SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1606.09282 · 2016

Learning without Forgetting

Derek Hoiem, Zhizhong Li

arXiv · PDF · Open in the Atlas

Code that ran

9 of 15 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (3 of them constructed an object rather than computing a result; 1 honoured the contract we drafted, 0 violated it and 3 have no contract verdict; 5 failures of our instrument; 6 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/1606.09282.json · badge: /api/ran/1606.09282.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
mmasana/FACIL community
Papers with Code archive
2 ran · 0 no recorded run
geox-lab/cmn community
Papers with Code archive
1 ran · 3 no recorded run
ngailapdi/LWF community
Papers with Code archive
1 ran · 2 no recorded run
GMvandeVen/continual-learning community
Papers with Code archive
1 ran · 1 no recorded run
ivclab/cvs community
Papers with Code archive
1 ran · 0 no recorded run
ContinualAI/avalanche community
Papers with Code archive
1 ran · 0 no recorded run
yaoyao-liu/mnemonics community
Papers with Code archive
1 ran · 0 no recorded run
wannabeOG/ExpertNet-Pytorch community
Papers with Code archive
1 ran · 0 no recorded run
FunctionStatusWhere it lives
Appr Ran
ran on a synthesized input · constructed an object; no result was computed
mmasana/FACIL/src/approach/lwf.py
code served (permissive licence) · get_code("a65368ea6628ffd0")
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, Appr: 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 DataLoader, Dataset
import numpy as np
from datetime import datetime
import tempfile
from argparse import Namespace

torch.manual_seed(0)

# Minimal model for testing
class SimpleModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.fc = nn.Linear(4, 10)
    
    def forward(self, x):
        return self.fc(x)

# Minimal exemplars dataset
class MinimalExemplarDataset:
    def __init__(self):
        self.images = []
        self.labels = []
    
    def _is_active(self):
        return False
    
    def collect_exemplars(self, model, trn_loader, selection_transform):
        pass

# Minimal logger
class MinimalLogger:
    def __init__(self):
        self.log_path = tempfile.gettempdir()
        self.exp_name = "test_exp"
        self.begin_time = datetime.now()
    
    def log_scalar(self, task, iter, name, value, group=None, curtime=None):
        pass
    
    def log_args(self, args):
        pass
    
    def log_result(self, array, name, step):
        pass
    
    def log_figure(self, name, iter, figure, curtime=None):
        pass
    
    def save_model(self, state_dict, task):
        pass

# Build the arguments dictionary
ARGS = {
    'model': SimpleModel(),
    'device': 'cpu',
    'nepochs': 10,
    'lr': 0.1,
    'lr_min': 0.0001,
    'lr_factor': 3,
    'lr_patience': 5,
    'clipgrad': 10000,
    'momentum': 0.9,
    'wd': 0.0001,
    'multi_softmax': False,
    'wu_nepochs': 0,
    'wu_lr_factor': 1,
    'fix_bn': False,
    'eval_on_train': False,
    'logger': MinimalLogger(),
    'exemplars_dataset': MinimalExemplarDataset()
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
PIL, numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Appr — 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 Appr 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
Inc_Learning_Appr Ran
ran on a synthesized input · constructed an object; no result was computed
mmasana/FACIL/src/approach/lwf.py
code served (permissive licence) · get_code("686c76397b75be4d")
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, Inc_Learning_Appr: 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 DataLoader, Dataset
import numpy as np
from datetime import datetime
import tempfile

torch.manual_seed(0)

# Minimal model
class TinyModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.fc1 = nn.Linear(4, 8)
        self.fc2 = nn.Linear(8, 10)
    
    def forward(self, x):
        x = torch.relu(self.fc1(x))
        x = self.fc2(x)
        return x

# Minimal dataset
class TinyDataset(Dataset):
    def __init__(self, size=8):
        self.size = size
    
    def __len__(self):
        return self.size
    
    def __getitem__(self, idx):
        x = torch.randn(4)
        y = torch.randint(0, 10, (1,)).item()
        return (x, y)

