SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2406.05658 · NeurIPS · 2024

Visual Prompt Tuning in Null Space for Continual Learning

Yue Lu, Peng Wang, Nannan Wang, Shizhou Zhang, De Cheng, Yinghui Xing, Yanning Zhang

arXiv · PDF · Open in the Atlas

Code that ran

12 of 19 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (4 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 1 violated it and 4 have no contract verdict; 7 failures of our instrument; 7 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/2406.05658.json · badge: /api/ran/2406.05658.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
zugexiaodui/VPTinNSforCL canonical (classifier label)
classifier basis: census
8 ran · 3 no recorded run
zugexiaodui/vptinnsforcl canonical (classifier label)
classifier basis: classifier_2025
4 ran · 4 no recorded run
FunctionStatusWhere it lives
ClassIncrementalManager Ran
ran on a synthesized input · constructed an object; no result was computed
zugexiaodui/vptinnsforcl/train_eval.py
code served (permissive licence) · get_code("3268683b13ffdab4")
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, ClassIncrementalManager: the run constructed the object and did not call it, so no result was computed. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

ARGS = {
    'class_list': [0, 1, 2, 3, 4, 5, 6, 7],
    'num_tasks': 2,
    'rand_seed': 0,
    'shuffle': True,
    'shuffle_level': 'class'
}
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 ClassIncrementalManager — 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 ClassIncrementalManager 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
ImagePathDataset Ran
ran on a synthesized input · constructed an object; no result was computed
zugexiaodui/vptinnsforcl/train_eval.py
code served (permissive licence) · get_code("078149a1266f4078")
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, ImagePathDataset: the run constructed the object and did not call it, so no result was computed. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn
from torch.utils.data import Dataset

torch.manual_seed(0)

# ImagePathDataset.__init__ takes no parameters beyond self
# So ARGS should be an empty dict since there are no parameters to configure

ARGS = {}
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 ImagePathDataset — 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 ImagePathDataset 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
ImagePathDatasetClassManager Ran
ran on a synthesized input · constructed an object; no result was computed
zugexiaodui/vptinnsforcl/train_eval.py
code served (permissive licence) · get_code("96c050fd4ac91bfd")
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, ImagePathDatasetClassManager: the run constructed the object and did not call it, so no result was computed. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import os
import tempfile
from functools import partial

# Create temporary directory structure to simulate datasets
temp_dir = tempfile.mkdtemp()

# Create CIFAR100 structure
cifar100_dir = os.path.join(temp_dir, 'data.CIFAR100')
cifar100_train = os.path.join(cifar100_dir, 'train')
os.makedirs(cifar100_train, exist_ok=True)

# Create 100 class directories with dummy files
for i in range(100):
    class_dir = os.path.join(cifar100_train, str(i))
    os.makedirs(class_dir, exist_ok=True)
    # Create a dummy image file
    with open(os.path.join(class_dir, 'img_0.jpg'), 'w') as f:
        f.write('dummy')

cifar100_val = os.path.join(cifar100_dir, 'val')
os.makedirs(cifar100_val, exist_ok=True)
for i in range(100):
    class_dir = os.path.join(cifar100_val, str(i))
    os.makedirs(class_dir, exist_ok=True)
    with open(os.path.join(class_dir, 'img_0.jpg'), 'w') as f:
        f.write('dummy')

# Create ImageNet-R structure
imagenet_r_dir = os.path.join(temp_dir, 'data.ImageNet-R')
imagenet_r_train = os.path.join(imagenet_r_dir, 'train')
os.makedirs(imagenet_r_train, exist_ok=True)

for i in range(200):
    class_dir = os.path.join(imagenet_r_train, f'class_{i:03d}')
    os.makedirs(class_dir, exist_ok=True)
    with open(os.path.join(class_dir, 'img_0.jpg'), 'w') as f:
        f.write('dummy')

imagenet_r_val = os.path.join(imagenet_r_dir, 'val')
os.makedirs(imagenet_r_val, exist_ok=True)
for i in range(200):
    class_dir = os.path.join(imagenet_r_val, f'class_{i:03d}')
    os.makedirs(class_dir, exist_ok=True)
    with open(os.path.join(class_dir, 'img_0.jpg'), 'w') as f:
        f.write('dummy')

