SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2410.04847 · NeurIPS · 2024

Causal Context Adjustment Loss for Learned Image Compression

Shengxi Li, Shuhang Gu, Minghao Han, Shiyin Jiang, Xin Deng, Mai Xu, Ce Zhu, Hyper Encoder

arXiv · PDF · Open in the Atlas

Code that ran

5 of 5 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, 0 violated it and 0 have no contract verdict; 4 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/2410.04847.json · badge: /api/ran/2410.04847.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
LabShuHangGU/CCA canonical (classifier label)
classifier basis: census
5 ran · 0 no recorded run
FunctionStatusWhere it lives
conv Ran
ran; our contract draft was wrong, not the code
LabShuHangGU/CCA/models/aux_em.py
pointer only (licence: MIT); the stored copy get_code serves, from micmic123/qmapcompression, is pointer only (licence: NONE) · get_code("f2cea3e9848ae52d")
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 micmic123/qmapcompression, 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.

deconv Ran
ran; our contract draft was wrong, not the code
LabShuHangGU/CCA/models/aux_em.py
pointer only (licence: MIT); the stored copy get_code serves, from micmic123/qmapcompression, is pointer only (licence: NONE) · get_code("1afd37c576493fea")
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 micmic123/qmapcompression, 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.

find_named_module Ran
ran; our contract draft was wrong, not the code
LabShuHangGU/CCA/models/vae.py
code served (permissive licence) · get_code("0b4d7c9d1baaf828")
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 Linear. 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 simple nested module structure
class SimpleModule(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear1 = nn.Linear(4, 3)
        self.linear2 = nn.Linear(3, 2)

class RootModule(nn.Module):
    def __init__(self):
        super().__init__()
        self.layer1 = SimpleModule()
        self.layer2 = nn.Linear(2, 1)

# Instantiate the root module
root_module = RootModule()

# The function uses module.named_modules() which returns tuples of (name, module)
# We need to provide a valid module and a query string that matches one of the named modules
# Valid queries for root_module would be: '', 'layer1', 'layer1.linear1', 'layer1.linear2', 'layer2'

ARGS = {
    "module": root_module,
    "query": "layer1.linear1"
}
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 Linear — a Linear output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Linear 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
get_scale_table Ran
ran, honoured the contract we drafted
LabShuHangGU/CCA/models/vae.py
pointer only (licence: MIT); the stored copy get_code serves, from xinjie-q/distributed-dvc, is pointer only (licence: NONE) · get_code("1eca81f63e28f103")
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 xinjie-q/distributed-dvc, 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.

ste_round Ran
ran; our contract draft was wrong, not the code
LabShuHangGU/CCA/models/aux_em.py
code served (permissive licence) · get_code("f5a3092a06493cb7")
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
from torch import Tensor

torch.manual_seed(0)

# ste_round takes a single parameter x which is a Tensor
# The function uses: torch.round(), .detach(), and basic arithmetic operations
# All of these work on any tensor, so we just need a simple tensor

x = torch.randn(2, 4)

ARGS = {
    "x": x
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 2, dtype family float, shape (2, 4) — shape (2, 4) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block

Also present in other papers' code

Of 5 samples lifted from this paper's linked repositories, 5 carry a function identity and 0 do not. Of the 5 distinct functions with an identity, 5 are also present in other papers whose code we hold (5 with at least one paper dated earlier, 0 the same month, 5 later) and 0 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
get_scale_table 16 other papers
14 dated earlier, 2 later
1802.01436 (2018-02)
Variational image compression with a scale hyperprior
2011.03029 (2020-11)
CompressAI: a PyTorch library and evaluation platform for end-to-end compression research
2011.06294 (2020-11)
Real-Time Intermediate Flow Estimation for Video Frame Interpolation
and 13 more not named here
conv 14 other papers
11 dated earlier, 3 later · 1 only as an assembly brick
Deng_Deep_Homography_for_Efficient_Stereo_Image_Compression_CVPR_2021_paper (2021 (year only))
Deep Homography for Efficient Stereo Image Compression
2108.03690 (2021-08)
Enhanced Invertible Encoding for Learned Image Compression
2108.09551 (2021-08)
Variable-Rate Deep Image Compression through Spatially-Adaptive Feature Transform
and 11 more not named here
find_named_module 8 other papers
7 dated earlier, 1 later
Deng_Deep_Homography_for_Efficient_Stereo_Image_Compression_CVPR_2021_paper (2021 (year only))
Deep Homography for Efficient Stereo Image Compression
2108.03690 (2021-08)
Enhanced Invertible Encoding for Learned Image Compression
aaai_19890 (2022 (year only))
OoDHDR-Codec: Out-of-Distribution Generalization for HDR Image Compression
and 5 more not named here
deconv 7 other papers
4 dated earlier, 3 later · 1 only as an assembly brick
2108.09551 (2021-08)
Variable-Rate Deep Image Compression through Spatially-Adaptive Feature Transform
2203.10886 (2022-03)
ELIC: Efficient Learned Image Compression with Unevenly Grouped Space-Channel Contextual Adaptive Coding
2306.05085 (2023-06)
TransTIC: Transferring Transformer-based Image Compression from Human Perception to Machine Perception
and 4 more not named here
ste_round 7 other papers
6 dated earlier, 1 later
aaai_19890 (2022 (year only))
OoDHDR-Codec: Out-of-Distribution Generalization for HDR Image Compression
2203.08450 (2022-03)
The Devil Is in the Details: Window-based Attention for Image Compression
2303.14978 (2023-03)
Learned Image Compression with Mixed Transformer-CNN Architectures
and 4 more not named here

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

In recent years, learned image compression (LIC) technologies have surpassed conventional methods notably in terms of rate-distortion (RD) performance. Most present learned techniques are VAE-based with an autoregressive entropy model, which obviously promotes the RD performance by utilizing the decoded causal context. However, extant methods are highly dependent on the fixed hand-crafted causal context. The question of how to guide the auto-encoder to generate a more effective causal context benefit for the autoregressive entropy models is worth exploring. In this paper, we make the first attempt in investigating the way to explicitly adjust the causal context with our proposed Causal Context Adjustment loss (CCA-loss). By imposing the CCA-loss, we enable the neural network to spontaneously adjust important information into the early stage of the autoregressive entropy model. Furthermore, as transformer technology develops remarkably, variants of which have been adopted by many state-of-the-art (SOTA) LIC techniques. The existing computing devices have not adapted the calculation of the attention mechanism well, which leads to a burden on computation quantity and inference latency. To overcome it, we establish a convolutional neural network (CNN) image compression model and adopt the unevenly channel-wise grouped strategy for high efficiency. Ultimately, the proposed CNN-based LIC network trained with our Causal Context Adjustment loss attains a great trade-off between inference latency and rate-distortion performance. The code is available here.

For agents

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

get_harvested_code_for_paper("2410.04847")
get_code_for_paper("2410.04847")
have("2410.04847")

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

curl https://syntology.ai/api/ran/2410.04847.json

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

[![Syntology run record](https://syntology.ai/api/ran/2410.04847.svg)](https://syntology.ai/paper/2410.04847)

Connect an agent — have() is free.