SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2206.14532 · ICML · 2022

Revisiting Label Smoothing and Knowledge Distillation Compatibility What was Missing?

Ngoc-Trung Tran, Ngai-Man Cheung, Keshigeyan Chandrasegaran, Yunqing Zhao

arXiv · PDF · Open in the Atlas

Code that ran

10 of 16 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 1 honoured the contract we drafted, 0 violated it and 8 have no contract verdict; 1 failure 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/2206.14532.json · badge: /api/ran/2206.14532.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
sutd-visual-computing-group/LS-KD-compatibility canonical (classifier label)
classifier basis not recorded on the link
10 ran · 6 no recorded run
FunctionStatusWhere it lives
accuracy Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
sutd-visual-computing-group/LS-KD-compatibility/src/image_classification/imagenet/official_evaluation.py
code served (permissive licence) · get_code("131a82fd65128218")
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 list, ndim 2, 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)

# Create small tensors for testing
batch_size = 2
num_classes = 4

# output: logits/scores tensor of shape (batch_size, num_classes)
output = torch.randn(batch_size, num_classes)

# target: class indices tensor of shape (batch_size,)
target = torch.tensor([1, 3])

# topk: tuple of k values to compute accuracy for
topk = (1, 2)

ARGS = {
    "output": output,
    "target": target,
    "topk": topk,
}
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 list, ndim 2, 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
adjust_learning_rate Ran
ran on a synthesized input
sutd-visual-computing-group/LS-KD-compatibility/src/image_classification/cub/train_student.py
code served (permissive licence) · get_code("4404239a85829756")
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 float, 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 as nn
from torch.optim import SGD

torch.manual_seed(0)

# Create a minimal model
class TinyModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear = nn.Linear(4, 3)
    
    def forward(self, x):
        return self.linear(x)

model = TinyModel()

# Create an optimizer
optimizer = SGD(model.parameters(), lr=0.1)

# Create args object with required attributes
class Args:
    def __init__(self):
        self.lr = 0.1

args = Args()

ARGS = {
    'optimizer': optimizer,
    'epoch': 5,
    'step': 10,
    'len_epoch': 100,
    'args': 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 float, 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
adjust_learning_rate Ran
ran on a synthesized input
sutd-visual-computing-group/LS-KD-compatibility/src/image_classification/cub/train_teacher.py
code served (permissive licence) · get_code("e93943edd5a8472f")
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 float, 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 as nn
import torch.optim as optim

torch.manual_seed(0)

# Create a simple model
class SimpleModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear = nn.Linear(4, 3)
    
    def forward(self, x):
        return self.linear(x)

model = SimpleModel()

# Create an optimizer
optimizer = optim.SGD(model.parameters(), lr=0.1)

# Create a simple args object with the required attributes
class Args:
    def __init__(self):
        self.lr = 0.1

args = Args()

# Define the ARGS dict with all parameters
ARGS = {
    'optimizer': optimizer,
    'epoch': 5,
    'step': 10,
    'len_epoch': 100,
    'args': 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 float, 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
adjust_learning_rate Ran
ran on a synthesized input
sutd-visual-computing-group/LS-KD-compatibility/src/image_classification/imagenet/official_evaluation.py
code served (permissive licence) · get_code("fa122dd6d0671258")
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 float, 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 as nn
from torch.optim import SGD

torch.manual_seed(0)

# Create a minimal model
model = nn.Linear(4, 3)

# Create an optimizer
optimizer = SGD(model.parameters(), lr=0.1)

# Create args with required attributes
class Args:
    def __init__(self):
        self.lr = 0.1

args = Args()

# Define ARGS dict mapping all parameters
ARGS = {
    'optimizer': optimizer,
    'epoch': 5,
    'args': 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 float, ndim 0, dtype family float, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
get_imagenet_trained_teacher_for_imagenet_student Ran
ran on a synthesized input
sutd-visual-computing-group/LS-KD-compatibility/src/image_classification/imagenet/utils.py
code served (permissive licence) · get_code("fbaa9853176fc776")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torchvision.models as models
import tempfile
import os

torch.manual_seed(0)

# Create a minimal checkpoint file with the required structure
temp_dir = tempfile.mkdtemp()
weights_path = os.path.join(temp_dir, 'teacher_checkpoint.pth')

# Create a minimal ResNet18 state dict
teacher_model = models.resnet18(pretrained=False)
state_dict = {}
for k, v in teacher_model.state_dict().items():
    state_dict[f'module.{k}'] = v

checkpoint = {
    'arch': 'resnet18',
    'state_dict': state_dict,
    'best_acc1': 76.130,
    'best_acc5': 92.862
}

torch.save(checkpoint, weights_path)

# Create args object with required attributes
class Args:
    def __init__(self, teacher_weights_path, teacher_alpha):
        self.teacher_weights_path = teacher_weights_path
        self.teacher_alpha = teacher_alpha

