SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2306.07280 · NeurIPS · 2023

Controlling Text-to-Image Diffusion by Orthogonal Finetuning

Bernhard Schölkopf, Adrian Weller, Dan Zhang, Zhen Liu, Weiyang Liu, Zeju Qiu, Haiwen Feng, Yuxuan Xue, Yao Feng

arXiv · PDF · Open in the Atlas

Code that ran

12 of 22 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (1 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 7 have no contract verdict; 5 failures of our instrument; 10 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/2306.07280.json · badge: /api/ran/2306.07280.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
zeju1997/oft community (classifier label)
classifier basis not recorded on the link
9 ran · 10 no recorded run
zqiu24/oft community
Papers with Code archive
3 ran · 0 no recorded run
FunctionStatusWhere it lives
OFTInjectedLinear Ran
ran on a synthesized input · constructed an object; no result was computed
zqiu24/oft/oft-control/oft.py
code served (permissive licence) · get_code("e85418d3d1fc5999")
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, OFTInjectedLinear: the run constructed the object and did not call it, so no result was computed. 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 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 = {
    "in_features": 8,
    "out_features": 6,
    "bias": False,
    "r": 4,
    "eps": 1e-05,
    "is_coft": True,
    "block_share": 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 OFTInjectedLinear — 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 OFTInjectedLinear 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
calculate_metrics Ran
ran on a synthesized input
zeju1997/oft/oft-control/eval_canny.py
code served (permissive licence) · get_code("a73565d61cbda080")
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 numpy as np

np.random.seed(0)

# Create small binary prediction and target arrays
pred = np.array([
    [1, 0, 1, 0],
    [1, 1, 0, 1]
], dtype=bool)

target = np.array([
    [1, 0, 0, 1],
    [1, 1, 1, 0]
], dtype=bool)

ARGS = {
    "pred": pred,
    "target": target
}
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, 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
collate_fn Ran
ran; our contract draft was wrong, not the code
zeju1997/oft/oft-db/hyperparameter_oft.py
pointer only (licence: MIT); the stored copy get_code serves, from vinairesearch/anti-dreambooth, is pointer only (licence: AGPL-3.0) · get_code("229d634611bdc645")
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 vinairesearch/anti-dreambooth, and AGPL-3.0, as GitHub detects it on that repository, is not in the permissive set Syntology serves inline. 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_directories_missing_folders Ran
ran on a synthesized input
zeju1997/oft/oft-db/check_empty.py
code served (permissive licence) · get_code("bca0e14a517d825c")
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 list, 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 os
import tempfile
import shutil

# Create a temporary directory structure for testing
temp_dir = tempfile.mkdtemp()

# Create some test directories matching the pattern used in the function
subject_names = ['backpack', 'backpack_dog', 'bear_plushie', 'berry_bowl', 'can', 'candle', 'cat', 'cat2', 'clock', 'colorful_sneaker', 'dog', 'dog2', 'dog3', 'dog5', 'dog6', 'dog7', 'dog8', 'duck_toy', 'fancy_boot', 'grey_sloth_plushie', 'monster_toy', 'pink_sunglasses', 'poop_emoji', 'rc_car', 'red_cartoon', 'robot_toy', 'shiny_sneaker', 'teapot', 'vase', 'wolf_plushie']

# Create some subdirectories with folders
for subject in subject_names[:3]:  # Create a few test subjects
    for y in range(3):  # Create a few variations
        subdir_name = f'{subject}-{y}'
        subdir_path = os.path.join(temp_dir, subdir_name)
        os.makedirs(subdir_path, exist_ok=True)
        
        # Create some folders inside (matching the folder_names list)
        for folder in ['folder1', 'folder2']:
            folder_path = os.path.join(subdir_path, folder)
            os.makedirs(folder_path, exist_ok=True)

# Create a directory that's missing some folders
incomplete_subdir = os.path.join(temp_dir, 'backpack-0')
os.makedirs(incomplete_subdir, exist_ok=True)
# Only create one folder instead of two
os.makedirs(os.path.join(incomplete_subdir, 'folder1'), exist_ok=True)

