SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2110.07604 · NeurIPS · 2021

NeRS: Neural Reflectance Surfaces for Sparse-view 3D Reconstruction in the Wild

Deva Ramanan, Shubham Tulsiani, Jason Zhang, Gengshan Yang

arXiv · PDF · Open in the Atlas

Code that ran

14 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; 0 honoured the contract we drafted, 0 violated it and 14 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/2110.07604.json · badge: /api/ran/2110.07604.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
jasonyzhang/ners canonical (classifier label)
classifier basis not recorded on the link
14 ran · 0 no recorded run
FunctionStatusWhere it lives
antialias Ran
ran on a synthesized input
jasonyzhang/ners/ners/utils/image.py
code served (permissive licence) · get_code("68f756675a769b50")
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 Image. 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 numpy as np
from PIL import Image

# Create a test PIL Image
pil_image = Image.new('RGB', (8, 6), color='red')

# Create a test numpy array (normalized to 0-1 range)
numpy_image = np.random.rand(6, 8, 3).astype(np.float32)

ARGS = {
    'image': pil_image
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
PIL, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Image — a Image output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Image output the recorded ndim 3 and dtype family int 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_crop_parameters Ran
ran on a synthesized input
jasonyzhang/ners/ners/pytorch3d/cameras.py
code served (permissive licence) · get_code("99e4cc514f8265d7")
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 numpy as np

# Construct valid arguments for compute_crop_parameters
image_size = 512  # Size of image (int)
bbox = np.array([100.0, 120.0, 300.0, 320.0])  # Square bounding box in xyxy format
image_center = np.array([256.0, 256.0])  # Principal point/image center

ARGS = {
    "image_size": image_size,
    "bbox": bbox,
    "image_center": image_center,
}
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 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
compute_crop_parameters Ran
ran on a synthesized input
jasonyzhang/ners/ners/utils/camera.py
code served (permissive licence) · get_code("4e1ef6a651cf45c9")
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 numpy as np

# Construct valid arguments for compute_crop_parameters

# image_size: can be int or tuple (W, H)
image_size = (640, 480)

# bbox: square bounding box in xyxy format (x1, y1, x2, y2)
bbox = [100, 150, 250, 300]

# image_center: center of projection/principal point (2,)
image_center = np.array([320.0, 240.0])

ARGS = {
    "image_size": image_size,
    "bbox": bbox,
    "image_center": image_center,
}
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 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
compute_distance_transform Ran
ran on a synthesized input
jasonyzhang/ners/ners/utils/masks.py
code served (permissive licence) · get_code("e95c6ae18f432126")
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 (4, 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
from scipy.ndimage import distance_transform_edt

# Create a small binary mask for testing
# The function expects a 2D or 3D array where 1 represents foreground and 0 represents background
mask = np.array([
    [0, 1, 1, 0],
    [1, 1, 1, 1],
    [1, 1, 0, 1],
    [0, 1, 1, 0]
], dtype=np.uint8)

ARGS = {
    "mask": mask
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, scipy — 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 (4, 4) — shape (4, 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
compute_lighting_specular Ran
ran on a synthesized input
jasonyzhang/ners/ners/pytorch3d/textures.py
code served (permissive licence) · get_code("b5fee1e18214bd34")
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, 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 torch.nn as nn
import numpy as np

torch.manual_seed(0)

class SimpleEnvMap(nn.Module):
    """Minimal environment map that takes ray directions and returns light intensities."""
    def __init__(self):
        super().__init__()
        self.linear = nn.Linear(3, 1)
    
    def forward(self, rays):
        # rays: (N, 3) -> output (N, 1)
        return self.linear(rays)

# Create environment map
env_map = SimpleEnvMap()

# N1 = 2 vertices, each with 3D normal
vertex_normals = torch.randn(2, 3)
vertex_normals = vertex_normals / torch.norm(vertex_normals, dim=-1, keepdim=True)

# View direction from vertices to camera
view_direction = torch.randn(2, 3)
view_direction = view_direction / torch.norm(view_direction, dim=-1, keepdim=True)

# N2 = 4 discrete ray samples
env_map_rays = torch.randn(4, 3)
env_map_rays = env_map_rays / torch.norm(env_map_rays, dim=-1, keepdim=True)

