SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2007.02442 · 2020

GRAF: Generative Radiance Fields for 3D-Aware Image Synthesis

arXiv · PDF · Open in the Atlas

Code that ran

10 of 12 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, 1 violated it and 6 have no contract verdict; 2 failures of our instrument; 2 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/2007.02442.json · badge: /api/ran/2007.02442.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
autonomousvision/graf canonical (classifier label)
classifier basis: census
10 ran · 2 no recorded run
FunctionStatusWhere it lives
color_depth_map Ran
ran on a synthesized input
autonomousvision/graf/graf/utils.py
code served (permissive licence) · get_code("71ec2a9308fa61f3")
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, 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 numpy as np

# Create a small depth map (HxW numpy array)
# Using distinct dimensions: 4 height, 3 width
depths = np.random.rand(4, 3) * 10.0

# Create ARGS dictionary with all parameters
ARGS = {
    "depths": depths,
    "scale": None
}
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 3, dtype family int, shape (4, 3, 3) — shape (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
count_trainable_parameters Ran
ran on a synthesized input
autonomousvision/graf/graf/utils.py
code served (permissive licence) · get_code("7de1ceff1fdc469d")
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 int64, 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
import numpy as np

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, 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 int64, 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
get_nsamples Ran
ran on a synthesized input
autonomousvision/graf/graf/utils.py
code served (permissive licence) · get_code("4cbceee1eebaf6d4")
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, 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)

# Create a simple data loader - a list of tensors
# The function calls next(iter(data_loader)) and expects tensors with .size(0)
data_loader = [
    torch.randn(2, 4),  # batch size 2, feature dim 4
    torch.randn(3, 4),  # batch size 3, feature dim 4
    torch.randn(2, 4),  # batch size 2, feature dim 4
]

N = 5  # Request 5 samples total

ARGS = {
    "data_loader": data_loader,
    "N": N,
}
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, 4) — shape (5, 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
load_blender_data Ran
ran on a synthesized input
autonomousvision/graf/submodules/nerf_pytorch/load_blender.py
code served (permissive licence) · get_code("83ed66ae61279dbc")
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 os
import torch
import numpy as np
import tempfile
import json
from PIL import Image

torch.manual_seed(0)
np.random.seed(0)

# Create a temporary directory structure with minimal Blender dataset
tmpdir = tempfile.mkdtemp()

# Create transforms files for train, val, test splits
for split in ['train', 'val', 'test']:
    transforms = {
        'camera_angle_x': 0.6911112070083618,
        'frames': [
            {
                'file_path': f'{split}_frame_0',
                'transform_matrix': np.eye(4).tolist()
            },
            {
                'file_path': f'{split}_frame_1',
                'transform_matrix': np.eye(4).tolist()
            }
        ]
    }
    with open(os.path.join(tmpdir, f'transforms_{split}.json'), 'w') as f:
        json.dump(transforms, f)

# Create minimal PNG images (4-channel RGBA)
for split in ['train', 'val', 'test']:
    for frame_idx in range(2):
        img_array = np.random.randint(0, 256, (64, 64, 4), dtype=np.uint8)
        img_path = os.path.join(tmpdir, f'{split}_frame_{frame_idx}.png')
        Image.fromarray(img_array).save(img_path)

ARGS = {
    'basedir': tmpdir,
    'half_res': False,
    'testskip': 1,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
PIL, cv2, imageio, numpy, 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
load_dv_data Ran
ran on a synthesized input
autonomousvision/graf/submodules/nerf_pytorch/load_deepvoxels.py
code served (permissive licence) · get_code("e69746042397f344")
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 os
import tempfile
import numpy as np
import imageio

# Create a temporary directory structure that mimics the deepvoxels dataset
temp_dir = tempfile.mkdtemp()
basedir = temp_dir

# Create directory structure
scene = 'cube'
train_dir = os.path.join(basedir, 'train', scene)
test_dir = os.path.join(basedir, 'test', scene)
val_dir = os.path.join(basedir, 'validation', scene)

for d in [train_dir, test_dir, val_dir]:
    os.makedirs(os.path.join(d, 'rgb'), exist_ok=True)
    os.makedirs(os.path.join(d, 'pose'), exist_ok=True)

# Create intrinsics.txt
intrinsics_content = """500.0 256.0 256.0
0.0 0.0 0.0
0.1
1.0
512.0 512.0
0
"""
with open(os.path.join(train_dir, 'intrinsics.txt'), 'w') as f:
    f.write(intrinsics_content)

