SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2305.12966 · NeurIPS · 2023

Hierarchical Integration Diffusion Model for Realistic Image Deblurring

Xin Yuan, Yulun Zhang, Jinjin Gu, Linghe Kong, Zheng Chen, Ding Liu, Bin Xia

arXiv · PDF · Open in the Atlas

Code that ran

11 of 13 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, 3 violated it and 0 have no contract verdict; 7 failures 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/2305.12966.json · badge: /api/ran/2305.12966.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
zhengchen1999/HI-Diff canonical (classifier label)
classifier basis: census
11 ran · 2 no recorded run
FunctionStatusWhere it lives
augment Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
zhengchen1999/HI-Diff/hi_diff/utils/transforms.py
code served (permissive licence) · get_code("7f124cd33fb750ee")
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
import cv2
import random

random.seed(0)
np.random.seed(0)

# Create sample images as ndarrays (height=4, width=3, channels=3)
img1 = np.random.randint(0, 256, (4, 3, 3), dtype=np.uint8)
img2 = np.random.randint(0, 256, (4, 3, 3), dtype=np.uint8)

# Create sample flows as ndarrays (height=4, width=3, 2 channels for x,y)
flow1 = np.random.randn(4, 3, 2).astype(np.float32)
flow2 = np.random.randn(4, 3, 2).astype(np.float32)

ARGS = {
    'imgs': [img1, img2],
    'hflip': True,
    'rotation': True,
    'flows': [flow1, flow2],
    'return_status': False
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
cv2, 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_psnr Ran
ran, violated the contract we drafted
zhengchen1999/HI-Diff/evaluate_realblur.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from invokerer/deeprft, is pointer only (licence: NOASSERTION) · get_code("91fa50e293bf2fde")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from invokerer/deeprft, and GitHub could not classify the licence of that repository (NOASSERTION). 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.

default Ran
ran, violated the contract we drafted
zhengchen1999/HI-Diff/hi_diff/utils/beta_schedule.py
code served (permissive licence) · get_code("fbf9ec7be545688e")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Sonnet 4.5, prompt python_fixture_v1) and returned str. 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 numpy as np
from inspect import isfunction

torch.manual_seed(0)

# Test case 1: val exists (not None), d is a regular value
val_case1 = 42
d_case1 = 100

# Test case 2: val is None, d is a regular value
val_case2 = None
d_case2 = "default_string"

# Test case 3: val is None, d is a function
val_case3 = None
def d_case3():
    return [1, 2, 3]

# Test case 4: val exists, d is a function (function should not be called)
val_case4 = torch.tensor([1.0, 2.0, 3.0])
def d_case4():
    return torch.tensor([4.0, 5.0, 6.0])

# We'll use case 2 for the main ARGS
ARGS = {
    'val': None,
    'd': "default_value"
}
Setup block drafted by
a language model, Claude Sonnet 4.5, prompt python_fixture_v1
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str 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
dict2str Ran
ran; our contract draft was wrong, not the code
zhengchen1999/HI-Diff/hi_diff/utils/options.py
code served (permissive licence) · get_code("769e9c110eeb4a0d")
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 str. 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

# Simple nested dictionary with various types of values
opt = {
    'model_name': 'resnet50',
    'learning_rate': 0.001,
    'batch_size': 32,
    'use_cuda': False,
    'training': {
        'epochs': 100,
        'optimizer': 'adam',
        'loss': {
            'type': 'cross_entropy',
            'weight': 0.5
        }
    },
    'augmentation': {
        'flip': True,
        'rotate': 90
    }
}

ARGS = {
    'opt': opt,
    'indent_level': 1
}
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 str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str 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
exists Ran
ran, violated the contract we drafted
zhengchen1999/HI-Diff/hi_diff/utils/beta_schedule.py
code served (permissive licence) · get_code("608e364a9d2376a3")
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 bool, ndim 0, dtype family float, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

def exists(x):
    return x is not None

ARGS = {
    "x": torch.tensor([1, 2, 3, 4])
}
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 bool, ndim 0, dtype family float, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
get_position_from_periods Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
zhengchen1999/HI-Diff/hi_diff/utils/lr_scheduler.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from AlifAshrafee/Restoring-Without-Forgetting, is pointer only (licence: NONE) · get_code("cd569444547de84f")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from AlifAshrafee/Restoring-Without-Forgetting, 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.