# Create DomainNet structure
domainnet_dir = os.path.join(temp_dir, 'data.DomainNet')
domainnet_train = os.path.join(domainnet_dir, 'train')
os.makedirs(domainnet_train, exist_ok=True)

for i in range(200):
    class_dir = os.path.join(domainnet_train, f'class_{i:03d}')
    os.makedirs(class_dir, exist_ok=True)
    with open(os.path.join(class_dir, 'img_0.jpg'), 'w') as f:
        f.write('dummy')

domainnet_val = os.path.join(domainnet_dir, 'val')
os.makedirs(domainnet_val, exist_ok=True)
for i in range(200):
    class_dir = os.path.join(domainnet_val, f'class_{i:03d}')
    os.makedirs(class_dir, exist_ok=True)
    with open(os.path.join(class_dir, 'img_0.jpg'), 'w') as f:
        f.write('dummy')

# Create tools directory with classname files
tools_dir = os.path.join(temp_dir, 'tools')
os.makedirs(tools_dir, exist_ok=True)

# Create CIFAR100 classnames file
with open(os.path.join(tools_dir, 'cifar100_classnames.txt'), 'w') as f:
    for i in range(100):
        f.write(f'class_{i}\n')

# Create ImageNet-R classnames file
with open(os.path.join(tools_dir, 'imagenet_r_classnames.txt'), 'w') as f:
    for i in range(200):
        f.write(f'class_{i:03d} classname_{i}\n')

# Mock the __file__ location for the classnames lookup
import sys
mock_module_dir = os.path.join(temp_dir, 'mock_module')
os.makedirs(mock_module_dir, exist_ok=True)

ARGS = {
    'cifar100': os.path.join(temp_dir, 'data.CIFAR100'),
    'imagenet_r': os.path.join(temp_dir, 'data.ImageNet-R'),
    'sdomainet': os.path.join(temp_dir, 'data.DomainNet'),
}
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 ImagePathDatasetClassManager — 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 ImagePathDatasetClassManager 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
SDomainNetPath Ran
ran on a synthesized input · constructed an object; no result was computed
zugexiaodui/vptinnsforcl/train_eval.py
code served (permissive licence) · get_code("8cfe390097de1132")
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, SDomainNetPath: the run constructed the object and did not call it, so no result was computed. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import os
import tempfile
import shutil

# Create a temporary directory structure that mimics DomainNet
temp_root = tempfile.mkdtemp()

# Create train and val directories
train_dir = os.path.join(temp_root, 'train')
val_dir = os.path.join(temp_root, 'val')
os.makedirs(train_dir)
os.makedirs(val_dir)

# Create 200 class directories with dummy image files
for split_path in [train_dir, val_dir]:
    for class_idx in range(200):
        class_dir = os.path.join(split_path, f'class_{class_idx:03d}')
        os.makedirs(class_dir)
        # Create a few dummy image files
        for img_idx in range(2):
            img_path = os.path.join(class_dir, f'image_{img_idx}.jpg')
            with open(img_path, 'w') as f:
                f.write('dummy')

ARGS = {
    'root_dir': temp_root,
    'train': 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 SDomainNetPath — 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 SDomainNetPath 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
basic_clean Ran
ran; our contract draft was wrong, not the code
zugexiaodui/VPTinNSforCL/clip/simple_tokenizer.py
pointer only (licence: MIT); the stored copy get_code serves, from ltpwy/MSCI, is pointer only (licence: NONE) · get_code("98f385d847636a3e")
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 ltpwy/MSCI, 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.

calc_accuracy Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
zugexiaodui/VPTinNSforCL/utils/misc.py
code served (permissive licence) · get_code("971c46b0aa73eb28")
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 torch

torch.manual_seed(0)

# Case 1: preds is long dtype with 2D shape (batch_size, num_classes)
# target is long dtype with 1D shape (batch_size,)
# topk is a tuple of integers

batch_size = 2
num_classes = 4

preds = torch.tensor([
    [0, 1, 2, 3],
    [3, 2, 1, 0]
], dtype=torch.long)

target = torch.tensor([1, 3], dtype=torch.long)

topk = (1, 2)

ARGS = {
    "preds": preds,
    "target": target,
    "topk": topk
}
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 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
format_duration Ran
ran; our contract draft was wrong, not the code
zugexiaodui/VPTinNSforCL/utils/misc.py
code served (permissive licence) · get_code("d06fe57a613a4c54")
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
import numpy as np

# Analyze the function:
# format_duration takes a single parameter delta_time of type float
# It converts delta_time to seconds and formats as HH:MM:SS
# The function only performs arithmetic operations on delta_time

ARGS = {
    'delta_time': 3661.7  # 1 hour, 1 minute, 1 second with decimal
}
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 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
get_init_weights_vit Ran
ran; our contract draft was wrong, not the code
zugexiaodui/VPTinNSforCL/utils/vit_builder.py
code served (permissive licence) · get_code("398c127f958b7a62")
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 partial. 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
from functools import partial

torch.manual_seed(0)

