SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2004.00288 · 2020

CurricularFace: Adaptive Curriculum Learning Loss for Deep Face Recognition

arXiv · PDF · Open in the Atlas

Code that ran

13 of 13 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; 2 honoured the contract we drafted, 0 violated it and 5 have no contract verdict; 6 failures of our instrument; 0 no recorded run), measured 2026-11-21; 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/2004.00288.json · badge: /api/ran/2004.00288.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
HuangYG123/CurricularFace canonical (classifier label)
classifier basis: census
13 ran · 0 no recorded run
FunctionStatusWhere it lives
ROC Ran
ran on a synthesized input
HuangYG123/CurricularFace/util/metrics.py
code served (permissive licence) · get_code("c386cdbee6b83849")
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 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 numpy as np

# Set random seed for reproducibility
np.random.seed(0)

# Create small test data
n_samples = 10
score_vec = np.random.rand(n_samples).astype(np.float32)
label_vec = np.array([True, False, True, False, True, True, False, False, True, False], dtype=np.bool_)

# Create thresholds (optional parameter)
thresholds = np.array([0.2, 0.5, 0.8], dtype=np.float32)

# Create FARs (optional parameter)
FARs = np.array([0.01, 0.1, 0.5], dtype=np.float32)

# Define ARGS dictionary with all parameters for ROC function
ARGS = {
    'score_vec': score_vec,
    'label_vec': label_vec,
    'thresholds': thresholds,
    'FARs': FARs,
    'get_false_indices': False
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 2, 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
ROC_by_mat Ran
ran on a synthesized input
HuangYG123/CurricularFace/util/metrics.py
code served (permissive licence) · get_code("71fd1b339eccf3b5")
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 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 numpy as np

# Set random seed for reproducibility
np.random.seed(0)

# Create a small pairwise score matrix (4x3)
score_mat = np.random.rand(4, 3).astype(np.float32)

# Create corresponding boolean label matrix (4x3)
label_mat = np.random.rand(4, 3) > 0.5

# Create thresholds
thresholds = np.array([0.2, 0.5, 0.8], dtype=np.float32)

# Create FARs
FARs = np.array([0.01, 0.1, 0.5], dtype=np.float32)

ARGS = {
    'score_mat': score_mat,
    'label_mat': label_mat,
    'thresholds': thresholds,
    'FARs': FARs,
    'get_false_indices': False,
    'triu_k': None
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 2, 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
ResNet_50 Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
HuangYG123/CurricularFace/backbone/model_resnet.py
code served (permissive licence) · get_code("de52f82c3df44b76")
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 ResNet. 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

torch.manual_seed(0)

ARGS = {
    "input_size": [112, 112],
}
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 ResNet — a ResNet output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a ResNet output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
calculate_accuracy Ran
ran, honoured the contract we drafted
HuangYG123/CurricularFace/util/verification.py
pointer only (licence: MIT); the stored copy get_code serves, from zrui94/insight_mx, is pointer only (licence: NONE) · get_code("f975336b5d5f1a80")
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 zrui94/insight_mx, 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.

calculate_roc Ran
ran on a synthesized input
HuangYG123/CurricularFace/util/verification.py
code served (permissive licence) · get_code("5f5160da896b0198")
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 numpy as np

np.random.seed(0)

# Create small test data
# embeddings1 and embeddings2: shape (n_pairs, embedding_dim)
n_pairs = 10
embedding_dim = 4

embeddings1 = np.random.randn(n_pairs, embedding_dim).astype(np.float32)
embeddings2 = np.random.randn(n_pairs, embedding_dim).astype(np.float32)

# actual_issame: boolean array of length n_pairs
actual_issame = np.array([True, False, True, False, True, False, True, False, True, False], dtype=bool)

# thresholds: array of threshold values to test
thresholds = np.array([0.1, 0.5, 1.0, 2.0, 5.0], dtype=np.float32)

# nrof_folds: number of k-fold splits
nrof_folds = 3

# pca: number of PCA components (0 means no PCA)
pca = 0

ARGS = {
    'thresholds': thresholds,
    'embeddings1': embeddings1,
    'embeddings2': embeddings2,
    'actual_issame': actual_issame,
    'nrof_folds': nrof_folds,
    'pca': pca
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, sklearn — 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
calculate_val Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
HuangYG123/CurricularFace/util/verification.py
pointer only (licence: MIT); the stored copy get_code serves, from dmonterom/face_recognition_TF2, is pointer only (licence: NONE) · get_code("f67f5f081381cc8a")
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 dmonterom/face_recognition_TF2, 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.

