SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1902.10107 · 2019

Utterance-level Aggregation For Speaker Recognition In The Wild

arXiv · PDF · Open in the Atlas

Code that ran

13 of 23 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 10 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/1902.10107.json · badge: /api/ran/1902.10107.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
WeidiXie/VGG-Speaker-Recognition canonical (classifier label)
classifier basis: classifier_2025
0 ran · 1 no recorded run
taylorlu/Speaker-Diarization community (classifier label)
classifier basis not recorded on the link
10 ran · 5 no recorded run
khassanoff/Speaker_Verification reimplementation (classifier label)
classifier basis: classifier_2025
3 ran · 0 no recorded run
zabir-nabil/tf2-speaker-recognition community (classifier label)
classifier basis not recorded on the link
0 ran · 4 no recorded run
FunctionStatusWhere it lives
append2dict Ran
ran on a synthesized input
taylorlu/Speaker-Diarization/speakerDiarization.py
code served (permissive licence) · get_code("1169b496229daf10")
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 dict. 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)

# speakerSlice: a dict that maps speaker IDs to lists of time dicts
speakerSlice = {
    "speaker_1": [
        {"start": 0, "stop": 10},
        {"start": 20, "stop": 30}
    ]
}

# spk_period: a dict with a single key-value pair where the value is a tuple/list of two floats
spk_period = {
    "speaker_1": [5.4, 15.7]
}

ARGS = {
    "speakerSlice": speakerSlice,
    "spk_period": spk_period
}
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 dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict 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
arrangeResult Ran
ran on a synthesized input
taylorlu/Speaker-Diarization/speakerDiarization.py
code served (permissive licence) · get_code("fb447405850e532c")
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 dict. 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)

# Analyzing the function `arrangeResult`:
# - labels: iterable that supports indexing and len(), contains label values
# - time_spec_rate: numeric value (int or float) used for multiplication

# The function iterates through labels, compares consecutive elements,
# and uses time_spec_rate to compute time boundaries.

# Create a simple labels list with some repeated and changing values
labels = [0, 0, 1, 1, 2, 2, 2, 1]

# time_spec_rate is a numeric multiplier for time calculations
time_spec_rate = 0.1

ARGS = {
    "labels": labels,
    "time_spec_rate": time_spec_rate
}
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 dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict 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
compute_sequence_match_accuracy Ran
ran on a synthesized input
taylorlu/Speaker-Diarization/uisrnn/evals.py
code served (permissive licence) · get_code("6a0b276952bdc83b")
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
from scipy import optimize

# The function compute_sequence_match_accuracy takes two sequences (lists)
# and computes matching accuracy using optimal assignment

# Create two simple sequences of integers with the same length
sequence1 = [0, 1, 0, 1, 2]
sequence2 = [1, 0, 1, 0, 2]

ARGS = {
    'sequence1': sequence1,
    'sequence2': sequence2,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, 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
conv_block Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
khassanoff/Speaker_Verification/speech_embedder_net.py
code served (permissive licence) · get_code("17437f2725b26ae5")
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 Sequential. 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)

# The function expects:
# - input_dim: an integer representing input channels
# - filters: a list of 3 integers [filter1, filter2, filter3]
# - strides: an integer or tuple for stride parameter

# Looking at the function, it creates a sequential model with:
# - Conv2d(input_dim, filters[0], ..., stride=strides)
# - Conv2d(filters[0], filters[1], ...)
# - Conv2d(filters[1], filters[2], ...)

# So filters must be a list of exactly 3 integers

ARGS = {
    'input_dim': 3,
    'filters': [16, 16, 32],
    'strides': 1
}
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 Sequential — a Sequential output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Sequential output the recorded ndim 1 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
genMap Ran
ran on a synthesized input
taylorlu/Speaker-Diarization/speakerDiarization.py
code served (permissive licence) · get_code("8f290b6646c54127")
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)

# The function expects 'intervals' which should be a tensor-like object
# that supports:
# - .tolist() method
# - indexing with [-1, -1]
# - iteration over rows

# Create a small tensor with intervals (start, end pairs)
intervals = torch.tensor([
    [0, 5],
    [5, 10],
    [10, 15]
], dtype=torch.float32)

ARGS = {
    "intervals": intervals
}
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
get_list_inverse_index Ran
ran on a synthesized input
taylorlu/Speaker-Diarization/uisrnn/evals.py
code served (permissive licence) · get_code("d60a5212d267941e")
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 dict. 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)

