SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2203.07293 · CVPR · 2022

InsetGAN for Full-Body Image Generation

Niloy Mitra, Jingwan Lu, Eli Shechtman, Peter Wonka, Krishna Singh, Anna Frühstück

arXiv · PDF · Open in the Atlas

Code that ran

6 of 6 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; 0 honoured the contract we drafted, 0 violated it and 6 have no contract verdict; 0 failures of our instrument; 0 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/2203.07293.json · badge: /api/ran/2203.07293.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
afruehstueck/insetGAN canonical (classifier label)
classifier basis not recorded on the link
6 ran · 0 no recorded run
FunctionStatusWhere it lives
l1_loss Ran
ran on a synthesized input
afruehstueck/insetGAN/run_insetgan.py
code served (permissive licence) · get_code("758b324c800925d0")
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. The code mentions cuda, and the harness aliases cuda to the CPU for every sample; whether this run needed that is 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)

# Create two tensors with the same shape
# The function uses: target.shape, optim.shape, and computes loss between them
# Shape: (batch=2, features=4)
target = torch.randn(2, 4)
optim = torch.randn(2, 4)

ARGS = {
    'target': target,
    'optim': optim
}
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
l2_loss Ran
ran on a synthesized input
afruehstueck/insetGAN/run_insetgan.py
code served (permissive licence) · get_code("713bed19c0b64133")
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. The code mentions cuda, and the harness aliases cuda to the CPU for every sample; whether this run needed that is 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)

# The function l2_loss takes two parameters: target and optim
# - target: used with .shape[-1], passed to loss_L2 (MSELoss)
# - optim: passed to loss_L2 (MSELoss)
# Both are tensors that MSELoss expects (same shape, numeric values)

# Create small tensors on CPU
batch_size = 2
dim = 4

target = torch.randn(batch_size, dim)
optim = torch.randn(batch_size, dim)

ARGS = {
    'target': target,
    'optim': optim
}
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
rgb2gray Ran
ran on a synthesized input
afruehstueck/insetGAN/run_insetgan.py
code served (permissive licence) · get_code("fefc5af0ccab2041")
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, 1, 4, 5) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# Create a small RGB image tensor
# Shape: (batch_size, channels, height, width)
# batch=2, channels=3 (RGB), height=4, width=5
rgb = torch.randn(2, 3, 4, 5)

ARGS = {
    'rgb': rgb
}
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, 1, 4, 5) — shape (2, 1, 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
show_tensor Ran
ran on a synthesized input
afruehstueck/insetGAN/optim_utils.py
code served (permissive licence) · get_code("6b2e1f58d09fbe73")
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 3, dtype family int, shape (4, 5, 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
import numpy as np
import matplotlib.pyplot as plt

torch.manual_seed(0)

# Create a small test tensor
# Shape: (3, 4, 5) - 3 channels, 4x5 spatial dimensions
# Values in range [-1, 1] as expected by the normalize logic
tensor = torch.randn(3, 4, 5)

# Create matplotlib axis for testing
fig, ax = plt.subplots(figsize=(6, 6))

ARGS = {
    'tensor': tensor,
    'ax': ax,
    'normalize': True,
    'text': 'Test image',
    'color': 'black',
    'size': 22
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
matplotlib, numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 3, dtype family int, shape (4, 5, 3) — shape (4, 5, 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
show_tensors Ran
ran on a synthesized input
afruehstueck/insetGAN/optim_utils.py
code served (permissive licence) · get_code("d07e9569906994e9")
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 3, dtype family int, shape (8, 3, 1) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import matplotlib.pyplot as plt
import numpy as np

torch.manual_seed(0)

# Create a small batch of tensors
# Shape: (batch_size=2, height=4, width=3) - 3D tensor
tensors = torch.randn(2, 4, 3)

# Grid dimensions
gh = 2  # grid height
gw = 1  # grid width

# Optional matplotlib axis
ax = None

# Normalize flag
normalize = True

# Title
title = "Test Image Grid"

ARGS = {
    "tensors": tensors,
    "gh": gh,
    "gw": gw,
    "ax": ax,
    "normalize": normalize,
    "title": title,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
matplotlib, numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 3, dtype family int, shape (8, 3, 1) — shape (8, 3, 1) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
tensor_to_int Ran
ran on a synthesized input
afruehstueck/insetGAN/optim_utils.py
code served (permissive licence) · get_code("60e1ec98bec7da79")
How it ran

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

import torch

torch.manual_seed(0)

# The function takes a tensor and performs denormalization
# It expects a tensor with values in range [-1, 1] and converts to [0, 255]
# The tensor needs .clone() method, so it must be a torch.Tensor

# Create a small tensor with values in the expected range [-1, 1]
tensor = torch.randn(2, 4) * 0.5  # Creates small random values roughly in [-1, 1] range

ARGS = {
    "tensor": tensor
}
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 (2, 4) — shape (2, 4) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block

Also present in other papers' code

Of 6 samples lifted from this paper's linked repositories, 6 carry a function identity and 0 do not. Of the 6 distinct functions with an identity, 0 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 0 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.

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

Figure 1. InsetGAN applications. Our full-body human generator is able to generate reasonable bodies at state-of-the-art resolution (1024×1024px) (a). However, some artifacts appear in the synthesized results, most visibly in extremities and faces. We make use of a second, specialized generator to seamlessly improve the face region (b). We can also use a given face as an input for unconditional generation of bodies (c). Furthermore, we can select both specific faces and bodies and compose them in a seamlessly merged output (d).

For agents

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

get_harvested_code_for_paper("2203.07293")
get_code_for_paper("2203.07293")
have("2203.07293")

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

curl https://syntology.ai/api/ran/2203.07293.json

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

[![Syntology run record](https://syntology.ai/api/ran/2203.07293.svg)](https://syntology.ai/paper/2203.07293)

Connect an agent — have() is free.