make_beta_schedule Ran
ran, honoured the contract we drafted
zhengchen1999/HI-Diff/hi_diff/utils/beta_schedule.py
code served (permissive licence) · get_code("78feb2965623416b")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned ndarray, ndim 1, dtype family float, shape (100,) 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
import math

torch.manual_seed(0)

ARGS = {
    'schedule': 'linear',
    'n_timestep': 100,
    'linear_start': 0.0001,
    'linear_end': 0.02,
    'cosine_s': 0.008
}
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 ndarray, ndim 1, dtype family float, shape (100,) — shape (100,) 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
mod_crop Ran
ran; our contract draft was wrong, not the code
zhengchen1999/HI-Diff/hi_diff/utils/transforms.py
code served (permissive licence) · get_code("223d0788b0725dfb")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned ndarray, ndim 3, dtype family int, shape (12, 16, 3) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

# Create a small test image (2D grayscale)
img_2d = np.random.randint(0, 256, size=(10, 15), dtype=np.uint8)

# Create a small test image (3D RGB)
img_3d = np.random.randint(0, 256, size=(12, 18, 3), dtype=np.uint8)

scale = 4

ARGS = {
    "img": img_3d,
    "scale": scale,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 3, dtype family int, shape (12, 16, 3) — shape (12, 16, 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
paired_random_crop Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
zhengchen1999/HI-Diff/hi_diff/utils/transforms.py
code served (permissive licence) · get_code("64cd160c586b1bef")
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 torch
import numpy as np
import random

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

# Create LQ (low quality) image with dimensions (batch, channels, height, width)
# LQ size: 2x3x8x8
img_lq_tensor = torch.randn(2, 3, 8, 8)

# Create GT (ground truth) image with scale factor 2, so GT should be 2x scale larger
# GT size: 2x3x16x16 (8*2 = 16)
img_gt_tensor = torch.randn(2, 3, 16, 16)

# Alternative numpy versions for testing (not used here but shown for reference)
# img_lq_numpy = np.random.randn(8, 8, 3)
# img_gt_numpy = np.random.randn(16, 16, 3)

ARGS = {
    'img_gts': img_gt_tensor,
    'img_lqs': img_lq_tensor,
    'gt_patch_size': 8,
    'scale': 2,
    'gt_path': 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 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
to_3d Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
zhengchen1999/HI-Diff/hi_diff/archs/Transformer_arch.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from gladzhang/Xformer, is pointer only (licence: NONE) · get_code("82a15cc1e46f7e4d")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from gladzhang/Xformer, 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.

to_4d Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
zhengchen1999/HI-Diff/hi_diff/archs/Transformer_arch.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from gladzhang/Xformer, is pointer only (licence: NONE) · get_code("b20f2a5df739a59e")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from gladzhang/Xformer, 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.

compute_ssim Not yet run
unverified (harvested, no recorded run)
zhengchen1999/HI-Diff/evaluate_realblur.py
code served (permissive licence) · get_code("64c776bb050ba1e2")
How it ran

no invocation recorded.

image_align Not yet run
unverified (harvested, no recorded run)
zhengchen1999/HI-Diff/evaluate_realblur.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from INVOKERer/AdaRevD, is pointer only (licence: NOASSERTION) · get_code("5aad1c1e328fd4ec")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from INVOKERer/AdaRevD, and GitHub could not classify the licence of that repository (NOASSERTION). 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. No invocation recorded.

