SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2002.08473 · ICML · 2020

Revisiting Training Strategies and Generalization Performance in Deep Metric Learning

Samarth Sinha, Karsten Roth, Timo Milbich, Prateek Gupta, Björn Ommer, Joseph Cohen

arXiv · PDF · Open in the Atlas

Code that ran

12 of 22 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (2 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; 3 failures of our instrument; 10 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/2002.08473.json · badge: /api/ran/2002.08473.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
Confusezius/Deep-Metric-Learning-Baselines canonical (classifier label)
classifier basis: census
4 ran · 4 no recorded run
Confusezius/Revisiting_Deep_Metric_Learning_PyTorch canonical (classifier label)
classifier basis: census
3 ran · 2 no recorded run
confusezius/revisiting_deep_metric_learning_pytorch canonical (classifier label)
classifier basis: classifier_2025
1 ran · 2 no recorded run
douxiaotian/metric community (classifier label)
classifier basis not recorded on the link
2 ran · 1 no recorded run
Confusezius/Deep_Metric_Learning_Baselines community
Papers with Code archive
1 ran · 1 no recorded run
Confusezius/ECCV2020_DiVA_MultiFeature_DML community
Papers with Code archive
1 ran · 0 no recorded run
FunctionStatusWhere it lives
Criterion Ran
ran on a synthesized input · constructed an object; no result was computed
Confusezius/ECCV2020_DiVA_MultiFeature_DML/criteria/stacked_self_distill.py
pointer only (licence: NONE) · get_code("03d1431453ef1a6b")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, Confusezius/ECCV2020_DiVA_MultiFeature_DML 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.

TupleSampler Ran
ran on a synthesized input · constructed an object; no result was computed
Confusezius/Deep_Metric_Learning_Baselines/losses.py
code served (permissive licence) · get_code("7eb35d3431b4daa0")
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, TupleSampler: 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 numpy as np
import random

torch.manual_seed(0)
random.seed(0)

ARGS = {
    'method': 'random'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
faiss, numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type TupleSampler — 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 TupleSampler 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
extension_parameters Ran
ran on a synthesized input
douxiaotian/metric/parameters.py
code served (permissive licence) · get_code("9ec84718a48b4a71")
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 ArgumentParser. 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 argparse
import torch

torch.manual_seed(0)

# Create a simple argument parser
parser = argparse.ArgumentParser()

# The function extension_parameters takes a parser and adds arguments to it
# We need to construct the parser argument
ARGS = {
    'parser': parser
}
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 ArgumentParser — a ArgumentParser output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a ArgumentParser 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
f1_score Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
Confusezius/Deep-Metric-Learning-Baselines/auxiliaries.py
code served (permissive licence) · get_code("d629034ac6132f80")
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 float64, ndim 0, 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

np.random.seed(0)

# Create synthetic data for clustering evaluation
n_samples = 8
embed_dim = 4
n_classes = 3

# Generate target labels (ground truth)
target_labels = np.array([0, 0, 1, 1, 2, 2, 0, 1], dtype=np.int64)

# Generate feature embeddings
feature_coll = np.random.randn(n_samples, embed_dim).astype(np.float32)

# Generate cluster labels from model
model_generated_cluster_labels = np.array([0, 0, 1, 1, 2, 2, 0, 1], dtype=np.int64)

# Generate computed centroids (one per cluster)
computed_centroids = np.random.randn(n_classes, embed_dim).astype(np.float32)

ARGS = {
    "model_generated_cluster_labels": model_generated_cluster_labels,
    "target_labels": target_labels,
    "feature_coll": feature_coll,
    "computed_centroids": computed_centroids,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, pandas, scipy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float64, ndim 0, dtype family float — no output shape is stated: observe_signature.py recorded float ndim 0, which disagrees with the fixture run
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
gimme_params Ran
ran, honoured the contract we drafted
Confusezius/Deep-Metric-Learning-Baselines/auxiliaries.py
code served (permissive licence) · get_code("23c18adfe2fa9b50")
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 int64, ndim 0, dtype family int, 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 numpy as np

torch.manual_seed(0)

class SimpleModel(nn.Module):
    def __init__(self):
        super(SimpleModel, self).__init__()
        self.linear1 = nn.Linear(4, 3)
        self.linear2 = nn.Linear(3, 2)
    
    def forward(self, x):
        x = self.linear1(x)
        x = self.linear2(x)
        return x

model = SimpleModel()

ARGS = {
    "model": model
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, pandas, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type int64, ndim 0, dtype family int, 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
gimme_save_string Ran
ran; our contract draft was wrong, not the code
Confusezius/Deep-Metric-Learning-Baselines/auxiliaries.py
code served (permissive licence) · get_code("88dd3c9b8a356d3a")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned str. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import argparse