ARGS = {
    'args': Args(
        teacher_weights_path=weights_path,
        teacher_alpha=0.95
    )
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch, torchvision — 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
to_python_float Ran
ran, honoured the contract we drafted
sutd-visual-computing-group/LS-KD-compatibility/src/image_classification/imagenet/train_student_dali.py
pointer only (licence: MIT); the stored copy get_code serves, from lengsicong/tell2design, is pointer only (licence: NONE) · get_code("6b5d363ab4dbf012")
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 lengsicong/tell2design, 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.

train Ran
ran on a synthesized input
sutd-visual-computing-group/LS-KD-compatibility/src/image_classification/cub/train_student.py
code served (permissive licence) · get_code("2a0d7e917240db2a")
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. 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
import torch.optim as optim

torch.manual_seed(0)

# Build a minimal model
class SimpleModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.fc = nn.Linear(4, 10)
    
    def forward(self, x):
        return self.fc(x)

# Build a minimal teacher model
class TeacherModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.fc = nn.Linear(4, 10)
    
    def forward(self, x):
        return self.fc(x)

# Create minimal train loader (list of batches)
train_loader = [
    (torch.randn(2, 4), torch.randint(0, 10, (2,))),
    (torch.randn(2, 4), torch.randint(0, 10, (2,))),
]

# Create model instances
model = SimpleModel()
teacher = TeacherModel()

# Create optimizer
optimizer = optim.SGD(model.parameters(), lr=0.1)

# Create criterion (KL divergence or MSE for teacher-student)
criterion = nn.KLDivLoss()

# Create scaler for mixed precision
scaler = torch.cuda.amp.GradScaler()

# Create args config
class Args:
    def __init__(self):
        self.gpu = None
        self.use_amp = False
        self.print_freq = 1
        self.lr = 0.1

args = Args()

ARGS = {
    'train_loader': train_loader,
    'model': model,
    'teacher': teacher,
    'criterion': criterion,
    'optimizer': optimizer,
    'scaler': scaler,
    'epoch': 0,
    'args': 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 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
validate Ran
ran on a synthesized input
sutd-visual-computing-group/LS-KD-compatibility/src/image_classification/cub/train_student.py
code served (permissive licence) · get_code("020b5acc2cea526a")
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. 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 a simple model for validation
class SimpleModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.fc = nn.Linear(4, 10)  # 10 classes for top-5 accuracy
    
    def forward(self, x):
        return self.fc(x)

# Create a simple validation loader (list of tuples)
val_loader = [
    (torch.randn(2, 4), torch.randint(0, 10, (2,))),
    (torch.randn(2, 4), torch.randint(0, 10, (2,))),
    (torch.randn(2, 4), torch.randint(0, 10, (2,))),
]

# Create model and criterion
model = SimpleModel()
criterion = nn.CrossEntropyLoss()

# Create args object
class Args:
    def __init__(self):
        self.gpu = None
        self.use_amp = False
        self.print_freq = 1

args = Args()

ARGS = {
    'val_loader': val_loader,
    'model': model,
    'criterion': criterion,
    'args': 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 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
validate Ran
ran on a synthesized input
sutd-visual-computing-group/LS-KD-compatibility/src/image_classification/imagenet/official_evaluation.py
code served (permissive licence) · get_code("889998947e765d45")
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. 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 a simple model that takes images and outputs logits
class SimpleModel(nn.Module):
    def __init__(self):
        super(SimpleModel, self).__init__()
        self.fc = nn.Linear(4, 10)  # 10 classes for top-5 accuracy
    
    def forward(self, x):
        return self.fc(x)

# Create a small validation dataset
val_loader = [
    (torch.randn(2, 4), torch.randint(0, 10, (2,))),
    (torch.randn(2, 4), torch.randint(0, 10, (2,))),
]

# Create model and criterion
model = SimpleModel()
criterion = nn.CrossEntropyLoss()

# Create args object with required attributes
class Args:
    def __init__(self):
        self.gpu = None
        self.print_freq = 1

args = Args()

ARGS = {
    'val_loader': val_loader,
    'model': model,
    'criterion': criterion,
    'args': 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 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
validate Ran
ran on a synthesized input
sutd-visual-computing-group/LS-KD-compatibility/src/image_classification/imagenet/.ipynb_checkpoints/official_evaluation-checkpoint.py
code served (permissive licence) · get_code("00c8fdc297e7e272")
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. 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
from unittest.mock import MagicMock

torch.manual_seed(0)

# Build a minimal model
class MinimalModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.fc = nn.Linear(4, 10)  # 10 classes for top-5 accuracy
    
    def forward(self, x):
        return self.fc(x)