Also present in other papers' code

Of 13 samples lifted from this paper's linked repositories, 13 carry a function identity and 0 do not. Of the 13 distinct functions with an identity, 12 are also present in other papers whose code we hold (7 with at least one paper dated earlier, 4 the same month, 12 later) and 1 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
default 207 other papers
51 dated earlier, 8 the same month, 146 later, 2 same year, order not determinable · 41 only as an assembly brick
1711.11585 (2017-11)
High-Resolution Image Synthesis and Semantic Manipulation with Conditional GANs
2006.11239 (2020-06)
Denoising Diffusion Probabilistic Models
2010.08162 (2020-10)
SidechainNet: An All-Atom Protein Structure Dataset for Machine Learning
and 204 more not named here
exists 144 other papers
37 dated earlier, 3 the same month, 103 later, 1 same year, order not determinable · 40 only as an assembly brick
1611.07004 (2016-11)
Image-to-Image Translation with Conditional Adversarial Networks
1711.11585 (2017-11)
High-Resolution Image Synthesis and Semantic Manipulation with Conditional GANs
2006.11239 (2020-06)
Denoising Diffusion Probabilistic Models
and 141 more not named here
to_3d 80 other papers
13 dated earlier, 3 the same month, 62 later, 2 same year, order not determinable · 18 only as an assembly brick
2003.06792 (2020-03)
Learning Enriched Features for Real Image Restoration and Enhancement
2102.02808 (2021-02)
Multi-Stage Progressive Image Restoration
2111.09881 (2021-11)
Restormer: Efficient Transformer for High-Resolution Image Restoration
and 77 more not named here
to_4d 79 other papers
13 dated earlier, 3 the same month, 61 later, 2 same year, order not determinable · 18 only as an assembly brick
2003.06792 (2020-03)
Learning Enriched Features for Real Image Restoration and Enhancement
2102.02808 (2021-02)
Multi-Stage Progressive Image Restoration
2111.09881 (2021-11)
Restormer: Efficient Transformer for High-Resolution Image Restoration
and 76 more not named here
get_position_from_periods 55 other papers
9 dated earlier, 44 later, 2 same year, order not determinable
2110.04627 (2021-10)
Vector-quantized Image Modeling with Improved VQGAN
136790177 (2022 (year only))
Frequency and Spatial Dual Guidance for Image Dehazing
2204.04676 (2022-04)
Simple Baselines for Image Restoration
and 52 more not named here
compute_psnr 5 other papers
1 dated earlier, 4 later
2201.02973 (2022-01)
MAXIM: Multi-Axis MLP for Image Processing
2401.00027 (2024-01)
Efficient Multi-scale Network with Learnable Discrete Wavelet Transform for Blind Motion Deblurring
2404.13153 (2024-04)
Motion-adaptive Separable Collaborative Filters for Blind Motion Deblurring
and 2 more not named here
image_align 5 other papers
1 dated earlier, 4 later
2201.02973 (2022-01)
MAXIM: Multi-Axis MLP for Image Processing
2401.00027 (2024-01)
Efficient Multi-scale Network with Learnable Discrete Wavelet Transform for Blind Motion Deblurring
2404.13153 (2024-04)
Motion-adaptive Separable Collaborative Filters for Blind Motion Deblurring
and 2 more not named here
mod_crop 3 other papers
3 later
2407.13987 (2024-07)
RealViformer: Investigating Attention for Real-World Video Super-Resolution
2411.13383 (2024-11)
Adversarial Diffusion Compression for Real-World Image Super-Resolution
augment 2 other papers
2 later
2411.13383 (2024-11)
Adversarial Diffusion Compression for Real-World Image Super-Resolution
paired_random_crop 2 other papers
2 later
2411.13383 (2024-11)
Adversarial Diffusion Compression for Real-World Image Super-Resolution
dict2str 1 other paper
1 later
2406.05723 (2024-06)
Binarized Diffusion Model for Image Super-Resolution
make_beta_schedule 1 other paper
1 later
2406.05723 (2024-06)
Binarized Diffusion Model for Image Super-Resolution

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

Diffusion models (DMs) have recently been introduced in image deblurring and exhibited promising performance, particularly in terms of details reconstruction. However, the diffusion model requires a large number of inference iterations to recover the clean image from pure Gaussian noise, which consumes massive computational resources. Moreover, the distribution synthesized by the diffusion model is often misaligned with the target results, leading to restrictions in distortionbased metrics. To address the above issues, we propose the Hierarchical Integration Diffusion Model (HI-Diff), for realistic image deblurring. Specifically, we perform the DM in a highly compacted latent space to generate the prior feature for the deblurring process. The deblurring process is implemented by a regression-based method to obtain better distortion accuracy. Meanwhile, the highly compact latent space ensures the efficiency of the DM. Furthermore, we design the hierarchical integration module to fuse the prior into the regression-based model from multiple scales, enabling better generalization in complex blurry scenarios. Comprehensive experiments on synthetic and real-world blur datasets demonstrate that our HI-Diff outperforms state-of-the-art methods. Code and trained models are available at https://github.com/zhengchen1999/HI-Diff.

For agents

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

get_harvested_code_for_paper("2305.12966")
get_code_for_paper("2305.12966")
have("2305.12966")

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

curl https://syntology.ai/api/ran/2305.12966.json

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

[![Syntology run record](https://syntology.ai/api/ran/2305.12966.svg)](https://syntology.ai/paper/2305.12966)

Connect an agent — have() is free.