ARGS = {
    'opt': argparse.Namespace(
        learning_rate=0.001,
        batch_size=32,
        epochs=10,
        model_name='resnet18',
        optimizer='adam',
        loss_function='cross_entropy',
        hyperparams={
            'momentum': 0.9,
            'weight_decay': 1e-4,
            'dropout': 0.5
        },
        data_config={
            'train_split': 0.8,
            'val_split': 0.1,
            'test_split': 0.1
        },
        device='cpu',
        seed=42,
        verbose=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 str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
googlenet Ran
ran on a synthesized input
Confusezius/Deep-Metric-Learning-Baselines/googlenet.py
code served (permissive licence) · get_code("890262140a75d9b9")
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 GoogLeNet. 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)

# Define the missing BasicConv2d class that GoogLeNet depends on
class BasicConv2d(nn.Module):
    def __init__(self, in_channels, out_channels, **kwargs):
        super(BasicConv2d, self).__init__()
        self.conv = nn.Conv2d(in_channels, out_channels, bias=False, **kwargs)
        self.bn = nn.BatchNorm2d(out_channels, eps=0.001)
    
    def forward(self, x):
        x = self.conv(x)
        x = self.bn(x)
        return torch.relu(x)

# Define the missing InceptionAux class that GoogLeNet depends on
class InceptionAux(nn.Module):
    def __init__(self, in_channels, num_classes):
        super(InceptionAux, self).__init__()
        self.conv = BasicConv2d(in_channels, 128, kernel_size=1)
        self.fc1 = nn.Linear(128 * 4 * 4, 1024)
        self.fc2 = nn.Linear(1024, num_classes)
    
    def forward(self, x):
        x = torch.adaptive_avg_pool2d(x, (4, 4))
        x = self.conv(x)
        x = x.view(x.size(0), -1)
        x = torch.relu(self.fc1(x))
        x = torch.dropout(x, p=0.7, train=self.training)
        x = self.fc2(x)
        return x

# Construct a small input tensor for the forward pass
# GoogLeNet expects 3-channel input (RGB images)
# Batch size 2, height 224, width 224 (standard ImageNet size, but we'll use smaller spatial dims)
input_tensor = torch.randn(2, 3, 224, 224)

# Define ARGS dictionary with all parameters for GoogLeNet
ARGS = {
    'num_classes': 1000,
    'aux_logits': True,
    'transform_input': False,
    'init_weights': True
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
scipy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type GoogLeNet — a GoogLeNet output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a GoogLeNet 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
name_adjust Ran
ran on a synthesized input
Confusezius/Revisiting_Deep_Metric_Learning_PyTorch/Result_Evaluations.py
code served (permissive licence) · get_code("014edb34f249bcfe")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned str. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import numpy as np

torch.manual_seed(0)

ARGS = {
    'n': 'Margin_b06_Distance_experiment_v1',
    'prep': 'Model_',
    'app': '_results',
    'for_plot': True
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
name_filter Ran
ran on a synthesized input
Confusezius/Revisiting_Deep_Metric_Learning_PyTorch/Result_Evaluations.py
code served (permissive licence) · get_code("08e4d248fe54f413")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned str. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

def name_filter(n):
    n = '_'.join(n.split('_')[1:])
    return n

ARGS = {
    'n': 'prefix_middle_suffix'
}
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 str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
wandb_parameters Ran
ran; our contract draft was wrong, not the code
confusezius/revisiting_deep_metric_learning_pytorch/parameters.py
code served (permissive licence) · get_code("5b09d2867ea2a10d")
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 ArgumentParser. 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 argparse
import torch

torch.manual_seed(0)

# Create a minimal ArgumentParser instance
parser = argparse.ArgumentParser(description='Test parser for wandb_parameters')

# Call the function to add arguments
def wandb_parameters(parser):
    parser.add_argument('--log_online', action='store_true', help='Flag. If set, run metrics are stored online in addition to offline logging. Should generally be set.')
    parser.add_argument('--wandb_key', default='<your_api_key_here>', type=str, help='API key for W&B.')
    parser.add_argument('--project', default='Sample_Project', type=str, help='Name of the project - relates to W&B project names. In --savename default setting part of the savename.')
    parser.add_argument('--group', default='Sample_Group', type=str, help='Name of the group - relates to W&B group names - all runs with same setup but different seeds are logged into one group. In --savename default setting part of the savename.')
    return parser

# Build the ARGS dict with valid arguments for wandb_parameters
ARGS = {
    'parser': parser
}
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 ArgumentParser — a ArgumentParser output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a ArgumentParser 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
wandb_parameters Ran
ran on a synthesized input
Confusezius/Revisiting_Deep_Metric_Learning_PyTorch/parameters.py
code served (permissive licence) · get_code("1825bfd5a626ed17")
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 MockArgumentParser. 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 argparse

class MockArgumentParser:
    def __init__(self):
        self.arguments = {}
    
    def add_argument(self, *args, **kwargs):
        # Store the argument configuration
        arg_name = args[0].lstrip('--')
        self.arguments[arg_name] = kwargs
        return self
    
    def parse_args(self):
        return self

parser = MockArgumentParser()

