SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2210.09477 · 2022

UniTune: Text-Driven Image Editing by Fine Tuning a Diffusion Model on a Single Image

arXiv · PDF · Open in the Atlas

Code that ran

10 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; 3 honoured the contract we drafted, 3 violated it and 0 have no contract verdict; 4 failures of our instrument; 3 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/2210.09477.json · badge: /api/ran/2210.09477.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
xuduo35/UniTune community (classifier label)
classifier basis: plan_row
10 ran · 3 no recorded run
FunctionStatusWhere it lives
chunk Ran
ran; our contract draft was wrong, not the code
xuduo35/UniTune/gr_stable_txt2img.py
pointer only (licence: MIT); the stored copy get_code serves, from vicgalle/stable-diffusion-aesthetic-gradients, is pointer only (licence: NOASSERTION) · get_code("8241c0562bc710fd")
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 vicgalle/stable-diffusion-aesthetic-gradients, 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
xuduo35/UniTune/ldm/modules/attention.py
code served (permissive licence) · get_code("424012cb37b31172")
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 int, ndim 0, dtype family int, 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.

from inspect import isfunction

def exists(val):
    return val is not None

def default(val, d):
    if exists(val):
        return val
    return d() if isfunction(d) else d

# Test case 1: val exists (not None), d is a regular value
ARGS_1 = {
    'val': 42,
    'd': 100
}

# Test case 2: val is None, d is a regular value
ARGS_2 = {
    'val': None,
    'd': 100
}

# Test case 3: val is None, d is a function
def default_func():
    return 200

ARGS_3 = {
    'val': None,
    'd': default_func
}

# Test case 4: val exists, d is a function
ARGS_4 = {
    'val': "existing_value",
    'd': lambda: "default_value"
}

# Use one of the test cases as the final ARGS
ARGS = ARGS_3
Setup block drafted by
a language model, Claude Sonnet 4.5, prompt python_fixture_v1
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type int, ndim 0, dtype family int, 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
exists Ran
ran, violated the contract we drafted
xuduo35/UniTune/ldm/modules/attention.py
pointer only (licence: MIT); the stored copy get_code serves, from ThomasMrY/VCT, is pointer only (licence: NONE) · get_code("aa5486a3650902d8")
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 ThomasMrY/VCT, 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.

get_bert_token_for_string Ran
ran, honoured the contract we drafted
xuduo35/UniTune/ldm/modules/embedding_manager.py
code served (permissive licence) · get_code("d5726f9e3a5684c3")
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 0, dtype family int, 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

torch.manual_seed(0)

class MockTokenizer:
    def __call__(self, string):
        # Return a tensor of shape (1, 3) with 3 non-zero tokens
        # This satisfies the assertion that torch.count_nonzero(token) == 3
        return torch.tensor([[101, 2054, 102]], dtype=torch.long)

tokenizer = MockTokenizer()
string = "hello world"

ARGS = {
    "tokenizer": tokenizer,
    "string": string,
}
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 0, dtype family int, 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_clip_token_for_string Ran
ran, honoured the contract we drafted
xuduo35/UniTune/ldm/modules/embedding_manager.py
code served (permissive licence) · get_code("41c6eb3a0aa3eb46")
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 0, dtype family int, 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

torch.manual_seed(0)

class MockTokenizer:
    def __call__(self, string, truncation=True, max_length=77, return_length=True, 
                 return_overflowing_tokens=False, padding='max_length', return_tensors='pt'):
        # Create a batch encoding that returns tokens
        # The function expects tokens where all values are 49407 except for exactly 2 positions
        # Position 0 and position 1 should have different values
        tokens = torch.full((1, 77), 49407, dtype=torch.long)
        tokens[0, 0] = 49406  # Different token at position 0
        tokens[0, 1] = 50256  # Different token at position 1 (this is what gets returned)
        
        return {
            'input_ids': tokens,
            'length': torch.tensor([77])
        }

tokenizer = MockTokenizer()
string = "test"