# Specularity and shininess (can be scalar or per-vertex)
specularity = torch.ones(2, 1)
shininess = torch.ones(2, 1) * 32.0

ARGS = {
    "env_map": env_map,
    "vertex_normals": vertex_normals,
    "view_direction": view_direction,
    "env_map_rays": env_map_rays,
    "specularity": specularity,
    "shininess": shininess,
    "device": None,
    "max_batch_size": None,
}
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 Tensor, 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
crop_image Ran
ran on a synthesized input
jasonyzhang/ners/ners/utils/image.py
code served (permissive licence) · get_code("c7b71b76c6df1fa6")
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 Image. 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 numpy as np
from PIL import Image

# Create a small test image in RGB mode
# The function expects a PIL Image with a mode attribute
test_image = Image.new('RGB', (100, 100), color=(200, 100, 50))

# Create a bounding box as tuple (x1, y1, x2, y2)
# This represents a crop region
bbox = (10, 20, 60, 80)

ARGS = {
    'image': test_image,
    'bbox': bbox
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
PIL, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Image — a Image output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Image output the recorded ndim 3 and dtype family int 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
dot_product Ran
ran on a synthesized input
jasonyzhang/ners/ners/pytorch3d/textures.py
code served (permissive licence) · get_code("0f6bb8ad1d96cbf6")
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, 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 torch.nn as nn

torch.manual_seed(0)

# Create two small tensors with shape (batch=2, dim=4)
a = torch.randn(2, 4)
b = torch.randn(2, 4)

ARGS = {
    "a": a,
    "b": b
}
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, 1) — shape (2, 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
get_bbox Ran
ran on a synthesized input
jasonyzhang/ners/ners/data.py
code served (permissive licence) · get_code("4d4e8165125c15e0")
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 int, shape (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

# Create a small test image with some non-zero values
# The function uses np.where to find non-zero elements and compute bounding box
img = np.zeros((5, 6), dtype=np.uint8)
# Add some non-zero pixels to create a bounding box
img[1:3, 2:5] = 255

ARGS = {
    "img": img
}
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 int, shape (4,) — shape (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_car_model Ran
ran on a synthesized input
jasonyzhang/ners/ners/models.py
code served (permissive licence) · get_code("70ecc56fb758a32d")
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 TemplateUV. 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 unittest.mock import MagicMock, patch

torch.manual_seed(0)

# Create a mock state_dict that would be loaded
mock_state_dict = {
    'positional_encoding.frequencies': torch.tensor([0.1, 0.2, 0.4, 0.8, 1.6, 3.2, 6.4, 12.8, 25.6, 51.2]),
    'mlp.0.weight': torch.randn(256, 60),
    'mlp.0.bias': torch.randn(256),
    'mlp.1.weight': torch.ones(256),
    'mlp.1.bias': torch.zeros(256),
    'mlp.3.weight': torch.randn(256, 256),
    'mlp.3.bias': torch.randn(256),
    'mlp.4.weight': torch.ones(256),
    'mlp.4.bias': torch.zeros(256),
    'mlp.6.weight': torch.randn(3, 256),
    'mlp.6.bias': torch.randn(3),
}

# Create a mock for torch.load that returns our mock state_dict
def mock_load(path):
    return mock_state_dict

ARGS = {
    'path': 'models/templates/car.pth'
}

# Patch torch.load to return our mock state_dict
torch.load = mock_load
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 TemplateUV — a TemplateUV output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a TemplateUV 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
matrix_to_rot6d Ran
ran on a synthesized input
jasonyzhang/ners/ners/utils/geometry.py
code served (permissive licence) · get_code("4d68a35bc9e25fbe")
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 (2, 3, 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 torch
import torch.nn as nn

torch.manual_seed(0)

# The function takes a rotation matrix of shape (B x 3 x 3)
# and returns the first two columns reshaped to (B x 3 x 2)
# We need to create a batch of valid rotation matrices

# Create a small batch of rotation matrices
# Using batch size 2, with 3x3 rotation matrices
B = 2
rotmat = torch.randn(B, 3, 3)

# To make them valid rotation matrices, we can use QR decomposition
# or just use them as-is since the function doesn't validate
# For simplicity, let's just use random tensors (function doesn't check validity)
rotmat = torch.randn(2, 3, 3)