# Create pose files (4x4 matrices as text)
pose_data = np.eye(4).astype(np.float32)
for split_dir in [train_dir, test_dir, val_dir]:
    for i in range(3):
        pose_file = os.path.join(split_dir, 'pose', f'pose_{i:03d}.txt')
        with open(pose_file, 'w') as f:
            f.write(' '.join(str(x) for x in pose_data.flatten()))

# Create dummy RGB images (512x512x3)
dummy_img = (np.random.rand(512, 512, 3) * 255).astype(np.uint8)
for split_dir in [train_dir, test_dir, val_dir]:
    for i in range(3):
        img_file = os.path.join(split_dir, 'rgb', f'image_{i:03d}.png')
        imageio.imwrite(img_file, dummy_img)

ARGS = {
    'scene': 'cube',
    'basedir': basedir,
    'testskip': 2
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
imageio, numpy — 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
normalize Ran
ran, violated the contract we drafted
autonomousvision/graf/submodules/nerf_pytorch/load_llff.py
code served (permissive licence) · get_code("1d6f8390d799f055")
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 1, dtype family float, shape (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 numpy as np

# The normalize function takes a single parameter 'x' which is expected to be
# a numpy array (or array-like). It divides x by its L2 norm.
# We need to provide a small numpy array as input.

x = np.array([3.0, 4.0])  # Simple 2D vector with norm=5, will normalize to [0.6, 0.8]

ARGS = {
    "x": x
}
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 1, dtype family float, shape (2,) — shape (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: none by the harness; any seed is in the setup block
pose_spherical Ran
ran, honoured the contract we drafted
autonomousvision/graf/submodules/nerf_pytorch/load_blender.py
pointer only (licence: MIT); the stored copy get_code serves, from runsong123/pcf-lift, is pointer only (licence: NONE) · get_code("8313c1fb69b28c8e")
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 runsong123/pcf-lift, 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.

ptstocam Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
autonomousvision/graf/submodules/nerf_pytorch/load_llff.py
code served (permissive licence) · get_code("36f5de7a0aae4a9b")
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 2, dtype family float, shape (2, 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

# Construct valid arguments for ptstocam function
# The function expects:
# - pts: points array that will be used in (pts - c2w[:3, 3])
# - c2w: camera-to-world transformation matrix with shape (4, 4) or similar
#        where c2w[:3, :3] is the rotation part and c2w[:3, 3] is the translation

# Create a c2w matrix (4x4 camera-to-world transformation)
c2w = np.array([
    [1.0, 0.0, 0.0, 0.5],
    [0.0, 1.0, 0.0, 1.0],
    [0.0, 0.0, 1.0, 2.0],
    [0.0, 0.0, 0.0, 1.0]
], dtype=np.float32)

# Create points array with shape (2, 3) - 2 points with 3 coordinates each
pts = np.array([
    [1.0, 2.0, 3.0],
    [4.0, 5.0, 6.0]
], dtype=np.float32)

ARGS = {
    "pts": pts,
    "c2w": c2w
}
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 2, dtype family float, shape (2, 3) — shape (2, 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
update_config Ran
ran on a synthesized input
autonomousvision/graf/graf/config.py
code served (permissive licence) · get_code("559fa10dfa807ea5")
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 dict. 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)

# Construct a nested config dict with various types
config = {
    'model': {
        'hidden_size': 128,
        'num_layers': 3,
        'dropout': 0.5,
        'use_attention': True,
        'optimizer': None
    },
    'learning_rate': 0.001,
    'batch_size': 32,
    'epochs': 100,
    'use_gpu': False
}

# Construct unknown arguments list with both nested (k1:k2) and flat (k) style arguments
unknown = [
    '--model:hidden_size', '256',
    '--model:num_layers', '5',
    '--model:use_attention', 'false',
    '--learning_rate', '0.0005',
    '--epochs', '200'
]

ARGS = {
    'config': config,
    'unknown': unknown
}
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 dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict 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
viewmatrix Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
autonomousvision/graf/submodules/nerf_pytorch/load_llff.py
code served (permissive licence) · get_code("aa52acee18f54d38")
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 2, dtype family float, shape (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 numpy as np

# Construct arguments for viewmatrix function
# The function expects:
# - z: a vector (used in normalize and cross product)
# - up: a vector (used in cross product)
# - pos: a vector (stacked into result)

# Create small 3D vectors as numpy arrays
z = np.array([0.0, 0.0, 1.0])
up = np.array([0.0, 1.0, 0.0])
pos = np.array([1.0, 2.0, 3.0])

ARGS = {
    "z": z,
    "up": up,
    "pos": pos
}
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 2, dtype family float, shape (3, 4) — shape (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
get_data Not yet run
unverified (harvested, no recorded run)
autonomousvision/graf/graf/config.py
code served (permissive licence) · get_code("856b87aabeff2342")
How it ran

no invocation recorded.

get_render_poses Not yet run
unverified (harvested, no recorded run)
autonomousvision/graf/graf/config.py
code served (permissive licence) · get_code("677c56fac6ae59a2")
How it ran

no invocation recorded.