# Test with a list of unique integer IDs
unique_ids_int = [10, 20, 30, 40]

ARGS = {
    'unique_ids': unique_ids_int
}
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 dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict 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
get_vggface2_imglist Ran
ran on a synthesized input
taylorlu/Speaker-Diarization/ghostvlad/toolkits.py
code served (permissive licence) · get_code("499126ea0fb9fe0a")
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
import numpy as np

# Create temporary directory and metadata files
temp_dir = tempfile.mkdtemp()
data_dir = tempfile.mkdtemp()

# Create training metadata file
trn_meta_path = os.path.join(temp_dir, 'train.txt')
with open(trn_meta_path, 'w') as f:
    f.write('img1.jpg 0\n')
    f.write('img2.jpg 1\n')
    f.write('img3.jpg 2\n')

# Create validation metadata file
val_meta_path = os.path.join(temp_dir, 'val.txt')
with open(val_meta_path, 'w') as f:
    f.write('img4.jpg 0\n')
    f.write('img5.jpg 1\n')

# Create a simple args class
class Args:
    def __init__(self, trn_meta, val_meta, data_path):
        self.trn_meta = trn_meta
        self.val_meta = val_meta
        self.data_path = data_path

ARGS = {
    'args': Args(
        trn_meta=trn_meta_path,
        val_meta=val_meta_path,
        data_path=data_dir
    )
}
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 — 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
identity_block Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
khassanoff/Speaker_Verification/speech_embedder_net.py
code served (permissive licence) · get_code("ba8e52a6186dbe68")
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 Sequential. 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)

# The function identity_block takes two parameters:
# 1. input_dim: an integer representing the number of input channels
# 2. filters: a list of 3 integers representing filter sizes for the three conv layers

# Looking at the function, it creates a sequential model with:
# - Conv2d(input_dim, filters[0], ...)
# - Conv2d(filters[0], filters[1], ...)
# - Conv2d(filters[1], filters[2], ...)

# So we need to provide valid input_dim and filters arguments

ARGS = {
    'input_dim': 64,
    'filters': [32, 32, 64]
}
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 Sequential — a Sequential output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Sequential output the recorded ndim 1 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
regularization_loss Ran
ran on a synthesized input
taylorlu/Speaker-Diarization/uisrnn/loss_func.py
code served (permissive licence) · get_code("23bad4461ca9d747")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 0, dtype family float, 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

torch.manual_seed(0)

# Create a small neural network with parameters
model = nn.Sequential(
    nn.Linear(4, 3),
    nn.ReLU(),
    nn.Linear(3, 2)
)

# Get all parameters from the model
params = list(model.parameters())

# Weight for the regularization term
weight = 0.01

ARGS = {
    "params": params,
    "weight": weight
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 0, dtype family float, 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
set_mp Ran
ran on a synthesized input
taylorlu/Speaker-Diarization/ghostvlad/toolkits.py
code served (permissive licence) · get_code("35a2331b27bb8ddd")
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 Pool. 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 multiprocessing as mp

ARGS = {
    "processes": 8
}
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 Pool — a Pool output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Pool 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
shortcut Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
khassanoff/Speaker_Verification/speech_embedder_net.py
code served (permissive licence) · get_code("460d697a4c9ca801")
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 Sequential. 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)

# The function signature is: shortcut(input_dim, filters, strides)
# Looking at the function body:
# - input_dim: integer representing input channels
# - filters: a list/array with at least 3 elements, where filters[2] is used as output channels
# - strides: integer or tuple for stride parameter in Conv2d

# Using the global filters variable as reference, each sublist has 3 elements
# Let's use the first one: [48, 48, 96]

ARGS = {
    'input_dim': 48,
    'filters': [48, 48, 96],
    'strides': 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 Sequential — a Sequential output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Sequential output the recorded ndim 1 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
sigma2_prior_loss Ran
ran on a synthesized input
taylorlu/Speaker-Diarization/uisrnn/loss_func.py
code served (permissive licence) · get_code("cd29a997e2d51ed2")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 0, dtype family float, 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

torch.manual_seed(0)

# Construct arguments for sigma2_prior_loss
num_non_zero = 10
sigma_alpha = 2.0
sigma_beta = 3.0
sigma2 = torch.tensor([1.5, 2.0, 0.8, 1.2])

ARGS = {
    "num_non_zero": num_non_zero,
    "sigma_alpha": sigma_alpha,
    "sigma_beta": sigma_beta,
    "sigma2": sigma2,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 0, dtype family float, 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
weighted_mse_loss Ran
ran on a synthesized input
taylorlu/Speaker-Diarization/uisrnn/loss_func.py
code served (permissive licence) · get_code("720fd03354ee7041")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 0, dtype family float, 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

torch.manual_seed(0)