ARGS = {
    'mode': 'jax',
    'head_bias': 0.0,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
timm, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type partial — a partial output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a partial 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_pairs Ran
ran; our contract draft was wrong, not the code
zugexiaodui/VPTinNSforCL/clip/simple_tokenizer.py
code served (permissive licence) · get_code("d919ae32e5e4e616")
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 set. 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)

# The function get_pairs takes a word parameter
# A word is represented as a tuple of symbols (variable-length strings)
# The function iterates through the word and creates pairs of consecutive symbols

word = ("hello", "world", "test")

ARGS = {
    "word": word
}
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 set — a set output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a set 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
resize_pos_embed Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
zugexiaodui/VPTinNSforCL/utils/vit_builder.py
code served (permissive licence) · get_code("aba59faf86f45da9")
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 (1, 10, 8) 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 math

torch.manual_seed(0)

# Construct position embedding tensors
# Old posemb: shape (1, num_tokens, embed_dim)
# For a 4x4 grid + 1 prefix token: 17 tokens total, embed_dim=8
gs_old = 4
num_prefix_tokens = 1
num_tokens_old = gs_old * gs_old + num_prefix_tokens  # 17
embed_dim = 8
posemb = torch.randn(1, num_tokens_old, embed_dim)

# New posemb: shape (1, num_tokens_new, embed_dim)
# For a 3x3 grid + 1 prefix token: 10 tokens total
gs_new_val = 3
num_tokens_new = gs_new_val * gs_new_val + num_prefix_tokens  # 10
posemb_new = torch.randn(1, num_tokens_new, embed_dim)

ARGS = {
    'posemb': posemb,
    'posemb_new': posemb_new,
    'num_prefix_tokens': 1,
    'gs_new': (3, 3),
    'interpolation': 'bicubic',
    'antialias': 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 Tensor, ndim 3, dtype family float, shape (1, 10, 8) — shape (1, 10, 8) 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
str2bool Ran
ran, violated the contract we drafted
zugexiaodui/VPTinNSforCL/utils/misc.py
code served (permissive licence) · get_code("29848dd47d2443ef")
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 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 argparse

def str2bool(s):
    if isinstance(s, bool):
        return s
    true_list = ('y', 'true', '1')
    false_list = ('n', 'false', '0')
    s = str(s).lower()
    if s in true_list:
        return True
    elif s in false_list:
        return False
    else:
        raise argparse.ArgumentTypeError(f"'{s}' should be in {true_list} or {false_list}")

# Test with different input types
ARGS = {
    "s": "true"
}
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 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
whitespace_clean Ran
ran; our contract draft was wrong, not the code
zugexiaodui/VPTinNSforCL/clip/simple_tokenizer.py
pointer only (licence: MIT); the stored copy get_code serves, from ltpwy/MSCI, is pointer only (licence: NONE) · get_code("9542161e9640b858")
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 ltpwy/MSCI, 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.

CIFAR100Path Not yet run
unverified (harvested, no recorded run)
zugexiaodui/vptinnsforcl/train_eval.py
code served (permissive licence) · get_code("ff2fbb0ab73e96ea")
How it ran

no invocation recorded.

GlobalVarsManager Not yet run
unverified (harvested, no recorded run)
zugexiaodui/vptinnsforcl/train_eval.py
code served (permissive licence) · get_code("a8edf98dd5c2d59b")
How it ran

no invocation recorded.

ImageNetRPath Not yet run
unverified (harvested, no recorded run)
zugexiaodui/vptinnsforcl/train_eval.py
code served (permissive licence) · get_code("474413fb5eedd59b")
How it ran

no invocation recorded.

build_model Not yet run
unverified (harvested, no recorded run)
zugexiaodui/VPTinNSforCL/clip/model.py
code served (permissive licence) · get_code("6a720146b65b3fd6")
How it ran

no invocation recorded.

define_dataset Not yet run
unverified (harvested, no recorded run)
zugexiaodui/VPTinNSforCL/utils/dataset_builder.py
code served (permissive licence) · get_code("35a10325e8443f2e")
How it ran

no invocation recorded.

get_update_projection_dict Not yet run
unverified (harvested, no recorded run)
zugexiaodui/vptinnsforcl/train_eval.py
code served (permissive licence) · get_code("fc5d7b56e09282a5")
How it ran

no invocation recorded.

load Not yet run
unverified (harvested, no recorded run)
zugexiaodui/VPTinNSforCL/clip/clip.py
code served (permissive licence) · get_code("fb0f7e6d8dacdf56")
How it ran

no invocation recorded.

