SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1906.02629 · 2019

When Does Label Smoothing Help?

arXiv · PDF · Open in the Atlas

Code that ran

6 of 7 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; 0 honoured the contract we drafted, 0 violated it and 5 have no contract verdict; 1 failure of our instrument; 1 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/1906.02629.json · badge: /api/ran/1906.02629.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
ElenaRyumina/EMO-AffectNetModel community (classifier label)
classifier basis not recorded on the link
4 ran · 1 no recorded run
copy not recorded community
not in the paper's declared repositories
1 ran · 0 no recorded run
seominseok0429/label-smoothing-visualization-pytorch reimplementation (classifier label)
classifier basis: classifier_2025
1 ran · 0 no recorded run
FunctionStatusWhere it lives
df_group Ran
ran on a synthesized input
ElenaRyumina/EMO-AffectNetModel/functions/sequences.py
code served (permissive licence) · get_code("ef172d9e9018893b")
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 DataFrame. 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 pandas as pd
import numpy as np

# Create a small DataFrame with frame column and some numeric columns
np.random.seed(0)
data = {
    'frame': [0, 0, 1, 1, 2],
    'feature_a': [1.0, 2.0, 3.0, 4.0, 5.0],
    'feature_b': [10.0, 20.0, 30.0, 40.0, 50.0],
    'feature_c': [100.0, 200.0, 300.0, 400.0, 500.0]
}
df = pd.DataFrame(data)

# label_model is used as an iterable to create dictionary keys in the agg() call
# The function iterates over it: {i: 'mean' for i in label_model}
# So label_model should be an iterable (list, tuple, etc.) of column names
label_model = ['feature_a', 'feature_b', 'feature_c']

ARGS = {
    'df': df,
    'label_model': label_model
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, pandas — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type DataFrame — a DataFrame output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a DataFrame output the recorded ndim 2 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
format_time Ran
ran; our contract draft was wrong, not the code
this paper's copy was not recorded; identical code first harvested from JunlinHan/YOCO
pointer only; the stored copy get_code serves, from JunlinHan/YOCO, is pointer only (licence: NONE) · get_code("8d7ec010c29e813b")
How it ran

The body of this sample is withheld: this paper's copy of the code was not recorded (its link names no repository or source path), so no licence was read for it. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

get_mean_and_std Ran
ran on a synthesized input
seominseok0429/label-smoothing-visualization-pytorch/utils.py
pointer only (licence: NONE) · get_code("3cbdd3b190b16350")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, seominseok0429/label-smoothing-visualization-pytorch 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.

preprocess_input Ran
ran on a synthesized input
ElenaRyumina/EMO-AffectNetModel/functions/utils.py
code served (permissive licence) · get_code("eee939c2db10a418")
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 4, dtype family float, shape (2, 4, 5, 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 tensorflow as tf

# Create a small batch of image data
# Shape: (batch_size, height, width, channels) for channels_last format
# Using distinct sizes: batch=2, height=4, width=5, channels=3
x = np.random.randn(2, 4, 5, 3).astype(np.float32)

# Set data_format - use channels_last as default
data_format = 'channels_last'

# Set version to 1
version = 1

ARGS = {
    'x': x,
    'data_format': data_format,
    'version': version,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 4, dtype family float, shape (2, 4, 5, 3) — shape (2, 4, 5, 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
resnet_conv_block Ran
ran on a synthesized input
ElenaRyumina/EMO-AffectNetModel/functions/get_models.py
code served (permissive licence) · get_code("4f8026759621b1d5")
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 EagerTensor, ndim 4, dtype family float, shape (2, 4, 4, 256) 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 tensorflow as tf
import numpy as np

# Create a small input tensor
# Shape: (batch_size, height, width, channels)
input_tensor = tf.keras.Input(shape=(8, 8, 3))

# Create a simple model to get an actual tensor
x = tf.keras.layers.Conv2D(3, (1, 1))(input_tensor)
model = tf.keras.Model(inputs=input_tensor, outputs=x)

# Get actual tensor by passing dummy input
dummy_input = tf.random.normal((2, 8, 8, 3))
actual_input_tensor = model(dummy_input)

ARGS = {
    'input_tensor': actual_input_tensor,
    'kernel_size': (3, 3),
    'filters': (64, 64, 256),
    'stage': 2,
    'block': 1,
    'strides': (2, 2),
    'bias': False
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type EagerTensor, ndim 4, dtype family float, shape (2, 4, 4, 256) — shape (2, 4, 4, 256) 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
sequences Ran
ran on a synthesized input
ElenaRyumina/EMO-AffectNetModel/functions/sequences.py
code served (permissive licence) · get_code("816fcd30b3119901")
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. 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

torch.manual_seed(0)

# all_path is indexed and sliced, extended with list operations
# It should be a list that supports slicing and extend
all_path = ["path_0", "path_1", "path_2", "path_3", "path_4"]

# all_feature is indexed with slicing, has .numpy() method, and .tolist() is called on slices
# It should be a torch tensor
all_feature = torch.randn(5, 4)

win = 10
step = 5

ARGS = {
    "all_path": all_path,
    "all_feature": all_feature,
    "win": win,
    "step": step
}
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 tuple — a tuple output has no array shape of its own; none is 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
resnet_identity_block Not yet run
unverified (harvested, no recorded run)
ElenaRyumina/EMO-AffectNetModel/functions/get_models.py
code served (permissive licence) · get_code("a396c058d14ded87")
How it ran

no invocation recorded.

Also present in other papers' code

Of 7 samples lifted from this paper's linked repositories, 7 carry a function identity and 0 do not. Of the 7 distinct functions with an identity, 2 are also present in other papers whose code we hold (2 with at least one paper dated earlier, 0 the same month, 2 later) and 5 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
format_time 37 other papers
5 dated earlier, 32 later
1603.05027 (2016-03)
Identity Mappings in Deep Residual Networks
1605.07146 (2016-05)
Wide Residual Networks
1805.11327 (2018-05)
Lightweight Probabilistic Deep Networks
and 34 more not named here
get_mean_and_std 18 other papers
3 dated earlier, 15 later
1603.05027 (2016-03)
Identity Mappings in Deep Residual Networks
1605.07146 (2016-05)
Wide Residual Networks
1805.11327 (2018-05)
Lightweight Probabilistic Deep Networks
and 15 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

The generalization and learning speed of a multi-class neural network can often be significantly improved by using soft targets that are a weighted average of the hard targets and the uniform distribution over labels. Smoothing the labels in this way prevents the network from becoming over-confident and label smoothing has been used in many state-of-the-art models, including image classification, language translation and speech recognition. Despite its widespread use, label smoothing is still poorly understood. Here we show empirically that in addition to improving generalization, label smoothing improves model calibration which can significantly improve beam-search. However, we also observe that if a teacher network is trained with label smoothing, knowledge distillation into a student network is much less effective. To explain these observations, we visualize how label smoothing changes the representations learned by the penultimate layer of the network. We show that label smoothing encourages the representations of training examples from the same class to group in tight clusters. This results in loss of information in the logits about resemblances between instances of different classes, which is necessary for distillation, but does not hurt generalization or calibration of the model's predictions.

For agents

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

get_harvested_code_for_paper("1906.02629")
get_code_for_paper("1906.02629")
have("1906.02629")

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

curl https://syntology.ai/api/ran/1906.02629.json

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

[![Syntology run record](https://syntology.ai/api/ran/1906.02629.svg)](https://syntology.ai/paper/1906.02629)

Connect an agent — have() is free.