conv1x1 Ran
ran; our contract draft was wrong, not the code
HuangYG123/CurricularFace/backbone/model_resnet.py
code served (permissive licence) · get_code("bfd3aee914279f48")
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 Conv2d. 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.nn import Linear, Conv2d, BatchNorm1d, BatchNorm2d, ReLU, Dropout, MaxPool2d, Sequential, Module

torch.manual_seed(0)

def conv1x1(in_planes, out_planes, stride=1):
    """1x1 convolution"""
    return Conv2d(in_planes, out_planes, kernel_size=1, stride=stride, bias=False)

ARGS = {
    'in_planes': 4,
    'out_planes': 3,
    'stride': 1
}
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 Conv2d — a Conv2d output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Conv2d 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
conv3x3 Ran
ran; our contract draft was wrong, not the code
HuangYG123/CurricularFace/backbone/model_resnet.py
code served (permissive licence) · get_code("999e0ea90670a179")
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 Conv2d. 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.nn import Linear, Conv2d, BatchNorm1d, BatchNorm2d, ReLU, Dropout, MaxPool2d, Sequential, Module

torch.manual_seed(0)

def conv3x3(in_planes, out_planes, stride=1):
    """3x3 convolution with padding"""
    return Conv2d(in_planes, out_planes, kernel_size=3, stride=stride, padding=1, bias=False)

ARGS = {
    "in_planes": 3,
    "out_planes": 64,
    "stride": 1
}
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 Conv2d — a Conv2d output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Conv2d 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
find_thresholds_by_FAR Ran
ran on a synthesized input
HuangYG123/CurricularFace/util/metrics.py
code served (permissive licence) · get_code("ffdd627e5c14a111")
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 1, dtype family float, shape (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

# Create sample score and label vectors
np.random.seed(0)

# score_vec: 1D array of scores
score_vec = np.random.rand(10).astype(np.float32)

# label_vec: 1D boolean array with same shape as score_vec
# Make sure we have at least one negative sample (False)
label_vec = np.array([True, False, True, False, True, False, True, False, True, False], dtype=np.bool_)

# FARs: list/array of False Alarm Rates (can be None or a list of floats between 0 and 1)
FARs = [0.01, 0.1, 0.5]

# epsilon: small value for numerical stability
epsilon = 1e-08

ARGS = {
    "score_vec": score_vec,
    "label_vec": label_vec,
    "FARs": FARs,
    "epsilon": epsilon,
}
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 ndarray, ndim 1, dtype family float, shape (3,) — shape (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
get_block Ran
ran, honoured the contract we drafted
HuangYG123/CurricularFace/backbone/model_irse.py
code served (permissive licence) · get_code("ad5dbf57f3ea2633")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Sonnet 4.5, prompt python_fixture_v1) and returned list, ndim 2, dtype family int; 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
from collections import namedtuple

torch.manual_seed(0)

# The function get_block takes in_channel, depth, num_units, and stride
# It returns a list of Bottleneck namedtuples

ARGS = {
    'in_channel': 64,
    'depth': 128,
    'num_units': 3,
    'stride': 2
}
Setup block drafted by
a language model, Claude Sonnet 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 int — a list output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
get_blocks Ran
ran; our contract draft was wrong, not the code
HuangYG123/CurricularFace/backbone/model_irse.py
pointer only (licence: MIT); the stored copy get_code serves, from VladAndronik/face_swap, is pointer only (licence: NONE) · get_code("ea7941eb1ea8ffb2")
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 VladAndronik/face_swap, 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.

l2_norm Ran
ran; our contract draft was wrong, not the code
HuangYG123/CurricularFace/backbone/model_irse.py
code served (permissive licence) · get_code("c54fea429589425d")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Sonnet 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. 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)

input_tensor = torch.randn(2, 4)
axis = 1

ARGS = {
    'input': input_tensor,
    'axis': axis
}
Setup block drafted by
a language model, Claude Sonnet 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
read_samples_from_record Ran
ran on a synthesized input
HuangYG123/CurricularFace/dataset/datasets.py
code served (permissive licence) · get_code("ea06a508974b1a38")
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 os
import tempfile
from collections import defaultdict

# Create a temporary directory and record file for testing
temp_dir = tempfile.mkdtemp()
root_dir = temp_dir

# Create some test image files (just empty files)
os.makedirs(os.path.join(root_dir, "images"), exist_ok=True)
open(os.path.join(root_dir, "images", "img1.jpg"), 'w').close()
open(os.path.join(root_dir, "images", "img2.jpg"), 'w').close()
open(os.path.join(root_dir, "images", "img3.jpg"), 'w').close()

# Create a record file with image paths and labels
record_file = os.path.join(temp_dir, "record.txt")
with open(record_file, 'w') as f:
    f.write("images/img1.jpg 0\n")
    f.write("images/img2.jpg 1\n")
    f.write("images/img3.jpg 0\n")

ARGS = {
    "root_dir": root_dir,
    "record_dir": record_file,
    "Train": 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 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