ARGS = {
    "rotmat": rotmat
}
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 3, dtype family float, shape (2, 3, 2) — shape (2, 3, 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
normalize Ran
ran on a synthesized input
jasonyzhang/ners/ners/pytorch3d/textures.py
code served (permissive licence) · get_code("9680239af0e7d99b")
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)

# x is a tensor that needs to have .norm() method and support division
# The function calls x.norm(dim=-1, keepdim=True) which works on tensors
# Create a small tensor with shape (2, 4) - batch size 2, dimension 4
x = torch.randn(2, 4)

ARGS = {
    "x": x,
    "eps": 1e-06
}
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
rle_to_binary_mask Ran
ran on a synthesized input
jasonyzhang/ners/ners/utils/masks.py
code served (permissive licence) · get_code("dad37729d8943913")
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

# Create a valid COCO RLE format dict with counts as a list
rle_single = {
    'counts': [0, 5, 3, 2, 1, 4],  # alternating: run of 0s, run of 1s, run of 0s, etc.
    'size': [3, 4]  # height=3, width=4, total=12 pixels
}

# Test with a list of RLE dicts
rle_list = [
    {
        'counts': [1, 2, 3, 1],
        'size': [2, 3]
    },
    {
        'counts': [0, 4, 2],
        'size': [2, 3]
    }
]

# Also test with string format counts
rle_string = {
    'counts': '2 3 1 2',
    'size': [2, 3]
}

ARGS = {
    'rle': rle_single
}
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
rot6d_to_matrix Ran
ran on a synthesized input
jasonyzhang/ners/ners/utils/geometry.py
code served (permissive licence) · get_code("e9c596da27a16cb8")
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 (2, 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 torch
import torch.nn.functional as F

torch.manual_seed(0)

# Create a small batch of 6D rotation representations
# Shape: (B x 6) where B=2 (batch size)
rot_6d = torch.randn(2, 6)

ARGS = {
    'rot_6d': rot_6d
}
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 3, dtype family float, shape (2, 3, 3) — shape (2, 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
visualize_masks Ran
ran on a synthesized input
jasonyzhang/ners/ners/utils/masks.py
code served (permissive licence) · get_code("a2821f4583ab8c8e")
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 float, shape (256, 256, 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 two boolean masks of shape (256, 256)
# mask: ground truth mask
mask = np.random.rand(256, 256) > 0.5

# mask_pred: predicted mask
mask_pred = np.random.rand(256, 256) > 0.5

ARGS = {
    "mask": mask,
    "mask_pred": mask_pred,
}
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 float, shape (256, 256, 3) — shape (256, 256, 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

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, 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 14 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

Recent history has seen a tremendous growth of work exploring implicit representations of geometry and radiance, popularized through Neural Radiance Fields (NeRF). Such works are fundamentally based on a (implicit) volumetric representation of occupancy, allowing them to model diverse scene structure including translucent objects and atmospheric obscurants. But because the vast majority of real-world scenes are composed of well-defined surfaces, we introduce a surface analog of such implicit models called Neural Reflectance Surfaces (NeRS). NeRS learns a neural shape representation of a closed surface that is diffeomorphic to a sphere, guaranteeing water-tight reconstructions. Even more importantly, surface parameterizations allow NeRS to learn (neural) bidirectional surface reflectance functions (BRDFs) that factorize view-dependent appearance into environmental illumination, diffuse color (albedo), and specular "shininess." Finally, rather than illustrating our results on synthetic scenes or controlled in-the-lab capture, we assemble a novel dataset of multi-view images from online marketplaces for selling goods. Such "in-the-wild" multi-view image sets pose a number of challenges, including a small number of views with unknown/rough camera estimates. We demonstrate that surface-based neural reconstructions enable learning from such data, outperforming volumetric neural rendering-based reconstructions. We hope that NeRS serves as a first step toward building scalable, high-quality libraries of real-world shape, materials, and illumination. The project page with code and video visualizations can be found at jasonyzhang.com/ners.

For agents

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

get_harvested_code_for_paper("2110.07604")
get_code_for_paper("2110.07604")
have("2110.07604")

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

curl https://syntology.ai/api/ran/2110.07604.json

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

[![Syntology run record](https://syntology.ai/api/ran/2110.07604.svg)](https://syntology.ai/paper/2110.07604)

Connect an agent — have() is free.