# Define the expected folder names
folder_names = ['folder1', 'folder2']

ARGS = {
    'directory': temp_dir,
    'folder_names': folder_names
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
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 list, ndim 1, dtype family float — a list 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
find_subdirectory_names_without_png Ran
ran on a synthesized input
zeju1997/oft/oft-db/check_empty.py
code served (permissive licence) · get_code("33aebbf50ad78324")
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 list, 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 os
import tempfile
import fnmatch

subject_names = ['backpack', 'backpack_dog', 'bear_plushie', 'berry_bowl', 'can', 'candle', 'cat', 'cat2', 'clock', 'colorful_sneaker', 'dog', 'dog2', 'dog3', 'dog5', 'dog6', 'dog7', 'dog8', 'duck_toy', 'fancy_boot', 'grey_sloth_plushie', 'monster_toy', 'pink_sunglasses', 'poop_emoji', 'rc_car', 'red_cartoon', 'robot_toy', 'shiny_sneaker', 'teapot', 'vase', 'wolf_plushie']

# Create a temporary directory structure for testing
temp_dir = tempfile.mkdtemp()

# Create some subdirectories with and without PNG files
os.makedirs(os.path.join(temp_dir, 'subdir_with_png'), exist_ok=True)
os.makedirs(os.path.join(temp_dir, 'subdir_without_png'), exist_ok=True)
os.makedirs(os.path.join(temp_dir, 'backpack-0'), exist_ok=True)

# Add a PNG file to one subdirectory
png_file = os.path.join(temp_dir, 'subdir_with_png', 'test.png')
with open(png_file, 'w') as f:
    f.write('dummy png content')

# The function expects a directory path as argument
ARGS = {
    'directory': temp_dir
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
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 list, ndim 1, dtype family float — a list 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
get_node_name Ran
ran; our contract draft was wrong, not the code
zeju1997/oft/oft-control/tool_add_control_oft.py
code served (permissive licence) · get_code("0dbf855dab1ab575")
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 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.

# Analyzing the function get_node_name
# It takes two string parameters: name and parent_name
# It performs string operations: len(), slicing, comparison
# Returns a tuple of (bool, str)

ARGS = {
    'name': 'parent_module.child_layer',
    'parent_name': 'parent_module.'
}
Setup block drafted by
a language model, Claude Haiku 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 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
get_params_to_save Ran
ran on a synthesized input
zeju1997/oft/oft-db/train_dreambooth_flax.py
code served (permissive licence) · get_code("320666d58c95245c")
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 dict. 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 jax
import jax.numpy as jnp

# The function uses jax.device_get and jax.tree_util.tree_map
# It applies a lambda that indexes the first element [0] of each leaf
# So params should be a pytree where each leaf is indexable (like an array)

# Create a nested pytree structure with jax arrays
# Each array needs to support indexing with [0]
params = {
    'layer1': {
        'weights': jnp.ones((2, 4)),
        'bias': jnp.ones((2, 3))
    },
    'layer2': {
        'weights': jnp.ones((2, 5)),
        'bias': jnp.ones((2, 6))
    },
    'scalar_array': jnp.ones((2,))
}

ARGS = {
    'params': params
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
jax — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict 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_state_dict Ran
ran; our contract draft was wrong, not the code
zeju1997/oft/oft-control/oldm/model.py
code served (permissive licence) · get_code("cfc17707f35e7ec4")
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 dict. 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

torch.manual_seed(0)

def get_state_dict(d):
    return d.get('state_dict', d)

# The function calls .get() on the parameter 'd', so it must be a dict-like object.
# It tries to get 'state_dict' key, and if not found, returns the dict itself.

ARGS = {
    'd': {'state_dict': {'layer1.weight': torch.randn(4, 3), 'layer1.bias': torch.randn(4)}}
}
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 dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict 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
plot_kpts Ran
ran on a synthesized input
zeju1997/oft/oft-control/eval_landmark.py
code served (permissive licence) · get_code("1ff5174d4258a8ab")
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 (200, 300, 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
import cv2

# Create a small test image (height=200, width=300, channels=3)
image = np.random.randint(0, 256, (200, 300, 3), dtype=np.uint8)

# Create keypoints array with shape (68, 3)
# Each row: [x, y, confidence]
np.random.seed(0)
kpts = np.random.rand(68, 3) * 100
kpts[:, 0] = np.random.rand(68) * 300  # x coordinates within image width
kpts[:, 1] = np.random.rand(68) * 200  # y coordinates within image height
kpts[:, 2] = np.random.rand(68)        # confidence scores

ARGS = {
    'image': image,
    'kpts': kpts,
    'color': 'g'
}
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 ndarray, ndim 3, dtype family int, shape (200, 300, 3) — shape (200, 300, 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
plot_points Ran
ran on a synthesized input
zeju1997/oft/oft-control/eval_landmark.py
code served (permissive licence) · get_code("9320471cb08ee42e")
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 (100, 150, 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
import cv2

