SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2110.08787 · ICCV · 2021

PixelPyramids: Exact Inference Models from Lossless Image Pyramids

Shweta Mahajan, Stefan Roth

arXiv · PDF · Open in the Atlas

Code that ran

11 of 11 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (8 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 8 have no contract verdict; 3 failures of our instrument; 0 no recorded run), measured 2026-11-22; 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/2110.08787.json · badge: /api/ran/2110.08787.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
visinf/pixelpyramids community
Papers with Code archive
8 ran · 0 no recorded run
copy not recorded community
not in the paper's declared repositories
3 ran · 0 no recorded run
FunctionStatusWhere it lives
Conv2d Ran
ran on a synthesized input · constructed an object; no result was computed
visinf/pixelpyramids/unet/pixelcnnpp.py
code served (permissive licence) · get_code("6387ca84ef0e560e")
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, Conv2d: 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)

# Conv2d inherits from nn.Conv2d and calls nn.utils.weight_norm(self)
# nn.Conv2d requires: in_channels, out_channels, kernel_size
# Optional parameters: stride, padding, dilation, groups, bias, padding_mode, device, dtype

ARGS = {
    'in_channels': 3,
    'out_channels': 4,
    'kernel_size': 3,
    'stride': 1,
    'padding': 0,
    'dilation': 1,
    'groups': 1,
    'bias': True,
    'padding_mode': 'zeros',
    'device': None,
    'dtype': None,
}
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 — 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 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 on a synthesized input · constructed an object; no result was computed
visinf/pixelpyramids/unet/pixelcnnpp.py
code served (permissive licence) · get_code("c0d558a50b50d636")
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, ConvTranspose2d: 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 = {
    'in_channels': 3,
    'out_channels': 16,
    '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 — 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 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
DownRightShiftedConv2d Ran
ran on a synthesized input · constructed an object; no result was computed
visinf/pixelpyramids/unet/pixelcnnpp.py
code served (permissive licence) · get_code("b26a0b96c0a3af73")
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, DownRightShiftedConv2d: 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)

# DownRightShiftedConv2d inherits from Conv2d which inherits from nn.Conv2d
# The __init__ takes all standard Conv2d arguments
# Looking at the forward method, it uses self.kernel_size to get Hk and Wk
# Then it pads x and calls super().forward(x)
# So we need valid Conv2d initialization parameters

ARGS = {
    "in_channels": 3,
    "out_channels": 16,
    "kernel_size": (3, 3),
    "stride": 1,
    "padding": 0,
    "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 DownRightShiftedConv2d — 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 DownRightShiftedConv2d 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
DownRightShiftedConvTranspose2d Ran
ran on a synthesized input · constructed an object; no result was computed
visinf/pixelpyramids/unet/pixelcnnpp.py
code served (permissive licence) · get_code("da898124203f12d6")
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, DownRightShiftedConvTranspose2d: 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)

# DownRightShiftedConvTranspose2d inherits from ConvTranspose2d which inherits from nn.ConvTranspose2d
# The forward method takes x (a tensor) and calls super().forward(x)
# It then slices the output based on kernel_size, stride, and output shape

# For nn.ConvTranspose2d, required positional args are: in_channels, out_channels, kernel_size
# Required or optional args: stride, padding, output_padding, groups, bias, dilation, padding_mode

ARGS = {
    "in_channels": 3,
    "out_channels": 4,
    "kernel_size": (3, 5),
    "stride": (2, 2),
    "padding": (0, 0),
    "output_padding": (0, 0),
    "groups": 1,
    "bias": True,
    "dilation": (1, 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 DownRightShiftedConvTranspose2d — 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 DownRightShiftedConvTranspose2d 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
DownShiftedConv2d Ran
ran on a synthesized input · constructed an object; no result was computed
visinf/pixelpyramids/unet/pixelcnnpp.py
code served (permissive licence) · get_code("15071eacd675873d")
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, DownShiftedConv2d: 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 = {
    'in_channels': 3,
    'out_channels': 16,
    'kernel_size': (3, 3),
    'stride': 1,
    'padding': 0,
    '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 DownShiftedConv2d — 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 DownShiftedConv2d 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
DownShiftedConvTranspose2d Ran
ran on a synthesized input · constructed an object; no result was computed
visinf/pixelpyramids/unet/pixelcnnpp.py
code served (permissive licence) · get_code("bf8af0a07b2898f7")
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, DownShiftedConvTranspose2d: 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 = {
    "in_channels": 4,
    "out_channels": 3,
    "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 DownShiftedConvTranspose2d — 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 DownShiftedConvTranspose2d 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
GatedResidualLayer Ran
ran on a synthesized input · constructed an object; no result was computed
visinf/pixelpyramids/unet/pixelcnnpp.py
code served (permissive licence) · get_code("9bf0b7848adc2035")
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, GatedResidualLayer: 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)

class SimpleConv(nn.Module):
    def __init__(self, in_channels, out_channels, kernel_size):
        super().__init__()
        self.conv = nn.Conv2d(in_channels, out_channels, kernel_size, padding=kernel_size//2)
    
    def forward(self, x):
        return self.conv(x)