# Minimal logger
class TinyLogger:
    def __init__(self, log_path, exp_name, begin_time=None):
        self.log_path = log_path
        self.exp_name = exp_name
        self.exp_path = log_path
        self.begin_time = begin_time if begin_time else datetime.now()
    
    def log_scalar(self, task, iter, name, value, group=None, curtime=None):
        pass
    
    def log_args(self, args):
        pass
    
    def log_result(self, array, name, step):
        pass
    
    def log_figure(self, name, iter, figure, curtime=None):
        pass
    
    def save_model(self, state_dict, task):
        pass

# Minimal exemplars dataset
class TinyExemplarsDataset:
    def __init__(self):
        self.images = []
        self.labels = []
        self.max_num_exemplars_per_class = 0
        self.max_num_exemplars = 0
    
    def collect_exemplars(self, model, trn_loader, selection_transform):
        pass
    
    def _is_active(self):
        return False

# Create instances
model = TinyModel()
device = torch.device('cpu')
logger = TinyLogger(tempfile.gettempdir(), 'test_exp')
exemplars_dataset = TinyExemplarsDataset()

ARGS = {
    'model': model,
    'device': device,
    'nepochs': 100,
    'lr': 0.05,
    'lr_min': 0.0001,
    'lr_factor': 3,
    'lr_patience': 5,
    'clipgrad': 10000,
    'momentum': 0,
    'wd': 0,
    'multi_softmax': False,
    'wu_nepochs': 0,
    'wu_lr_factor': 1,
    'fix_bn': False,
    'eval_on_train': False,
    'logger': logger,
    'exemplars_dataset': exemplars_dataset
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
PIL, numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Inc_Learning_Appr — 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 Inc_Learning_Appr 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
KDLoss Ran
ran, honoured the contract we drafted
ivclab/cvs/kdloss.py
pointer only (licence: NONE) · get_code("28946d6ab7de665b")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, ivclab/cvs 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.

LearningWithoutForgetting Ran
ran on a synthesized input · constructed an object; no result was computed
ContinualAI/avalanche/avalanche/training/regularization.py
code served (permissive licence) · get_code("68bc5e317dccf467")
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, LearningWithoutForgetting: 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 collections import defaultdict
import copy

torch.manual_seed(0)

# Build a minimal model that supports the expected interface
class SimpleModel(nn.Module):
    def __init__(self, input_dim=4, output_dim=5):
        super().__init__()
        self.linear = nn.Linear(input_dim, output_dim)
    
    def forward(self, x):
        return self.linear(x)

# Build a MultiTaskModule stand-in
class MultiTaskModule(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear = nn.Linear(4, 5)
    
    def forward(self, x):
        return self.linear(x)

# Helper function stub (referenced in the code)
def avalanche_forward(model, x, task_id):
    if isinstance(model, MultiTaskModule):
        return {0: model(x)}
    else:
        return {0: model(x)}

# Build minimal agent and experience objects
class MinimalExperience:
    def __init__(self):
        self.dataset = None

class MinimalAgent:
    def __init__(self):
        self.model = SimpleModel(input_dim=4, output_dim=5)

# Create test tensors
mb_x = torch.randn(2, 4)  # batch_size=2, input_dim=4
mb_pred = torch.randn(2, 5)  # batch_size=2, num_classes=5
model = SimpleModel(input_dim=4, output_dim=5)

# Create the LearningWithoutForgetting instance with various parameter combinations
ARGS = {
    'alpha': 0.5,
    'temperature': 2
}
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 LearningWithoutForgetting — 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 LearningWithoutForgetting 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
MultiClassCrossEntropy Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
ngailapdi/LWF/model.py
code served (permissive licence) · get_code("b43737d7d32870ed")
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 float; no output shape is stated. 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 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
from torch.autograd import Variable

torch.manual_seed(0)

# Create small tensors for logits and labels
# logits: batch_size=2, num_classes=4
logits = torch.randn(2, 4, requires_grad=True)

# labels: batch_size=2, num_classes=4 (soft targets)
labels = torch.randn(2, 4)
labels = torch.softmax(labels, dim=1)  # Make them valid probability distributions

# Temperature parameter
T = 1.0

ARGS = {
    "logits": logits,
    "labels": labels,
    "T": T
}
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 float — no output shape is stated: observe_signature.py could not observe it (call:AssertionError)
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_one_hot Ran
ran; our contract draft was wrong, not the code
geox-lab/cmn/LwF-CIFAR100/LwF.py
code served (permissive licence) · get_code("dc209a0ea5b09a3f")
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 2, dtype family float, shape (2, 4) as a later run recorded it. 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, 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)

device = torch.device('cpu')