Also present in other papers' code

Of 12 samples lifted from this paper's linked repositories, 12 carry a function identity and 0 do not. Of the 12 distinct functions with an identity, 9 are also present in other papers whose code we hold (4 with at least one paper dated earlier, 0 the same month, 9 later) and 3 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
normalize 19 other papers
1 dated earlier, 18 later
2003.08934 (2020-03)
NeRF: Representing Scenes as Neural Radiance Fields for View Synthesis
2104.04073 (2021-04)
Direct-PoseNet: Absolute Pose Regression with Photometric Consistency
2111.14292 (2021-11)
Deblur-NeRF: Neural Radiance Fields from Blurry Images
and 16 more not named here
viewmatrix 17 other papers
1 dated earlier, 16 later
2003.08934 (2020-03)
NeRF: Representing Scenes as Neural Radiance Fields for View Synthesis
2104.04073 (2021-04)
Direct-PoseNet: Absolute Pose Regression with Photometric Consistency
2111.14292 (2021-11)
Deblur-NeRF: Neural Radiance Fields from Blurry Images
and 14 more not named here
pose_spherical 11 other papers
11 later
2205.14870 (2022-05)
Compressible-composable NeRF via Rank-residual Decomposition
2205.15285 (2022-05)
Fast Dynamic Radiance Fields with Time-Aware Neural Voxels
and 8 more not named here
ptstocam 10 other papers
1 dated earlier, 9 later
2003.08934 (2020-03)
NeRF: Representing Scenes as Neural Radiance Fields for View Synthesis
2104.04073 (2021-04)
Direct-PoseNet: Absolute Pose Regression with Photometric Consistency
2111.14292 (2021-11)
Deblur-NeRF: Neural Radiance Fields from Blurry Images
and 7 more not named here
load_dv_data 2 other papers
1 dated earlier, 1 later
2003.08934 (2020-03)
NeRF: Representing Scenes as Neural Radiance Fields for View Synthesis
color_depth_map 1 other paper
1 later
2202.01020 (2022-02)
MedNeRF: Medical Neural Radiance Fields for Reconstructing 3D-aware CT-Projections from a Single X-ray
count_trainable_parameters 1 other paper
1 later
2202.01020 (2022-02)
MedNeRF: Medical Neural Radiance Fields for Reconstructing 3D-aware CT-Projections from a Single X-ray
get_render_poses 1 other paper
1 later
2202.01020 (2022-02)
MedNeRF: Medical Neural Radiance Fields for Reconstructing 3D-aware CT-Projections from a Single X-ray
update_config 1 other paper
1 later
2202.01020 (2022-02)
MedNeRF: Medical Neural Radiance Fields for Reconstructing 3D-aware CT-Projections from a Single X-ray

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

While 2D generative adversarial networks have enabled high-resolution image synthesis, they largely lack an understanding of the 3D world and the image formation process. Thus, they do not provide precise control over camera viewpoint or object pose. To address this problem, several recent approaches leverage intermediate voxel-based representations in combination with differentiable rendering. However, existing methods either produce low image resolution or fall short in disentangling camera and scene properties, e.g., the object identity may vary with the viewpoint. In this paper, we propose a generative model for radiance fields which have recently proven successful for novel view synthesis of a single scene. In contrast to voxel-based representations, radiance fields are not confined to a coarse discretization of the 3D space, yet allow for disentangling camera and scene properties while degrading gracefully in the presence of reconstruction ambiguity. By introducing a multi-scale patch-based discriminator, we demonstrate synthesis of high-resolution images while training our model from unposed 2D images alone. We systematically analyze our approach on several challenging synthetic and real-world datasets. Our experiments reveal that radiance fields are a powerful representation for generative image synthesis, leading to 3D consistent models that render with high fidelity.

For agents

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

get_harvested_code_for_paper("2007.02442")
get_code_for_paper("2007.02442")
have("2007.02442")

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

curl https://syntology.ai/api/ran/2007.02442.json

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

[![Syntology run record](https://syntology.ai/api/ran/2007.02442.svg)](https://syntology.ai/paper/2007.02442)

Connect an agent — have() is free.