SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2307.16508 · ICCV · 2023

Towards General Low-Light Raw Noise Synthesis and Modeling

Nong Sang, Xinran Liu, Bin Xu, Changxin Gao, Feng Zhang, Zhiqiang Li, Qingbo Lu

arXiv · PDF · Open in the Atlas

Code that ran

15 of 20 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; 4 honoured the contract we drafted, 0 violated it and 0 have no contract verdict; 11 failures of our instrument; 5 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/2307.16508.json · badge: /api/ran/2307.16508.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
fengzhang427/LRD canonical (classifier label)
classifier basis: census
15 ran · 5 no recorded run
FunctionStatusWhere it lives
apply_ccms Ran
ran; our contract draft was wrong, not the code
fengzhang427/LRD/raw_process/LRD_post_process.py
code served (permissive licence) · get_code("70e90e19a49b549d")
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, 7, 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)

# images: shape (batch, channels, height, width)
# After permute(0, 2, 3, 1): (batch, height, width, channels)
# After [:, :, :, None, :]: (batch, height, width, 1, channels)
images = torch.randn(2, 3, 4, 5)  # batch=2, channels=3, height=4, width=5

# ccms: shape (batch, output_channels, input_channels)
# After [:, None, None, :, :]: (batch, 1, 1, output_channels, input_channels)
# Must match input channels (3) and have some output channels
ccms = torch.randn(2, 7, 3)  # batch=2, output_channels=7, input_channels=3

ARGS = {
    "images": images,
    "ccms": ccms,
}
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, 7, 4, 5) — shape (2, 7, 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
apply_gains Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
fengzhang427/LRD/raw_process/SID_post_process.py
code served (permissive licence) · get_code("63cc917598a67f63")
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, 4, 5, 6) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import numpy as np

torch.manual_seed(0)

# Analyze the function:
# - bayer_images: used with .shape to get N, C, and other dims; used in multiplication
# - wbs: used with .view(1, C, 1, 1) and .repeat_interleave(N, dim=0)
# So bayer_images is a 4D tensor (N, C, H, W)
# and wbs is a 1D tensor of shape (C,) containing per-channel gains

# Create bayer_images: shape (N=2, C=4, H=5, W=6)
# Bayer images typically have 4 channels (RGGB pattern)
bayer_images = torch.randn(2, 4, 5, 6)

# Create wbs: shape (C=4,) - white balance gains for each channel
wbs = torch.randn(4)

ARGS = {
    "bayer_images": bayer_images,
    "wbs": wbs,
}
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 4, dtype family float, shape (2, 4, 5, 6) — shape (2, 4, 5, 6) 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
demosaic Ran
ran; our contract draft was wrong, not the code
fengzhang427/LRD/raw_process/LRD_post_process.py
code served (permissive licence) · get_code("1da403b404f61152")
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, 3, 12, 8) 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 a small RGGB Bayer image batch
# Shape: (batch, height, width, channels) where channels=4 for RGGB
# We need height and width to be even numbers since the function uses factor-2 operations
# Using batch=2, height=4, width=6 for small but valid dimensions
bayer_images = torch.randn(2, 4, 6, 4)

ARGS = {
    'bayer_images': bayer_images
}
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, 3, 12, 8) — shape (2, 3, 12, 8) 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
drop_path Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
fengzhang427/LRD/ViT_helper.py
pointer only (licence: MIT); the stored copy get_code serves, from fudan-zvg/SeaFormer, is pointer only (licence: NONE) · get_code("3ac6b7d76e8e3584")
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 fudan-zvg/SeaFormer, 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.

forward_patches Ran
ran; our contract draft was wrong, not the code
fengzhang427/LRD/train_denoise_LRD.py
code served (permissive licence) · get_code("135aad478fe73cf5")
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, 3, 128, 96) 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)

# Create a minimal model that takes an input tensor and returns an output tensor
class SimpleDenoiser(nn.Module):
    def __init__(self):
        super().__init__()
        self.conv = nn.Conv2d(3, 3, kernel_size=3, padding=1)
    
    def forward(self, x):
        return self.conv(x)

# Create the model instance
model = SimpleDenoiser()
model.eval()

# Create a small input image tensor (batch=2, channels=3, height=128, width=96)
# Using distinct dimensions to allow probing later
input_img = torch.rand(2, 3, 128, 96)

