - N: batch size
# - C: number of channels
# - H: height
# - W: width

# The function uses these values to:
# 1. Unpack the shape tuple: n, c, h, w = shape
# 2. Create a zeros tensor of shape (n, c, h, w)
# 3. Create arange tensors for height and width dimensions
# 4. Concatenate them along dimension 1

# Using distinct sizes for different dimensions to allow probing
N = 2  # batch size
C = 4  # channels
H = 3  # height
W = 5  # width

shape = (N, C, H, W)

ARGS = {
    "shape": shape
}
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, 8, 3, 5) — shape (2, 8, 3, 5) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
load_image Ran
ran on a synthesized input
eyalbetzalel/pytorch-generative/pytorch_generative/colab_utils.py
code served (permissive licence) · get_code("6e93d4d24f711246")
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 3, dtype family float, shape (3, 3, 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 tempfile
import PIL
import torch
from torchvision import transforms

torch.manual_seed(0)

# Create a temporary test image file
with tempfile.NamedTemporaryFile(suffix='.png', delete=False) as tmp:
    # Create a simple RGB image (3x4 pixels)
    img = PIL.Image.new('RGB', (4, 3), color=(255, 0, 0))
    img.save(tmp.name)
    temp_image_path = tmp.name

ARGS = {
    'path': temp_image_path,
    'size': None,
    'remove_alpha_channel': True
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
PIL, torch, torchvision — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 3, dtype family float, shape (3, 3, 4) — shape (3, 3, 4) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
loss_fn Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
kamenbliznashki/pixel_models/pixelcnn.py
pointer only (licence: NONE) · get_code("bd3e6d44a59d468d")
How it ran

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

mask Ran
ran; our contract draft was wrong, not the code
chandlersupple/Tensorflow-PixelCNN/pixelcnn.py
code served (permissive licence) · get_code("1e14e324863c2ccd")
How it ran

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

import numpy as np

ARGS = {
    'mask_type': 'v',
    'num_inputs': 4,
    'num_outputs': 3,
    'kernel_size': (5, 5),
    'num_channels': 2
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 4, dtype family float, shape (5, 5, 4, 3) — shape (5, 5, 4, 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
pixelcnn_gate Ran
ran, honoured the contract we drafted
kamenbliznashki/pixel_models/pixelcnn.py
pointer only (licence: NONE) · get_code("68cdfbf37a172d13")
How it ran

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

same_padding Ran
ran, honoured the contract we drafted
kkleidal/gatedpixelcnnpytorch/models/components/pixelcnn.py
code served (permissive licence) · get_code("b2d883de2fd93619")
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)

# Test case 1: kernel_size as an integer
kernel_size_int = 5

# Test case 2: kernel_size as a tuple
kernel_size_tuple = (5, 3)

# We'll create ARGS with the integer case as the primary test
ARGS = {
    "kernel_size": 5
}
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
generate_fn Not yet run
unverified (harvested, no recorded run)
kamenbliznashki/pixel_models/pixelcnn.py
pointer only (licence: NONE) · get_code("a88f8690106036a3")
How it ran

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

negative_log_likelihood Not yet run
unverified (harvested, no recorded run)
kkleidal/gatedpixelcnnpytorch/mnist_pixelcnn_train.py
code served (permissive licence) · get_code("e1fa075299151e23")
How it ran

no invocation recorded.

preprocess Not yet run
unverified (harvested, no recorded run)
kkleidal/gatedpixelcnnpytorch/mnist_pixelcnn_train.py
code served (permissive licence) · get_code("48694d3cb7b430fb")
How it ran

no invocation recorded.

quantisize Not yet run
unverified (harvested, no recorded run)
kkleidal/gatedpixelcnnpytorch/mnist_pixelcnn_train.py
code served (permissive licence) · get_code("1c783a15bde67d51")
How it ran

no invocation recorded.

reproduce Not yet run
unverified (harvested, no recorded run)
EugenHotaj/pytorch-generative/pytorch_generative/models/autoregressive/gated_pixel_cnn.py
code served (permissive licence) · get_code("f5e294f7c52471d8")
How it ran

no invocation recorded.

reproduce Not yet run
unverified (harvested, no recorded run)
EugenHotaj/pytorch-generative/pytorch_generative/models/autoregressive/pixel_cnn.py
code served (permissive licence) · get_code("1eef18f20b4c3e7d")
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, 2 are also present in other papers whose code we hold (1 with at least one paper dated earlier, 0 the same month, 1 later) and 12 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
quantisize 2 other papers
2 later
1609.03499 (2016-09)
WAVENET: A GENERATIVE MODEL FOR RAW AUDIO Aäron van den Oord
2310.09604 (2023-10)
Learning Hierarchical Features with Joint Latent Space Energy-Based Prior
pixelcnn_gate 1 other paper
1 dated earlier
1601.06759 (2016-01)
Pixel Recurrent Neural Networks Aäron van den Oord

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

This work explores conditional image generation with a new image density model based on the PixelCNN architecture. The model can be conditioned on any vector, including descriptive labels or tags, or latent embeddings created by other networks. When conditioned on class labels from the ImageNet database, the model is able to generate diverse, realistic scenes representing distinct animals, objects, landscapes and structures. When conditioned on an embedding produced by a convolutional network given a single image of an unseen face, it generates a variety of new portraits of the same person with different facial expressions, poses and lighting conditions. We also show that conditional PixelCNN can serve as a powerful decoder in an image autoencoder. Additionally, the gated convolutional layers in the proposed model improve the log-likelihood of PixelCNN to match the state-of-the-art performance of PixelRNN on ImageNet, with greatly reduced computational cost.

For agents

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

get_harvested_code_for_paper("1606.05328")
get_code_for_paper("1606.05328")
have("1606.05328")

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

curl https://syntology.ai/api/ran/1606.05328.json

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

[![Syntology run record](https://syntology.ai/api/ran/1606.05328.svg)](https://syntology.ai/paper/1606.05328)

Connect an agent — have() is free.