# Create small tensors for input and target
# observation_dim = 4 (last dimension)
# batch size = 2
input_tensor = torch.randn(2, 4)
target_tensor = torch.randn(2, 4)

# weight is used as a diagonal weight matrix with nelement() called on it
# It should have shape (4,) to match observation_dim
weight = torch.ones(4)

ARGS = {
    "input_tensor": input_tensor,
    "target_tensor": target_tensor,
    "weight": weight,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 0, dtype family float, 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
amsoftmax_loss Not yet run
unverified (harvested, no recorded run)
taylorlu/Speaker-Diarization/ghostvlad/model.py
code served (permissive licence) · get_code("6518429bdea14293")
How it ran

no invocation recorded.

amsoftmax_loss Not yet run
unverified (harvested, no recorded run)
zabir-nabil/tf2-speaker-recognition/model.py
code served (permissive licence) · get_code("bf118be3e0b08f1e")
How it ran

no invocation recorded.

conv_block_2D Not yet run
unverified (harvested, no recorded run)
taylorlu/Speaker-Diarization/ghostvlad/backbone.py
code served (permissive licence) · get_code("ba579e59da9e4e3f")
How it ran

no invocation recorded.

conv_block_2D Not yet run
unverified (harvested, no recorded run)
zabir-nabil/tf2-speaker-recognition/backbone.py
code served (permissive licence) · get_code("61a6cf5e47d41f7e")
How it ran

no invocation recorded.

identity_block_2D Not yet run
unverified (harvested, no recorded run)
taylorlu/Speaker-Diarization/ghostvlad/backbone.py
code served (permissive licence) · get_code("43330ce78044b94f")
How it ran

no invocation recorded.

identity_block_2D Not yet run
unverified (harvested, no recorded run)
zabir-nabil/tf2-speaker-recognition/backbone.py
code served (permissive licence) · get_code("c81b6fabb296ede5")
How it ran

no invocation recorded.

initialize_GPU Not yet run
unverified (harvested, no recorded run)
taylorlu/Speaker-Diarization/ghostvlad/toolkits.py
code served (permissive licence) · get_code("d560100c1112033e")
How it ran

no invocation recorded.

resnet_2D_v1 Not yet run
unverified (harvested, no recorded run)
taylorlu/Speaker-Diarization/ghostvlad/backbone.py
code served (permissive licence) · get_code("ea4dfc945d7a3187")
How it ran

no invocation recorded.

resnet_2D_v1 Not yet run
unverified (harvested, no recorded run)
zabir-nabil/tf2-speaker-recognition/backbone.py
code served (permissive licence) · get_code("4b2de008af319598")
How it ran

no invocation recorded.

set_result_path Not yet run
unverified (harvested, no recorded run)
WeidiXie/VGG-Speaker-Recognition/src/predict.py
pointer only (licence: NONE) · get_code("f8272c80971e5978")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, WeidiXie/VGG-Speaker-Recognition 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. No invocation recorded.

Also present in other papers' code

Of 23 samples lifted from this paper's linked repositories, 23 carry a function identity and 0 do not. Of the 19 distinct functions with an identity, 3 are also present in other papers whose code we hold (3 with at least one paper dated earlier, 0 the same month, 0 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
regularization_loss 1 other paper
1 dated earlier
1710.10467 (2017-10)
Generalized End-to-End Loss for Speaker Verification
sigma2_prior_loss 1 other paper
1 dated earlier
1710.10467 (2017-10)
Generalized End-to-End Loss for Speaker Verification
weighted_mse_loss 1 other paper
1 dated earlier
1710.10467 (2017-10)
Generalized End-to-End Loss for Speaker Verification

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 objective of this paper is speaker recognition "in the wild"-where utterances may be of variable length and also contain irrelevant signals. Crucial elements in the design of deep networks for this task are the type of trunk (frame level) network, and the method of temporal aggregation. We propose a powerful speaker recognition deep network, using a "thin-ResNet" trunk architecture, and a dictionary-based NetVLAD or GhostVLAD layer to aggregate features across time, that can be trained end-to-end. We show that our network achieves state of the art performance by a significant margin on the VoxCeleb1 test set for speaker recognition, whilst requiring fewer parameters than previous methods. We also investigate the effect of utterance length on performance, and conclude that for "in the wild" data, a longer length is beneficial.

For agents

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

get_harvested_code_for_paper("1902.10107")
get_code_for_paper("1902.10107")
have("1902.10107")

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

curl https://syntology.ai/api/ran/1902.10107.json

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

[![Syntology run record](https://syntology.ai/api/ran/1902.10107.svg)](https://syntology.ai/paper/1902.10107)

Connect an agent — have() is free.