ARGS = {
    "model": model,
    "input_img": input_img,
    "patch_size": 256,
    "pad": 32
}
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 4, dtype family float, shape (2, 3, 128, 96) — shape (2, 3, 128, 96) 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
gamma_compression Ran
ran; our contract draft was wrong, not the code
fengzhang427/LRD/raw_process/SID_post_process.py
code served (permissive licence) · get_code("f8c8b670d9bb8c81")
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, 3, 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 torch

torch.manual_seed(0)

# Create a small tensor of linear space images
# Shape: (batch=2, channels=3, height=4, width=4) - typical image tensor
images = torch.rand(2, 3, 4, 4)

# gamma parameter for the compression
gamma = 2.2

ARGS = {
    "images": images,
    "gamma": gamma,
}
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, 3, 4, 4) — shape (2, 3, 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
get_psnr Ran
ran, honoured the contract we drafted
fengzhang427/LRD/utils.py
code served (permissive licence) · get_code("6c071f2514ce8a40")
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 float64, ndim 0, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim 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 skimage.metrics

torch.manual_seed(0)

# Create two small image arrays with values in range [0, 255]
# PSNR requires two images of the same shape
im1 = np.random.randint(0, 256, size=(4, 3), dtype=np.uint8)
im2 = np.random.randint(0, 256, size=(4, 3), dtype=np.uint8)

ARGS = {
    "im1": im1,
    "im2": im2,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, skimage, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float64, ndim 0, dtype family float — no output shape is stated: observe_signature.py recorded float ndim 0, which disagrees with the fixture run
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_scheduler Ran
ran; our contract draft was wrong, not the code
fengzhang427/LRD/utils.py
code served (permissive licence) · get_code("5fa7698e84d3a88c")
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 LambdaLR. 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 torch.optim import Adam

torch.manual_seed(0)

# Create a minimal model
model = nn.Linear(4, 3)

# Create an optimizer
optimizer = Adam(model.parameters(), lr=0.001)

# Create hyperparameters dict for 'linear' policy
hyperparameters = {
    'lr_policy': 'linear',
    'epoch_count': 0,
    'n_epochs': 100,
    'n_epochs_decay': 100,
}

# Set iterations to -1 as default
iterations = -1

ARGS = {
    'optimizer': optimizer,
    'hyperparameters': hyperparameters,
    'iterations': iterations,
}
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 LambdaLR — a LambdaLR output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a LambdaLR 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
pack_raw_DJI Ran
ran, honoured the contract we drafted
fengzhang427/LRD/noise_profile/noise_profile_SonyA7S2.py
code served (permissive licence) · get_code("ed42d1958af188dc")
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 (4, 2, 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

class MockRawImage:
    def __init__(self):
        # Create a raw image with shape (4, 4) - small test image
        self.raw_image_visible = np.random.randint(0, 65535, size=(4, 4), dtype=np.uint16)
        
        # raw_pattern defines the Bayer pattern layout
        # For a 2x2 pattern that repeats: [0=R, 1=G1, 2=B, 3=G2]
        self.raw_pattern = np.array([[0, 1], [3, 2]], dtype=np.uint8)
        
        # black_level_per_channel for RGGB channels (4 channels)
        self.black_level_per_channel = np.array([100, 100, 100, 100], dtype=np.uint16)

raw = MockRawImage()

ARGS = {
    "raw": raw
}
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 (4, 2, 2) — shape (4, 2, 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
pack_raw_SID Ran
ran, honoured the contract we drafted
fengzhang427/LRD/datasets/Denoise_dataset.py
code served (permissive licence) · get_code("37d9c6eba769af47")
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 (4, 2, 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

class MockRaw:
    def __init__(self):
        # Create a small 4x4 raw image with Bayer pattern
        self.raw_image_visible = np.array([
            [100, 150, 120, 160],
            [110, 155, 125, 165],
            [105, 152, 122, 162],
            [115, 157, 130, 167]
        ], dtype=np.uint16)
        
        # Standard Bayer pattern: R G B G
        # 0=R, 1=G1, 2=B, 3=G2
        self.raw_pattern = np.array([
            [0, 1],
            [3, 2]
        ], dtype=np.uint8)
        
        self.white_level = 1023
        self.black_level_per_channel = np.array([64, 64, 64, 64], dtype=np.uint16)

raw = MockRaw()