Also present in other papers' code

Of 13 samples lifted from this paper's linked repositories, 13 carry a function identity and 0 do not. Of the 13 distinct functions with an identity, 8 are also present in other papers whose code we hold (7 with at least one paper dated earlier, 0 the same month, 8 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
l2_norm 29 other papers
6 dated earlier, 23 later · 2 only as an assembly brick
1804.07573 (2018-04)
MobileFaceNets: Efficient CNNs for Accurate Real-Time Face Verification on Mobile Devices
1903.08527 (2019-03)
Accurate 3D Face Reconstruction with Weakly-Supervised Learning: From Single Image to Image Set
1905.11946 (2019-05)
EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks
and 26 more not named here
get_block 27 other papers
6 dated earlier, 21 later
1804.07573 (2018-04)
MobileFaceNets: Efficient CNNs for Accurate Real-Time Face Verification on Mobile Devices
1903.08527 (2019-03)
Accurate 3D Face Reconstruction with Weakly-Supervised Learning: From Single Image to Image Set
1907.10830 (2019-07)
U-GAT-IT: Unsupervised Generative Attentional Networks with Adaptive Layer-Instance Normalization for Image-to-Image Translation
and 24 more not named here
get_blocks 16 other papers
6 dated earlier, 10 later
1804.07573 (2018-04)
MobileFaceNets: Efficient CNNs for Accurate Real-Time Face Verification on Mobile Devices
1903.08527 (2019-03)
Accurate 3D Face Reconstruction with Weakly-Supervised Learning: From Single Image to Image Set
1907.10830 (2019-07)
U-GAT-IT: Unsupervised Generative Attentional Networks with Adaptive Layer-Instance Normalization for Image-to-Image Translation
and 13 more not named here
calculate_accuracy 12 other papers
7 dated earlier, 5 later
1503.03832 (2015-03)
FaceNet: A Unified Embedding for Face Recognition and Clustering
1607.08221 (2016-07)
MS-Celeb-1M: A Dataset and Benchmark for Large-Scale Face Recognition
1703.07737 (2017-03)
In Defense of the Triplet Loss for Person Re-Identification
and 9 more not named here
conv3x3 9 other papers
1 dated earlier, 8 later
2001.06838 (2020-01)
Towards Stabilizing Batch Statistics in Backward Propagation of Batch Normalization
2107.08621 (2021-07)
Face.evoLVe: A High-Performance Face Recognition Library
2210.10090 (2022-10)
How to Boost Face Recognition with StyleGAN?
and 6 more not named here
conv1x1 8 other papers
1 dated earlier, 7 later
2001.06838 (2020-01)
Towards Stabilizing Batch Statistics in Backward Propagation of Batch Normalization
2008.04115 (2020-08)
T-GD: Transferable GAN-generated Images Detection Framework
2107.08621 (2021-07)
Face.evoLVe: A High-Performance Face Recognition Library
and 5 more not named here
ResNet_50 4 other papers
4 later
2107.08621 (2021-07)
Face.evoLVe: A High-Performance Face Recognition Library
2210.10090 (2022-10)
How to Boost Face Recognition with StyleGAN?
2408.05939 (2024-08)
UniPortrait: A Unified Framework for Identity-Preserving Single- and Multi-Human Image Personalization
and 1 more not named here
calculate_val 2 other papers
1 dated earlier, 1 later
1801.07698 (2018-01)
ArcFace: Additive Angular Margin Loss for Deep Face Recognition
2103.14803 (2021-03)
Face Transformer for Recognition

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 2026-11-21.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

As an emerging topic in face recognition, designing margin-based loss functions can increase the feature margin between different classes for enhanced discriminability. More recently, the idea of mining-based strategies is adopted to emphasize the misclassified samples, achieving promising results. However, during the entire training process, the prior methods either do not explicitly emphasize the sample based on its importance that renders the hard samples not fully exploited; or explicitly emphasize the effects of semi-hard/hard samples even at the early training stage that may lead to convergence issue. In this work, we propose a novel Adaptive Curriculum Learning loss (CurricularFace) that embeds the idea of curriculum learning into the loss function to achieve a novel training strategy for deep face recognition, which mainly addresses easy samples in the early training stage and hard ones in the later stage. Specifically, our CurricularFace adaptively adjusts the relative importance of easy and hard samples during different training stages. In each stage, different samples are assigned with different importance according to their corresponding difficultness. Extensive experimental results on popular benchmarks demonstrate the superiority of our CurricularFace over the state-of-the-art competitors.

For agents

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

get_harvested_code_for_paper("2004.00288")
get_code_for_paper("2004.00288")
have("2004.00288")

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

curl https://syntology.ai/api/ran/2004.00288.json

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

[![Syntology run record](https://syntology.ai/api/ran/2004.00288.svg)](https://syntology.ai/paper/2004.00288)

Connect an agent — have() is free.