# Create a small test image (BGR format, as cv2 expects)
image = np.random.randint(0, 256, size=(100, 150, 3), dtype=np.uint8)

# Create keypoints array with shape (n, 3)
# 68 keypoints would be typical for facial landmarks, but we'll use fewer for testing
kpts = np.array([
    [25.5, 30.2, 0.9],
    [50.1, 40.3, 0.8],
    [75.8, 35.5, 0.85],
    [40.2, 60.1, 0.9],
], dtype=np.float32)

# Color parameter - one of 'r', 'g', 'b', 'y', 'w'
color = 'w'

ARGS = {
    'image': image,
    'kpts': kpts,
    'color': color,
}
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 ndarray, ndim 3, dtype family int, shape (100, 150, 3) — shape (100, 150, 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
project Ran
ran; our contract draft was wrong, not the code
zqiu24/oft/oft-control/oft.py
code served (permissive licence) · get_code("ccec8c1b07c1a3b6")
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 (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)

# R should be a square tensor (the function accesses R.size(0) twice and creates an identity matrix of that size)
# eps should be a float threshold for the norm
R = torch.randn(4, 4, dtype=torch.float32)
eps = 0.5

ARGS = {
    "R": R,
    "eps": eps
}
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 (4, 4) — shape (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
project_batch Ran
ran; our contract draft was wrong, not the code
zqiu24/oft/oft-control/oft.py
code served (permissive licence) · get_code("4cd1d6571b9e35a1")
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 (2, 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)

# R should be a batch of square matrices
# Looking at the function:
# - R.shape[0] is the batch size
# - R.size(1) and R.size(2) should be equal (square matrices)
# - Operations: R.shape[0], R.size(1), R.size(2), R.device, R.dtype
# - torch.norm with dim=(1, 2) expects 3D tensor (batch, rows, cols)

batch_size = 2
matrix_dim = 4

# Create a batch of 2x4x4 matrices (batch of square matrices)
R = torch.randn(batch_size, matrix_dim, matrix_dim, dtype=torch.float32)

ARGS = {
    "R": R,
    "eps": 1e-05
}
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 (2, 4, 4) — shape (2, 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
clip_image Not yet run
unverified (harvested, no recorded run)
zeju1997/oft/oft-db/eval_ablation.py
code served (permissive licence) · get_code("862b8dc1d24afa80")
How it ran

no invocation recorded.

clip_text Not yet run
unverified (harvested, no recorded run)
zeju1997/oft/oft-db/eval_ablation.py
code served (permissive licence) · get_code("028c10fb01328ebd")
How it ran

no invocation recorded.

dino Not yet run
unverified (harvested, no recorded run)
zeju1997/oft/oft-db/eval_ablation.py
code served (permissive licence) · get_code("8c1f9091bcfcc35f")
How it ran

no invocation recorded.

get_full_repo_name Not yet run
unverified (harvested, no recorded run)
zeju1997/oft/oft-db/train_dreambooth_flax.py
pointer only (licence: MIT); the stored copy get_code serves, from isno0907/isodiff, is pointer only (licence: NONE) · get_code("6431708d8f3ff177")
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. No invocation recorded.

import_model_class_from_model_name_or_path Not yet run
unverified (harvested, no recorded run)
zeju1997/oft/oft-db/hyperparameter_oft.py
code served (permissive licence) · get_code("2f00b39db5a81466")
How it ran

no invocation recorded.

inject_trainable_oft Not yet run
unverified (harvested, no recorded run)
zeju1997/oft/oft-control/oft.py
code served (permissive licence) · get_code("5cfa761d3c77fa2d")
How it ran

no invocation recorded.

load_state_dict Not yet run
unverified (harvested, no recorded run)
zeju1997/oft/oft-control/oldm/model.py
code served (permissive licence) · get_code("df733a879693145d")
How it ran

no invocation recorded.

parse_args Not yet run
unverified (harvested, no recorded run)
zeju1997/oft/oft-db/hyperparameter_oft.py
code served (permissive licence) · get_code("9a1b8f02d12708b3")
How it ran

no invocation recorded.

parse_args Not yet run
unverified (harvested, no recorded run)
zeju1997/oft/oft-db/train_dreambooth.py
code served (permissive licence) · get_code("f5132e29bee34d4d")
How it ran

no invocation recorded.

parse_args Not yet run
unverified (harvested, no recorded run)
zeju1997/oft/oft-db/train_dreambooth_lora.py
code served (permissive licence) · get_code("2f3af25e8d9d78f5")
How it ran

no invocation recorded.

