SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2103.15061 · CVPR · 2021

Invertible Image Signal Processing

Yazhou Xing, Qifeng Chen, Zian Qian

arXiv · PDF · Open in the Atlas

Code that ran

4 of 6 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (3 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 3 have no contract verdict; 1 failure of our instrument; 2 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/2103.15061.json · badge: /api/ran/2103.15061.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
yzxing87/Invertible-ISP marked official in Papers with Code
Papers with Code archive
4 ran · 2 no recorded run
FunctionStatusWhere it lives
DenseBlock Ran
ran on a synthesized input · constructed an object; no result was computed
yzxing87/Invertible-ISP/model/model.py
code served (permissive licence) · get_code("bd8fe4d81407950c")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned an instance of its own class, DenseBlock: the run constructed the object and did not call it, so no result was computed. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn

torch.manual_seed(0)

ARGS = {
    'channel_in': 32,
    'channel_out': 64,
    'init': 'xavier',
    'gc': 16,
    'bias': True,
}
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 DenseBlock — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a DenseBlock 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
InvBlock Ran
ran on a synthesized input · constructed an object; no result was computed
yzxing87/Invertible-ISP/model/model.py
code served (permissive licence) · get_code("556edaf281c888db")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned an instance of its own class, InvBlock: the run constructed the object and did not call it, so no result was computed. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn

torch.manual_seed(0)

class SimpleSubnet(nn.Module):
    """Minimal subnet that takes input_channels and outputs output_channels"""
    def __init__(self, input_channels, output_channels):
        super().__init__()
        self.linear = nn.Linear(input_channels, output_channels)
    
    def forward(self, x):
        # x shape: (batch, channels, height, width)
        b, c, h, w = x.shape
        # Flatten spatial dimensions
        x_flat = x.view(b, c, -1)  # (batch, channels, h*w)
        x_flat = x_flat.transpose(1, 2)  # (batch, h*w, channels)
        out = self.linear(x_flat)  # (batch, h*w, output_channels)
        out = out.transpose(1, 2)  # (batch, output_channels, h*w)
        out = out.view(b, self.linear.out_features, h, w)  # (batch, output_channels, h, w)
        return out

def subnet_constructor(in_channels, out_channels):
    """Factory function for creating subnets"""
    return SimpleSubnet(in_channels, out_channels)

ARGS = {
    'subnet_constructor': subnet_constructor,
    'channel_num': 6,
    'channel_split_num': 2,
    'clamp': 0.8
}
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 InvBlock — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a InvBlock 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
InvertibleConv1x1 Ran
ran on a synthesized input · constructed an object; no result was computed
yzxing87/Invertible-ISP/model/model.py
code served (permissive licence) · get_code("e6b1592a851a4e6a")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned an instance of its own class, InvertibleConv1x1: the run constructed the object and did not call it, so no result was computed. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn

torch.manual_seed(0)

# Create a minimal test case for InvertibleConv1x1
num_channels = 4

# Test with LU_decomposed=False
conv_no_lu = nn.Module()

# Test with LU_decomposed=True
conv_with_lu = nn.Module()

# Create test input: (batch=2, channels=4, height=3, width=5)
test_input = torch.randn(2, 4, 3, 5)

# Create test logdet
test_logdet = torch.tensor(0.0)

ARGS = {
    "num_channels": 4,
    "LU_decomposed": False,
}
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 InvertibleConv1x1 — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a InvertibleConv1x1 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
subnet Ran
ran; our contract draft was wrong, not the code
yzxing87/Invertible-ISP/model/model.py
code served (permissive licence) · get_code("6e4e60f0269cef68")
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 function. 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
import torch.nn.init as init

torch.manual_seed(0)

class DenseBlock(nn.Module):

    def __init__(self, channel_in, channel_out, init='xavier', gc=32, bias=True):
        super(DenseBlock, self).__init__()
        self.conv1 = nn.Conv2d(channel_in, gc, 3, 1, 1, bias=bias)
        self.conv2 = nn.Conv2d(channel_in + gc, gc, 3, 1, 1, bias=bias)
        self.conv3 = nn.Conv2d(channel_in + 2 * gc, gc, 3, 1, 1, bias=bias)
        self.conv4 = nn.Conv2d(channel_in + 3 * gc, gc, 3, 1, 1, bias=bias)
        self.conv5 = nn.Conv2d(channel_in + 4 * gc, channel_out, 3, 1, 1, bias=bias)
        self.lrelu = nn.LeakyReLU(negative_slope=0.2, inplace=True)
        if init == 'xavier':
            initialize_weights_xavier([self.conv1, self.conv2, self.conv3, self.conv4], 0.1)
        else:
            initialize_weights([self.conv1, self.conv2, self.conv3, self.conv4], 0.1)
        initialize_weights(self.conv5, 0)