ARGS = {
    'parser': parser
}
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 MockArgumentParser — a MockArgumentParser output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a MockArgumentParser 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
wandb_parameters Ran
ran on a synthesized input
douxiaotian/metric/parameters.py
code served (permissive licence) · get_code("0d2b44d3cf1009d0")
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 ArgumentParser. 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 argparse

def create_parser():
    """Create a minimal ArgumentParser for testing wandb_parameters"""
    return argparse.ArgumentParser()

ARGS = {
    'parser': create_parser()
}
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 ArgumentParser — a ArgumentParser output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a ArgumentParser 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
MarginLoss Not yet run
unverified (harvested, no recorded run)
Confusezius/Deep_Metric_Learning_Baselines/losses.py
code served (permissive licence) · get_code("135e67099969c757")
How it ran

no invocation recorded.

basic_training_parameters Not yet run
unverified (harvested, no recorded run)
confusezius/revisiting_deep_metric_learning_pytorch/parameters.py
code served (permissive licence) · get_code("1edf35955d77e643")
How it ran

no invocation recorded.

basic_training_parameters Not yet run
unverified (harvested, no recorded run)
douxiaotian/metric/parameters.py
code served (permissive licence) · get_code("aca58e0b4399a88e")
How it ran

no invocation recorded.

get_data Not yet run
unverified (harvested, no recorded run)
Confusezius/Revisiting_Deep_Metric_Learning_PyTorch/Result_Evaluations.py
code served (permissive licence) · get_code("3d33c198f5dfd1d8")
How it ran

no invocation recorded.

give_CARS196_datasets Not yet run
unverified (harvested, no recorded run)
Confusezius/Deep-Metric-Learning-Baselines/datasets.py
code served (permissive licence) · get_code("94a2b3fc9d173339")
How it ran

no invocation recorded.

give_CUB200_datasets Not yet run
unverified (harvested, no recorded run)
Confusezius/Deep-Metric-Learning-Baselines/datasets.py
code served (permissive licence) · get_code("4a2b2ab542b49c29")
How it ran

no invocation recorded.

give_dataloaders Not yet run
unverified (harvested, no recorded run)
Confusezius/Deep-Metric-Learning-Baselines/datasets.py
code served (permissive licence) · get_code("cea101048260a728")
How it ran

no invocation recorded.

loss_select Not yet run
unverified (harvested, no recorded run)
Confusezius/Deep-Metric-Learning-Baselines/losses.py
code served (permissive licence) · get_code("719ec941b055b8e4")
How it ran

no invocation recorded.

loss_specific_parameters Not yet run
unverified (harvested, no recorded run)
confusezius/revisiting_deep_metric_learning_pytorch/parameters.py
code served (permissive licence) · get_code("e7d6234c596e2482")
How it ran

no invocation recorded.

loss_specific_parameters Not yet run
unverified (harvested, no recorded run)
Confusezius/Revisiting_Deep_Metric_Learning_PyTorch/parameters.py
code served (permissive licence) · get_code("6d771473f57274c0")
How it ran

no invocation recorded.

Also present in other papers' code

Of 22 samples lifted from this paper's linked repositories, 22 carry a function identity and 0 do not. Of the 20 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 16 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
gimme_params 3 other papers
3 later
2003.11113 (2020-03)
PADS: Policy-Adapted Sampling for Visual Similarity Learning
2108.11179 (2021-08)
Recall@k Surrogate Loss with Large Batches and Similarity Mixup
gimme_save_string 3 other papers
3 later
2003.11113 (2020-03)
PADS: Policy-Adapted Sampling for Visual Similarity Learning
2108.11179 (2021-08)
Recall@k Surrogate Loss with Large Batches and Similarity Mixup
f1_score 2 other papers
2 later
2108.11179 (2021-08)
Recall@k Surrogate Loss with Large Batches and Similarity Mixup
basic_training_parameters 1 other paper
1 later
2009.08348 (2020-09)
S2SD: Simultaneous Similarity-based Self-Distillation for Deep Metric 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

Deep Metric Learning (DML) is arguably one of the most influential lines of research for learning visual similarities with many proposed approaches every year. Although the field benefits from the rapid progress, the divergence in training protocols, architectures, and parameter choices make an unbiased comparison difficult. To provide a consistent reference point, we revisit the most widely used DML objective functions and conduct a study of the crucial parameter choices as well as the commonly neglected mini-batch sampling process. Under consistent comparison, DML objectives show much higher saturation than indicated by literature. Further based on our analysis, we uncover a correlation between the embedding space density and compression to the generalization performance of DML models. Exploiting these insights, we propose a simple, yet effective, training regularization to reliably boost the performance of ranking-based DML models on various standard benchmark datasets. Code and a publicly accessible WandB-repo are available at https://github.com/Confusezius/ Revisiting Deep Metric Learning PyTorch.

For agents

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

get_harvested_code_for_paper("2002.08473")
get_code_for_paper("2002.08473")
have("2002.08473")

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

curl https://syntology.ai/api/ran/2002.08473.json

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

[![Syntology run record](https://syntology.ai/api/ran/2002.08473.svg)](https://syntology.ai/paper/2002.08473)

Connect an agent — have() is free.