# Create a small target tensor with class indices
# Shape: (batch_size=2, ) with values in range [0, num_class)
target = torch.tensor([0, 2])

# Number of classes
num_class = 4

ARGS = {
    'target': target,
    'num_class': num_class,
}
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 2, dtype family float, shape (2, 4) — shape (2, 4) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
incremental_train_and_eval Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
yaoyao-liu/mnemonics/mnemonics-training/1_train/trainer/incremental.py
code served (permissive licence) · get_code("a99cc76d5a5b058b")
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 SimpleModel. 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
import torch.nn.functional as F

torch.manual_seed(0)

# Define a minimal target model
class SimpleModel(nn.Module):
    def __init__(self, num_classes=10):
        super().__init__()
        self.conv1 = nn.Conv2d(3, 8, kernel_size=3, padding=1)
        self.bn1 = nn.BatchNorm2d(8)
        self.fc = nn.Linear(8 * 32 * 32, num_classes)
    
    def forward(self, x):
        x = self.bn1(self.conv1(x))
        x = x.view(x.size(0), -1)
        x = self.fc(x)
        return x

# Define a reference model (for iteration > start_iteration case)
class RefModel(nn.Module):
    def __init__(self, num_classes=5):
        super().__init__()
        self.conv1 = nn.Conv2d(3, 8, kernel_size=3, padding=1)
        self.bn1 = nn.BatchNorm2d(8)
        self.fc = nn.Linear(8 * 32 * 32, num_classes)
    
    def forward(self, x):
        x = self.bn1(self.conv1(x))
        x = x.view(x.size(0), -1)
        x = self.fc(x)
        return x

# Create models
tg_model = SimpleModel(num_classes=10)
ref_model = RefModel(num_classes=5)
free_model = SimpleModel(num_classes=10)
ref_free_model = RefModel(num_classes=5)

# Create optimizer and scheduler
tg_optimizer = torch.optim.SGD(tg_model.parameters(), lr=0.1)
tg_lr_scheduler = torch.optim.lr_scheduler.StepLR(tg_optimizer, step_size=1, gamma=0.1)

# Create dummy data loaders
def create_dataloader(batch_size=2, num_batches=2, num_classes=10, img_size=32):
    data = []
    for _ in range(num_batches):
        inputs = torch.randn(batch_size, 3, img_size, img_size)
        targets = torch.randint(0, num_classes, (batch_size,))
        data.append((inputs, targets))
    return data

trainloader = create_dataloader(batch_size=2, num_batches=2, num_classes=10)
testloader = create_dataloader(batch_size=2, num_batches=1, num_classes=10)

# Create weight_per_class
weight_per_class = torch.ones(10)

ARGS = {
    'epochs': 1,
    'tg_model': tg_model,
    'ref_model': ref_model,
    'free_model': free_model,
    'ref_free_model': ref_free_model,
    'tg_optimizer': tg_optimizer,
    'tg_lr_scheduler': tg_lr_scheduler,
    'trainloader': trainloader,
    'testloader': testloader,
    'iteration': 0,
    'start_iteration': 0,
    'lamda': 1.0,
    'dist': 'cosine',
    'K': 5,
    'lw_mr': 0.5,
    'fix_bn': False,
    'weight_per_class': weight_per_class,
    'device': torch.device('cpu')
}
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 SimpleModel — a SimpleModel output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a SimpleModel 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
loss_fn_kd Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
GMvandeVen/continual-learning/models/utils/loss_functions.py
code served (permissive licence) · get_code("b76db52297524e52")
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 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
import torch.nn.functional as F

torch.manual_seed(0)

# Create small tensors for scores and target_scores
# scores: shape (batch_size=2, num_classes=4)
# target_scores: shape (batch_size=2, num_classes=3)
scores = torch.randn(2, 4)
target_scores = torch.randn(2, 3)

# Create weights for the batch (one weight per sample)
weights = [0.6, 0.4]

# Temperature parameter
T = 2.0

# Normalization flag
norm = True

ARGS = {
    "scores": scores,
    "target_scores": target_scores,
    "T": T,
    "weights": weights,
    "norm": norm,
}
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 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
model_criterion Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
wannabeOG/ExpertNet-Pytorch/utils/model_utils.py
code served (permissive licence) · get_code("25e295f6e060e434")
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 float; no output shape is stated. 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 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)