Also present in other papers' code

Of 22 samples lifted from this paper's linked repositories, 22 carry a function identity and 0 do not. Of the 22 distinct functions with an identity, 6 are also present in other papers whose code we hold (6 with at least one paper dated earlier, 1 the same month, 6 later) and 16 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_state_dict 29 other papers
4 dated earlier, 25 later
2304.04269 (2023-04)
HumanSD: A Native Skeleton-Guided Diffusion Model for Human Image Generation
2305.02034 (2023-05)
SAMRS: Scaling-up Remote Sensing Segmentation Dataset with Segment Anything Model
2305.11147 (2023-05)
UniControl: A Unified Diffusion Model for Controllable Visual Generation In the Wild
and 26 more not named here
load_state_dict 25 other papers
4 dated earlier, 21 later
2304.04269 (2023-04)
HumanSD: A Native Skeleton-Guided Diffusion Model for Human Image Generation
2305.02034 (2023-05)
SAMRS: Scaling-up Remote Sensing Segmentation Dataset with Segment Anything Model
2305.11147 (2023-05)
UniControl: A Unified Diffusion Model for Controllable Visual Generation In the Wild
and 22 more not named here
import_model_class_from_model_name_or_path 24 other papers
1 dated earlier, 1 the same month, 22 later
2212.04489 (2022-12)
SINE: SINgle Image Editing with Text-to-Image Diffusion Models
2306.06874 (2023-06)
VillanDiffusion: A Unified Backdoor Attack Framework for Diffusion Models
2310.04687 (2023-10)
Targeted Attack Improves Protection against Unauthorized Diffusion Customization
and 21 more not named here
get_full_repo_name 17 other papers
9 dated earlier, 8 later
2210.09477 (2022-10)
UniTune: Text-Driven Image Editing by Fine Tuning a Diffusion Model on a Single Image
2211.15388 (2022-11)
Shifted Diffusion for Text-to-image Generation
2302.07944 (2023-02)
Effective Data Augmentation With Diffusion Models
and 14 more not named here
get_node_name 15 other papers
2 dated earlier, 13 later
2304.04269 (2023-04)
HumanSD: A Native Skeleton-Guided Diffusion Model for Human Image Generation
2305.11147 (2023-05)
UniControl: A Unified Diffusion Model for Controllable Visual Generation In the Wild
2309.03453 (2023-09)
SyncDreamer: Generating Multiview-consistent Images from a Single-view Image
and 12 more not named here
collate_fn 7 other papers
2 dated earlier, 5 later
2208.12242 (2022-08)
DreamBooth: Fine Tuning Text-to-Image Diffusion Models for Subject-Driven Generation
2303.15433 (2023-03)
Anti-DreamBooth: Protecting users from personalized text-to-image synthesis
2310.04687 (2023-10)
Targeted Attack Improves Protection against Unauthorized Diffusion Customization
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

Large text-to-image diffusion models have impressive capabilities in generating photorealistic images from text prompts. How to effectively guide or control these powerful models to perform different downstream tasks becomes an important open problem. To tackle this challenge, we introduce a principled finetuning method -Orthogonal Finetuning (OFT), for adapting text-to-image diffusion models to downstream tasks. Unlike existing methods, OFT can provably preserve hyperspherical energy which characterizes the pairwise neuron relationship on the unit hypersphere. We find that this property is crucial for preserving the semantic generation ability of text-to-image diffusion models. To improve finetuning stability, we further propose Constrained Orthogonal Finetuning (COFT) which imposes an additional radius constraint to the hypersphere. Specifically, we consider two important finetuning text-to-image tasks: subject-driven generation where the goal is to generate subject-specific images given a few images of a subject and a text prompt, and controllable generation where the goal is to enable the model to take in additional control signals. We empirically show that our OFT framework outperforms existing methods in generation quality and convergence speed.

For agents

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

get_harvested_code_for_paper("2306.07280")
get_code_for_paper("2306.07280")
have("2306.07280")

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

curl https://syntology.ai/api/ran/2306.07280.json

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

[![Syntology run record](https://syntology.ai/api/ran/2306.07280.svg)](https://syntology.ai/paper/2306.07280)

Connect an agent — have() is free.