# Build a minimal criterion
criterion = nn.CrossEntropyLoss()

# Build a minimal val_loader (list of (images, target) tuples)
# Batch size 2, image feature dim 4, 10 classes
val_loader = [
    (torch.randn(2, 4), torch.randint(0, 10, (2,))),
    (torch.randn(2, 4), torch.randint(0, 10, (2,))),
]

# Build model
model = MinimalModel()

# Build args object with necessary attributes
class Args:
    def __init__(self):
        self.gpu = None  # CPU only
        self.print_freq = 1  # Print every batch

args = Args()

ARGS = {
    'val_loader': val_loader,
    'model': model,
    'criterion': criterion,
    'args': 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 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_cub_trained_teacher_for_cub_student Not yet run
unverified (harvested, no recorded run)
sutd-visual-computing-group/LS-KD-compatibility/src/image_classification/cub/utils.py
code served (permissive licence) · get_code("205d6e975fc4fca7")
How it ran

no invocation recorded.

get_imagenet_trained_teacher_for_cub_teacher Not yet run
unverified (harvested, no recorded run)
sutd-visual-computing-group/LS-KD-compatibility/src/image_classification/cub/utils.py
code served (permissive licence) · get_code("35f1a5ee27e2d2a0")
How it ran

no invocation recorded.

get_penultimate_fts Not yet run
unverified (harvested, no recorded run)
sutd-visual-computing-group/LS-KD-compatibility/src/visualization/alpha-LS-KD_imagenet_centroids.py
code served (permissive licence) · get_code("06ec1746d47a9901")
How it ran

no invocation recorded.

get_test_loader Not yet run
unverified (harvested, no recorded run)
sutd-visual-computing-group/LS-KD-compatibility/src/visualization/data_loader_imagenet.py
code served (permissive licence) · get_code("afe574de7fe140de")
How it ran

no invocation recorded.

load_model_kd Not yet run
unverified (harvested, no recorded run)
sutd-visual-computing-group/LS-KD-compatibility/src/visualization/alpha-LS-KD_imagenet_centroids.py
code served (permissive licence) · get_code("33ec9768e54f5c4a")
How it ran

no invocation recorded.

load_model_ls Not yet run
unverified (harvested, no recorded run)
sutd-visual-computing-group/LS-KD-compatibility/src/visualization/alpha-LS-KD_imagenet_centroids.py
code served (permissive licence) · get_code("1517519d4d7d27cf")
How it ran

no invocation recorded.

Also present in other papers' code

Of 16 samples lifted from this paper's linked repositories, 16 carry a function identity and 0 do not. Of the 16 distinct functions with an identity, 2 are also present in other papers whose code we hold (2 with at least one paper dated earlier, 1 the same month, 2 later) and 14 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
accuracy 116 other papers
44 dated earlier, 7 the same month, 61 later, 4 same year, order not determinable · 8 only as an assembly brick
1502.03044 (2015-02)
Show, Attend and Tell: Neural Image Caption Generation with Visual Attention
1610.05820 (2016-10)
Membership Inference Attacks Against Machine Learning Models
1905.03696 (2019-05)
HAWQ: Hessian AWare Quantization of Neural Networks with Mixed-Precision
and 113 more not named here
to_python_float 9 other papers
4 dated earlier, 4 later, 1 same year, order not determinable
1904.00962 (2019-04)
Large Batch Optimization for Deep Learning: Training BERT in 76 minutes
1909.03329 (2019-09)
LAMOL: LAnguage MOdeling for Lifelong Language Learning
2109.04332 (2021-09)
PPT: Pre-trained Prompt Tuning for Few-shot Learning
and 6 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

This work investigates the compatibility between label smoothing (LS) and knowledge distillation (KD). Contemporary findings addressing this thesis statement take dichotomous standpoints: Müller et al. (2019); Shen et al. (2021b). Critically, there is no effort to understand and resolve these contradictory findings, leaving the primal questionto smooth or not to smooth a teacher network?unanswered. The main contributions of our work are the discovery, analysis and validation of systematic diffusion as the missing concept which is instrumental in understanding and resolving these contradictory findings. This systematic diffusion essentially curtails the benefits of distilling from an LS-trained teacher, thereby rendering KD at increased temperatures ineffective. Our discovery is comprehensively supported by large-scale experiments, analyses and case studies including image classification, neural machine translation and compact student distillation tasks spanning across multiple datasets and teacherstudent architectures. Based on our analysis, we suggest practitioners to use an LS-trained teacher with a low-temperature transfer to achieve high performance students. Code and models are available at https://keshik6.github.io/ revisiting-ls-kd-compatibility/

For agents

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

get_harvested_code_for_paper("2206.14532")
get_code_for_paper("2206.14532")
have("2206.14532")

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

curl https://syntology.ai/api/ran/2206.14532.json

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

[![Syntology run record](https://syntology.ai/api/ran/2206.14532.svg)](https://syntology.ai/paper/2206.14532)

Connect an agent — have() is free.