# Create small tensors for predictions and labels
batch_size = 2
num_classes = 4

# For 'CE' flag: preds should be logits, labels should be class indices
preds_ce = torch.randn(batch_size, num_classes)
labels_ce = torch.randint(0, num_classes, (batch_size,))

# For 'Distill' flag: both preds and labels should be score tensors (logits or similar)
preds_distill = torch.randn(batch_size, num_classes)
labels_distill = torch.randn(batch_size, num_classes)

# Test with CE flag
ARGS = {
    'preds': preds_ce,
    'labels': labels_ce,
    'flag': 'CE',
    'T': 2
}
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 float — no output shape is stated: observe_signature.py could not observe it (call:AssertionError)
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
LwFmodel Not yet run
unverified (harvested, no recorded run)
geox-lab/cmn/LwF-CIFAR100/LwF.py
code served (permissive licence) · get_code("f08f055b7c6c13c0")
How it ran

no invocation recorded.

Model Not yet run
unverified (harvested, no recorded run)
ngailapdi/LWF/model.py
code served (permissive licence) · get_code("8bb63a717b07a3b8")
How it ran

no invocation recorded.

iCIFAR100 Not yet run
unverified (harvested, no recorded run)
geox-lab/cmn/LwF-CIFAR100/LwF.py
code served (permissive licence) · get_code("97e769f51fb2d580")
How it ran

no invocation recorded.

kaiming_normal_init Not yet run
unverified (harvested, no recorded run)
ngailapdi/LWF/model.py
code served (permissive licence) · get_code("a8174ba24a4f43da")
How it ran

no invocation recorded.

network Not yet run
unverified (harvested, no recorded run)
geox-lab/cmn/LwF-CIFAR100/LwF.py
code served (permissive licence) · get_code("da1b0e40683dca7a")
How it ran

no invocation recorded.

weighted_average Not yet run
unverified (harvested, no recorded run)
GMvandeVen/continual-learning/models/utils/loss_functions.py
code served (permissive licence) · get_code("1179cbb68303ab65")
How it ran

no invocation recorded.

Also present in other papers' code

Of 15 samples lifted from this paper's linked repositories, 15 carry a function identity and 0 do not. Of the 15 distinct functions with an identity, 4 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 4 later) and 11 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
Inc_Learning_Appr
brick of this paper's assembly: helper
2 other papers
2 later
1611.07725 (2016-11)
iCaRL: Incremental Classifier and Representation Learning
1612.00796 (2016-12)
Overcoming catastrophic forgetting in neural networks
get_one_hot
brick of this paper's assembly: helper
1 other paper
1 later
1611.07725 (2016-11)
iCaRL: Incremental Classifier and Representation Learning
incremental_train_and_eval
brick of this paper's assembly: entry
1 other paper
1 later
2010.05063 (2020-10)
Adaptive Aggregation Networks for Class-Incremental Learning
network
brick of this paper's assembly: model
1 other paper
1 later · 1 only as an assembly brick
1611.07725 (2016-11)
iCaRL: Incremental Classifier and Representation 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

When building a unified vision system or gradually adding new capabilities to a system, the usual assumption is that training data for all tasks is always available. However, as the number of tasks grows, storing and retraining on such data becomes infeasible. A new problem arises where we add new capabilities to a Convolutional Neural Network (CNN), but the training data for its existing capabilities are unavailable. We propose our Learning without Forgetting method, which uses only new task data to train the network while preserving the original capabilities. Our method performs favorably compared to commonly used feature extraction and fine-tuning adaption techniques and performs similarly to multitask learning that uses original task data we assume unavailable. A more surprising observation is that Learning without Forgetting may be able to replace fine-tuning with similar old and new task datasets for improved new task performance.

For agents

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

get_harvested_code_for_paper("1606.09282")
get_code_for_paper("1606.09282")
have("1606.09282")

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

curl https://syntology.ai/api/ran/1606.09282.json

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

[![Syntology run record](https://syntology.ai/api/ran/1606.09282.svg)](https://syntology.ai/paper/1606.09282)

Connect an agent — have() is free.