ARGS = {
    'conv': SimpleConv,
    'n_channels': 4,
    'kernel_size': 3,
    'drop_rate': 0.1,
    'shortcut_channels': 2,
    'n_cond_classes': 5,
    'relu_fn': None
}
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 GatedResidualLayer — 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 GatedResidualLayer 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
PixelCNNpp Ran
ran on a synthesized input · constructed an object; no result was computed
visinf/pixelpyramids/unet/pixelcnnpp.py
code served (permissive licence) · get_code("5e5e8e9bdd22d08b")
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, PixelCNNpp: 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 = {
    "image_dims": (3, 4, 4),
    "n_channels": 32,
    "n_res_layers": 2,
    "n_classes": 10,
    "n_cond_classes": None,
    "drop_rate": 0.0,
}
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 PixelCNNpp — 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 PixelCNNpp 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
concat_elu Ran
ran; our contract draft was wrong, not the code
this paper's copy was not recorded; identical code first harvested from kamenbliznashki/pixel_models
pointer only; the stored copy get_code serves, from kamenbliznashki/pixel_models, is pointer only (licence: NONE) · get_code("1bc8478089c8a7c9")
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.

down_shift Ran
ran; our contract draft was wrong, not the code
this paper's copy was not recorded; identical code first harvested from kamenbliznashki/pixel_models
pointer only; the stored copy get_code serves, from kamenbliznashki/pixel_models, is pointer only (licence: NONE) · get_code("a65cc793ef279a03")
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.

right_shift Ran
ran; our contract draft was wrong, not the code
this paper's copy was not recorded; identical code first harvested from kamenbliznashki/pixel_models
pointer only; the stored copy get_code serves, from kamenbliznashki/pixel_models, is pointer only (licence: NONE) · get_code("1074bb941737a16e")
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.

Also present in other papers' code

Of 11 samples lifted from this paper's linked repositories, 11 carry a function identity and 0 do not. Of the 11 distinct functions with an identity, 3 are also present in other papers whose code we hold (3 with at least one paper dated earlier, 0 the same month, 0 later) and 8 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
concat_elu
brick of this paper's assembly: helper
2 other papers
2 dated earlier
1701.05517 (2017-01)
PixelCNN++: Improving the PixelCNN with Discretized Logistic Mixture Likelihood and Other Modifications
1712.09763 (2017-12)
PixelSNAIL: An Improved Autoregressive Generative Model
down_shift
brick of this paper's assembly: helper
2 other papers
2 dated earlier
1701.05517 (2017-01)
PixelCNN++: Improving the PixelCNN with Discretized Logistic Mixture Likelihood and Other Modifications
1712.09763 (2017-12)
PixelSNAIL: An Improved Autoregressive Generative Model
right_shift
brick of this paper's assembly: helper
2 other papers
2 dated earlier
1701.05517 (2017-01)
PixelCNN++: Improving the PixelCNN with Discretized Logistic Mixture Likelihood and Other Modifications
1712.09763 (2017-12)
PixelSNAIL: An Improved Autoregressive Generative Model

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

Autoregressive models are a class of exact inference approaches with highly flexible functional forms, yielding state-of-the-art density estimates for natural images. Yet, the sequential ordering on the dimensions makes these models computationally expensive and limits their applicability to low-resolution imagery. In this work, we propose Pixel-Pyramids, 1 a block-autoregressive approach employing a lossless pyramid decomposition with scale-specific representations to encode the joint distribution of image pixels. Crucially, it affords a sparser dependency structure compared to fully autoregressive approaches. Our PixelPyramids yield state-of-the-art results for density estimation on various image datasets, especially for high-resolution data. For CelebA-HQ 1024 × 1024, we observe that the density estimates (in terms of bits/dim) are improved to ∼44 % of the baseline despite sampling speeds superior even to easily parallelizable flow-based models.

For agents

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

get_harvested_code_for_paper("2110.08787")
get_code_for_paper("2110.08787")
have("2110.08787")

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

curl https://syntology.ai/api/ran/2110.08787.json

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

[![Syntology run record](https://syntology.ai/api/ran/2110.08787.svg)](https://syntology.ai/paper/2110.08787)

Connect an agent — have() is free.