Also present in other papers' code

Of 19 samples lifted from this paper's linked repositories, 19 carry a function identity and 0 do not. Of the 19 distinct functions with an identity, 10 are also present in other papers whose code we hold (4 with at least one paper dated earlier, 5 the same month, 10 later) and 9 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_pairs 258 other papers
152 dated earlier, 11 the same month, 87 later, 7 same year, order not determinable, 1 undated · 4 only as an assembly brick
1802.06474 (2018-02)
A Closed-form Solution to Photorealistic Image Stylization
1803.01128 (2018-03)
Seq2Sick: Evaluating the Robustness of Sequence-to-Sequence Models with Adversarial Examples
1809.00125 (2018-09)
Simple Fusion: Return of the Language Model
and 255 more not named here
whitespace_clean 231 other papers
133 dated earlier, 9 the same month, 81 later, 7 same year, order not determinable, 1 undated · 4 only as an assembly brick
1901.07441 (2019-01)
PadChest: A large chest x-ray image dataset with multi-label annotated reports
2107.06383 (2021-07)
How Much Can CLIP Benefit Vision-and-Language Tasks?
2108.10904 (2021-08)
SimVLM: Simple Visual Language Model Pretraining with Weak Supervision
and 228 more not named here
basic_clean 218 other papers
133 dated earlier, 10 the same month, 67 later, 7 same year, order not determinable, 1 undated · 4 only as an assembly brick
1901.07441 (2019-01)
PadChest: A large chest x-ray image dataset with multi-label annotated reports
2107.06383 (2021-07)
How Much Can CLIP Benefit Vision-and-Language Tasks?
2108.10904 (2021-08)
SimVLM: Simple Visual Language Model Pretraining with Weak Supervision
and 215 more not named here
load 18 other papers
7 dated earlier, 3 the same month, 8 later
2107.06383 (2021-07)
How Much Can CLIP Benefit Vision-and-Language Tasks?
2311.11261 (2023-11)
Adversarial Prompt Tuning for Vision-Language Models
2311.18231 (2023-11)
TCP:Textual-based Class-aware Prompt tuning for Visual-Language Model
and 15 more not named here
get_init_weights_vit 2 other papers
1 the same month, 1 later
2406.18051 (2024-06)
ViT-1.58b: Mobile Vision Transformers in the 1-bit Era
2505.22389 (2025-05)
Train with Perturbation, Infer after Merging: A Two-Stage Framework for Continual Learning
calc_accuracy 1 other paper
1 later
2505.22389 (2025-05)
Train with Perturbation, Infer after Merging: A Two-Stage Framework for Continual Learning
define_dataset 1 other paper
1 later
2505.22389 (2025-05)
Train with Perturbation, Infer after Merging: A Two-Stage Framework for Continual Learning
format_duration 1 other paper
1 later
2505.22389 (2025-05)
Train with Perturbation, Infer after Merging: A Two-Stage Framework for Continual Learning
resize_pos_embed 1 other paper
1 later
2505.22389 (2025-05)
Train with Perturbation, Infer after Merging: A Two-Stage Framework for Continual Learning
str2bool 1 other paper
1 later
2505.22389 (2025-05)
Train with Perturbation, Infer after Merging: A Two-Stage Framework for Continual Learning

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

Existing prompt-tuning methods have demonstrated impressive performances in continual learning (CL), by selecting and updating relevant prompts in the visiontransformer models. On the contrary, this paper aims to learn each task by tuning the prompts in the direction orthogonal to the subspace spanned by previous tasks' features, so as to ensure no interference on tasks that have been learned to overcome catastrophic forgetting in CL. However, different from the orthogonal projection in the traditional CNN architecture, the prompt gradient orthogonal projection in the ViT architecture shows completely different and greater challenges, i.e., 1) the highorder and non-linear self-attention operation; 2) the drift of prompt distribution brought by the LayerNorm in the transformer block. Theoretically, we have finally deduced two consistency conditions to achieve the prompt gradient orthogonal projection, which provide a theoretical guarantee of eliminating interference on previously learned knowledge via the self-attention mechanism in visual prompt tuning. In practice, an effective null-space-based approximation solution has been proposed to implement the prompt gradient orthogonal projection. Extensive experimental results demonstrate the effectiveness of anti-forgetting on four classincremental benchmarks with diverse pre-trained baseline models, and our approach achieves superior performances to state-of-the-art methods. Our code is available at https://github.com/zugexiaodui/VPTinNSforCL .

For agents

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

get_harvested_code_for_paper("2406.05658")
get_code_for_paper("2406.05658")
have("2406.05658")

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

curl https://syntology.ai/api/ran/2406.05658.json

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

[![Syntology run record](https://syntology.ai/api/ran/2406.05658.svg)](https://syntology.ai/paper/2406.05658)

Connect an agent — have() is free.