    def forward(self, x):
        x1 = self.lrelu(self.conv1(x))
        x2 = self.lrelu(self.conv2(torch.cat((x, x1), 1)))
        x3 = self.lrelu(self.conv3(torch.cat((x, x1, x2), 1)))
        x4 = self.lrelu(self.conv4(torch.cat((x, x1, x2, x3), 1)))
        x5 = self.conv5(torch.cat((x, x1, x2, x3, x4), 1))
        return x5

def initialize_weights(net_l, scale=1):
    if not isinstance(net_l, list):
        net_l = [net_l]
    for net in net_l:
        for m in net.modules():
            if isinstance(m, nn.Conv2d):
                init.kaiming_normal_(m.weight, a=0, mode='fan_in')
                m.weight.data *= scale
                if m.bias is not None:
                    m.bias.data.zero_()
            elif isinstance(m, nn.Linear):
                init.kaiming_normal_(m.weight, a=0, mode='fan_in')
                m.weight.data *= scale
                if m.bias is not None:
                    m.bias.data.zero_()
            elif isinstance(m, nn.BatchNorm2d):
                init.constant_(m.weight, 1)
                init.constant_(m.bias.data, 0.0)

def initialize_weights_xavier(net_l, scale=1):
    if not isinstance(net_l, list):
        net_l = [net_l]
    for net in net_l:
        for m in net.modules():
            if isinstance(m, nn.Conv2d):
                init.xavier_normal_(m.weight)
                m.weight.data *= scale
                if m.bias is not None:
                    m.bias.data.zero_()
            elif isinstance(m, nn.Linear):
                init.xavier_normal_(m.weight)
                m.weight.data *= scale
                if m.bias is not None:
                    m.bias.data.zero_()
            elif isinstance(m, nn.BatchNorm2d):
                init.constant_(m.weight, 1)
                init.constant_(m.bias.data, 0.0)

def subnet(net_structure, init='xavier'):

    def constructor(channel_in, channel_out):
        if net_structure == 'DBNet':
            if init == 'xavier':
                return DenseBlock(channel_in, channel_out, init)
            else:
                return DenseBlock(channel_in, channel_out)
        else:
            return None
    return constructor


ARGS = {
    'net_structure': 'DBNet',
    'init': 'xavier'
}
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 function — a function output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a function 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
InvISPNet Not yet run
unverified (harvested, no recorded run)
yzxing87/Invertible-ISP/model/model.py
code served (permissive licence) · get_code("fda23ef78edc9d9f")
How it ran

no invocation recorded.

initialize_weights_xavier Not yet run
unverified (harvested, no recorded run)
yzxing87/Invertible-ISP/model/model.py
code served (permissive licence) · get_code("6f1bb896f0c61c30")
How it ran

no invocation recorded.

Also present in other papers' code

Of 6 samples lifted from this paper's linked repositories, 6 carry a function identity and 0 do not. Of the 6 distinct functions with an identity, 1 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 1 later) and 5 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
initialize_weights_xavier
brick of this paper's assembly: helper
1 other paper
1 later
2304.12300 (2023-04)
Large-capacity and Flexible Video Steganography via Invertible Neural Network

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

Unprocessed RAW data is a highly valuable image format for image editing and computer vision. However, since the file size of RAW data is huge, most users can only get access to processed and compressed sRGB images. To bridge this gap, we design an Invertible Image Signal Processing (InvISP) pipeline, which not only enables rendering visually appealing sRGB images but also allows recovering nearly perfect RAW data. Due to our framework's inherent reversibility, we can reconstruct realistic RAW data instead of synthesizing RAW data from sRGB images without any memory overhead. We also integrate a differentiable JPEG compression simulator that empowers our framework to reconstruct RAW data from JPEG images. Extensive quantitative and qualitative experiments on two DSLR demonstrate that our method obtains much higher quality in both rendered sRGB images and reconstructed RAW data than alternative methods.

For agents

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

get_harvested_code_for_paper("2103.15061")
get_code_for_paper("2103.15061")
have("2103.15061")

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

curl https://syntology.ai/api/ran/2103.15061.json

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

[![Syntology run record](https://syntology.ai/api/ran/2103.15061.svg)](https://syntology.ai/paper/2103.15061)

Connect an agent — have() is free.