ARGS = {
    "raw": raw
}
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 (4, 2, 2) — shape (4, 2, 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
pack_raw_bayer Ran
ran, honoured the contract we drafted
fengzhang427/LRD/noise_profile/noise_profile_SonyA7S2.py
code served (permissive licence) · get_code("77dad890b0d037bc")
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 (4, 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

class MockRawImage:
    def __init__(self):
        # Create a small raw image (4x6 to have even dimensions for Bayer pattern)
        self.raw_image_visible = np.random.randint(100, 5000, size=(4, 6), dtype=np.uint16)
        
        # Standard Bayer pattern: [[0, 1], [3, 2]] (R, G1, G2, B)
        # Repeated to match image dimensions
        self.raw_pattern = np.array([[0, 1, 0, 1, 0, 1],
                                      [3, 2, 3, 2, 3, 2],
                                      [0, 1, 0, 1, 0, 1],
                                      [3, 2, 3, 2, 3, 2]], dtype=np.uint8)
        
        # Black level per channel (R, G1, B, G2)
        self.black_level_per_channel = np.array([100, 110, 120, 105], dtype=np.float32)

ARGS = {
    "raw": MockRawImage()
}
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 (4, 2, 3) — shape (4, 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
raw2stack_RGGB Ran
ran; our contract draft was wrong, not the code
fengzhang427/LRD/train_denoise_LRD.py
code served (permissive licence) · get_code("12e66c962c260d27")
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 (4, 2, 3) 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

torch.manual_seed(0)

# The function raw2stack_RGGB expects:
# - var: a 2D tensor with even dimensions (h, w)
# The function:
# - Checks var.shape to get h, w
# - Checks var.is_cuda to determine device
# - Creates a 4D output tensor of shape (4, h//2, w//2)
# - Indexes into var using slicing [0::2, 1::2] etc.

# Create a small 2D tensor with even dimensions
var = torch.randn(4, 6)

ARGS = {
    "var": var
}
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 (4, 2, 3) — shape (4, 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
read_dorf Ran
ran; our contract draft was wrong, not the code
fengzhang427/LRD/raw_process/EMoR/EMoR.py
code served (permissive licence) · get_code("031ab536c74127bb")
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 tempfile
import os
import numpy as np

# Create a temporary file with valid DORF format
temp_dir = tempfile.mkdtemp()
temp_file = os.path.join(temp_dir, "test.dorf")

# Write DORF format data: curve_names on line 0, Es on line 3, Bs on line 5, repeated every 6 lines
with open(temp_file, 'w') as f:
    # First curve (lines 0-5)
    f.write("curve_1\n")
    f.write("dummy_line_1\n")
    f.write("dummy_line_2\n")
    f.write("1.0 2.0 3.0 4.0\n")
    f.write("dummy_line_4\n")
    f.write("5.0 6.0 7.0 8.0\n")
    
    # Second curve (lines 6-11)
    f.write("curve_2\n")
    f.write("dummy_line_1\n")
    f.write("dummy_line_2\n")
    f.write("9.0 10.0 11.0 12.0\n")
    f.write("dummy_line_4\n")
    f.write("13.0 14.0 15.0 16.0\n")

ARGS = {
    "address": temp_file
}
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
stack2raw_RGGB Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
fengzhang427/LRD/train_denoise_LRD.py
code served (permissive licence) · get_code("7b94bb50a39ba2fd")
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 (6, 10) 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

torch.manual_seed(0)

# The function expects a tensor with shape (4, h, w)
# where the 4 channels represent R, G, G, B components
# It will create an output of shape (2*h, 2*w)
var = torch.randn(4, 3, 5)

ARGS = {
    "var": var
}
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 (6, 10) — shape (6, 10) 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
trunc_normal_ Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
fengzhang427/LRD/ViT_helper.py
code served (permissive licence) · get_code("02566da69866c48c")
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) 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
import math

torch.manual_seed(0)

# Create a small tensor to be filled with truncated normal values
tensor = torch.empty(2, 4)

# Define arguments for trunc_normal_
ARGS = {
    'tensor': tensor,
    'mean': 0.0,
    'std': 1.0,
    'a': -2.0,
    'b': 2.0
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
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
apply_gains Not yet run
unverified (harvested, no recorded run)
fengzhang427/LRD/raw_process/LRD_post_process.py
code served (permissive licence) · get_code("e6501a9602eb526b")
How it ran

no invocation recorded.