ARGS = {
    'tokenizer': tokenizer,
    'string': string
}
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 0, dtype family int, 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_clip_token_for_string Ran
ran, honoured the contract we drafted
xuduo35/UniTune/merge_embeddings.py
code served (permissive licence) · get_code("e466f56ae7a8bea6")
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 0, dtype family int, 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

torch.manual_seed(0)

class MockTokenizer:
    def __call__(self, string, truncation=True, max_length=77, return_length=True, return_overflowing_tokens=False, padding='max_length', return_tensors='pt'):
        # Return a batch encoding dict with input_ids
        # The function checks if torch.count_nonzero(tokens - 49407) == 2
        # This means exactly 2 tokens are not equal to 49407
        # Let's create tokens where most are 49407, but 2 are different
        tokens = torch.full((1, 77), 49407, dtype=torch.long)
        # Set positions 1 and some other position to different values
        tokens[0, 1] = 50000  # Different from 49407
        tokens[0, 50] = 50001  # Different from 49407
        
        return {
            'input_ids': tokens,
            'length': torch.tensor([77])
        }

tokenizer = MockTokenizer()
string = "test string"

ARGS = {
    'tokenizer': tokenizer,
    'string': string
}
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 0, dtype family int, 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_embedding_for_clip_token Ran
ran; our contract draft was wrong, not the code
xuduo35/UniTune/ldm/modules/embedding_manager.py
code served (permissive licence) · get_code("71c2df28a1008d03")
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 1, dtype family float, 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 torch
import torch.nn as nn

torch.manual_seed(0)

class SimpleEmbedder(nn.Module):
    def __init__(self):
        super().__init__()
        self.embedding = nn.Linear(1, 4)
    
    def forward(self, x):
        # x shape: (batch_size, 1)
        # output shape: (batch_size, 1, 4)
        out = self.embedding(x)
        return out.unsqueeze(1)

embedder = SimpleEmbedder()
token = torch.tensor([1], dtype=torch.float32)

ARGS = {
    "embedder": embedder,
    "token": token,
}
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 1, dtype family float, 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
get_placeholder_loop Ran
ran; our contract draft was wrong, not the code
xuduo35/UniTune/merge_embeddings.py
code served (permissive licence) · get_code("cb1a19e61d70046b")
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, ndim 1, 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 torch.nn as nn

torch.manual_seed(0)

# Stand-in tokenizer for CLIP (SD)
class CLIPTokenizer:
    def __call__(self, string, truncation=True, max_length=77, return_length=True, 
                 return_overflowing_tokens=False, padding='max_length', return_tensors='pt'):
        # Return a batch_encoding dict with input_ids
        # We need to construct tokens such that:
        # - torch.count_nonzero(tokens - 49407) == 2 for the function to return a token
        # - This means exactly 2 non-zero differences from 49407
        # Create a token tensor where most values are 49407 (pad token) and exactly 2 differ
        tokens = torch.full((1, 77), 49407, dtype=torch.long)
        tokens[0, 0] = 101  # Start token
        tokens[0, 1] = 1000  # Actual token (will be returned)
        return {'input_ids': tokens}

# Stand-in tokenizer for BERT
class BertTokenizer:
    def __call__(self, string):
        # Return a token tensor where exactly 3 elements are non-zero
        tokens = torch.zeros((1, 10), dtype=torch.long)
        tokens[0, 0] = 101    # CLS token
        tokens[0, 1] = 2054   # Token to return
        tokens[0, 2] = 102    # SEP token
        return tokens

# Stand-in embedder for SD (uses CLIP tokenizer)
class SDEmbedder:
    def __init__(self):
        self.tokenizer = CLIPTokenizer()

# Stand-in embedder for BERT
class BertEmbedder:
    def __init__(self):
        self.tknz_fn = BertTokenizer()

# Mock input function - return a valid placeholder string
original_input = input
def mock_input(prompt):
    return "valid_placeholder"

# Replace input with mock during setup
import builtins
builtins.input = mock_input

ARGS = {
    'placeholder_string': 'old_placeholder',
    'embedder': SDEmbedder(),
    'is_sd': 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 tuple, ndim 1, dtype family float — a tuple output has no array shape of its own; none is stated
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
numpy_to_pil Ran
ran, violated the contract we drafted
xuduo35/UniTune/gr_stable_txt2img.py
pointer only (licence: MIT); the stored copy get_code serves, from isno0907/isodiff, is pointer only (licence: NONE) · get_code("a3f07f0c1a9623c3")
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 isno0907/isodiff, 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.

uniq Ran
ran; our contract draft was wrong, not the code
xuduo35/UniTune/ldm/modules/attention.py
pointer only (licence: MIT); the stored copy get_code serves, from JingWu321/MUNBa, is pointer only (licence: NONE) · get_code("9a299fe5ae09e407")
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 JingWu321/MUNBa, 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.

get_bert_token_for_string Not yet run
unverified (harvested, no recorded run)
xuduo35/UniTune/merge_embeddings.py
code served (permissive licence) · get_code("91610d6e2f2a67b4")
How it ran

no invocation recorded.

get_full_repo_name Not yet run
unverified (harvested, no recorded run)
xuduo35/UniTune/train_unitune_colab.py
pointer only (licence: MIT); the stored copy get_code serves, from guoqiang-zhang-x/BDIA, is pointer only (licence: NONE) · get_code("d7a3942d0d5c6041")
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 guoqiang-zhang-x/BDIA, 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. No invocation recorded.

parse_args Not yet run
unverified (harvested, no recorded run)
xuduo35/UniTune/train_unitune_colab.py
code served (permissive licence) · get_code("fadd22e702a7c11d")
How it ran