forward_patches Not yet run
unverified (harvested, no recorded run)
fengzhang427/LRD/train_denoise_SID.py
code served (permissive licence) · get_code("785f3a8049b134d2")
How it ran

no invocation recorded.

get_ssim Not yet run
unverified (harvested, no recorded run)
fengzhang427/LRD/utils.py
code served (permissive licence) · get_code("08a9da5a3c00e119")
How it ran

no invocation recorded.

log Not yet run
unverified (harvested, no recorded run)
fengzhang427/LRD/noise_profile/noise_profile_SonyA7S2.py
code served (permissive licence) · get_code("fe0468e2d3603fb0")
How it ran

no invocation recorded.

read_emor Not yet run
unverified (harvested, no recorded run)
fengzhang427/LRD/raw_process/EMoR/EMoR.py
code served (permissive licence) · get_code("dfa0c71130d11161")
How it ran

no invocation recorded.

Also present in other papers' code

Of 20 samples lifted from this paper's linked repositories, 20 carry a function identity and 0 do not. Of the 20 distinct functions with an identity, 5 are also present in other papers whose code we hold (4 with at least one paper dated earlier, 2 the same month, 3 later) and 15 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
drop_path 224 other papers
98 dated earlier, 2 the same month, 117 later, 6 same year, order not determinable, 1 undated · 38 only as an assembly brick
2003.12039 (2020-03)
RAFT: Recurrent All-Pairs Field Transforms for Optical Flow
2007.09384 (2020-07)
Multi-Scale Positive Sample Refinement for Few-Shot Object Detection
2010.11929 (2020-10)
An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale
and 221 more not named here
trunc_normal_ 160 other papers
62 dated earlier, 3 the same month, 93 later, 2 same year, order not determinable · 36 only as an assembly brick
1911.10194 (2019-11)
Panoptic-DeepLab: A Simple, Strong, and Fast Baseline for Bottom-Up Panoptic Segmentation
2005.10242 (2020-05)
Understanding Contrastive Representation Learning through Alignment and Uniformity on the Hypersphere
2007.12911 (2020-07)
Tighter risk certificates for neural networks
and 157 more not named here
log 1 other paper
1 later
Li_Learning_Hierarchical_Line_Buffer_for_Image_Processing_ICCV_2025_paper (2025 (year only))
Learning Hierarchical Line Buffer for Image Processing
read_dorf 1 other paper
1 dated earlier
2108.02158 (2021-08)
Physics-based Noise Modeling for Extreme Low-light Photography
read_emor 1 other paper
1 dated earlier
2108.02158 (2021-08)
Physics-based Noise Modeling for Extreme Low-light Photography

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

Modeling and synthesizing low-light raw noise is a fundamental problem for computational photography and image processing applications. Although most recent works have adopted physics-based models to synthesize noise, the signal-independent noise in low-light conditions is far more complicated and varies dramatically across camera sensors, which is beyond the description of these models. To address this issue, we introduce a new perspective to synthesize the signal-independent noise by a generative model. Specifically, we synthesize the signal-dependent and signalindependent noise in a physics-and learning-based manner, respectively. In this way, our method can be considered as a general model, that is, it can simultaneously learn different noise characteristics for different ISO levels and generalize to various sensors. Subsequently, we present an effective multi-scale discriminator termed Fourier transformer discriminator (FTD) to distinguish the noise distribution accurately. Additionally, we collect a new low-light raw denoising (LRD) dataset for training and benchmarking. Qualitative validation shows that the noise generated by our proposed noise model can be highly similar to the real noise in terms of distribution. Furthermore, extensive denoising experiments demonstrate that our method performs favorably against state-of-the-art methods on different sensors.

For agents

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

get_harvested_code_for_paper("2307.16508")
get_code_for_paper("2307.16508")
have("2307.16508")

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

curl https://syntology.ai/api/ran/2307.16508.json

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

[![Syntology run record](https://syntology.ai/api/ran/2307.16508.svg)](https://syntology.ai/paper/2307.16508)

Connect an agent — have() is free.