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
exists 486 other papers
116 dated earlier, 9 the same month, 358 later, 1 same year, order not determinable, 2 undated · 78 only as an assembly brick
1603.08155 (2016-03)
Perceptual Losses for Real-Time Style Transfer and Super-Resolution
1611.07004 (2016-11)
Image-to-Image Translation with Conditional Adversarial Networks
1804.04637 (2018-04)
EMBER: An Open Dataset for Training Static PE Malware Machine Learning Models
and 483 more not named here
default 207 other papers
23 dated earlier, 2 the same month, 182 later · 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
uniq 115 other papers
4 dated earlier, 1 the same month, 110 later
2112.10752 (2021-12)
High-Resolution Image Synthesis with Latent Diffusion Models
2204.06125 (2022-04)
Hierarchical Text-Conditional Image Generation with CLIP Latents
2206.02779 (2022-06)
Blended Latent Diffusion
and 112 more not named here
chunk 40 other papers
5 dated earlier, 1 the same month, 34 later
2106.14490 (2021-06)
Making Images Real Again: A Comprehensive Survey on Deep Image Composition
2112.10752 (2021-12)
High-Resolution Image Synthesis with Latent Diffusion Models
2204.11824 (2022-04)
Semi-Parametric Neural Image Synthesis
and 37 more not named here
get_full_repo_name 17 other papers
17 later
2211.15388 (2022-11)
Shifted Diffusion for Text-to-image Generation
2302.07944 (2023-02)
Effective Data Augmentation With Diffusion Models
2303.09522 (2023-03)
P+: Extended Textual Conditioning in Text-to-Image Generation
and 14 more not named here
get_bert_token_for_string 11 other papers
1 dated earlier, 10 later
2208.01618 (2022-08)
An Image is Worth One Word: Personalizing Text-to-Image Generation using Textual Inversion
2211.09800 (2022-11)
InstructPix2Pix: Learning to Follow Image Editing Instructions
2212.04489 (2022-12)
SINE: SINgle Image Editing with Text-to-Image Diffusion Models
and 8 more not named here
get_clip_token_for_string 9 other papers
1 dated earlier, 8 later
2208.01618 (2022-08)
An Image is Worth One Word: Personalizing Text-to-Image Generation using Textual Inversion
2211.09800 (2022-11)
InstructPix2Pix: Learning to Follow Image Editing Instructions
2212.04489 (2022-12)
SINE: SINgle Image Editing with Text-to-Image Diffusion Models
and 6 more not named here
get_embedding_for_clip_token 9 other papers
1 dated earlier, 8 later
2208.01618 (2022-08)
An Image is Worth One Word: Personalizing Text-to-Image Generation using Textual Inversion
2211.09800 (2022-11)
InstructPix2Pix: Learning to Follow Image Editing Instructions
2212.04489 (2022-12)
SINE: SINgle Image Editing with Text-to-Image Diffusion Models
and 6 more not named here
get_bert_token_for_string 4 other papers
4 later
2304.08818 (2023-04)
Align your Latents: High-Resolution Video Synthesis with Latent Diffusion Models
2306.00926 (2023-06)
Inserting Anybody in Diffusion Models via Celeb Basis
2307.08585 (2023-07)
Identity-Preserving Aging of Face Images via Latent Diffusion Models
and 1 more not named here
get_clip_token_for_string 4 other papers
4 later
2304.08818 (2023-04)
Align your Latents: High-Resolution Video Synthesis with Latent Diffusion Models
2306.00926 (2023-06)
Inserting Anybody in Diffusion Models via Celeb Basis
2307.08585 (2023-07)
Identity-Preserving Aging of Face Images via Latent Diffusion Models
and 1 more not named here
get_placeholder_loop 4 other papers
4 later
2304.08818 (2023-04)
Align your Latents: High-Resolution Video Synthesis with Latent Diffusion Models
2306.00926 (2023-06)
Inserting Anybody in Diffusion Models via Celeb Basis
2307.08585 (2023-07)
Identity-Preserving Aging of Face Images via Latent Diffusion Models
and 1 more not named here
numpy_to_pil 4 other papers
4 later
2307.11410 (2023-07)
Subject-Diffusion:Open Domain Personalized Text-to-Image Generation without Test-time Fine-tuning
2308.12038 (2023-08)
Large Multilingual Models Pivot Zero-Shot Multimodal Learning across Languages
2407.11451 (2024-07)
Isometric Representation Learning for Disentangled Latent Space of Diffusion Models
and 1 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

Text-driven image generation methods have shown impressive results recently, allowing casual users to generate high quality images by providing textual descriptions. However, similar capabilities for editing existing images are still out of reach. Text-driven image editing methods usually need edit masks, struggle with edits that require significant visual changes and cannot easily keep specific details of the edited portion. In this paper we make the observation that image-generation models can be converted to image-editing models simply by fine-tuning them on a single image. We also show that initializing the stochastic sampler with a noised version of the base image before the sampling and interpolating relevant details from the base image after sampling further increase the quality of the edit operation. Combining these observations, we propose UniTune, a novel image editing method. UniTune gets as input an arbitrary image and a textual edit description, and carries out the edit while maintaining high fidelity to the input image. UniTune does not require additional inputs, like masks or sketches, and can perform multiple edits on the same image without retraining. We test our method using the Imagen model in a range of different use cases. We demonstrate that it is broadly applicable and can perform a surprisingly wide range of expressive editing operations, including those requiring significant visual changes that were previously impossible.

For agents

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

get_harvested_code_for_paper("2210.09477")
get_code_for_paper("2210.09477")
have("2210.09477")

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

curl https://syntology.ai/api/ran/2210.09477.json

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

[![Syntology run record](https://syntology.ai/api/ran/2210.09477.svg)](https://syntology.ai/paper/2210.09477)

Connect an agent — have() is free.