SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1906.06579 · 2019

EXTD: Extremely Tiny Face Detector via Iterative Filter Reuse

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; 0 honoured the contract we drafted, 0 violated it and 7 have no contract verdict; 6 failures of our instrument; 0 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/1906.06579.json · badge: /api/ran/1906.06579.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
clovaai/EXTD_Pytorch canonical (classifier label)
classifier basis: census
11 ran · 0 no recorded run
SeungyounShin/EXTD reimplementation (classifier label)
classifier basis: classifier_2025
2 ran · 0 no recorded run
FunctionStatusWhere it lives
add_extras_mobileFace Ran
ran on a synthesized input
clovaai/EXTD_Pytorch/EXTD_32.py
code served (permissive licence) · get_code("f7dd0fc85b99dadd")
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 list, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn

torch.manual_seed(0)

ARGS = {
    "in_channel": 32
}
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 list, ndim 1, dtype family float — a list output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
bbox_predictor Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
SeungyounShin/EXTD/EXTD.py
pointer only (licence: NONE) · get_code("f2e50f49c109c032")
How it ran

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

center_size Ran
ran on a synthesized input
clovaai/EXTD_Pytorch/layers/bbox_utils.py
code served (permissive licence) · get_code("1283157b9af10fce")
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 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)

# The function takes boxes in point_form (xmin, ymin, xmax, ymax)
# and converts to center-size form (cx, cy, w, h)
# boxes should be a tensor of shape (batch_size, 4)

boxes = torch.tensor([
    [10.0, 20.0, 30.0, 40.0],
    [5.0, 15.0, 25.0, 35.0]
])

ARGS = {
    "boxes": boxes
}
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 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
class_predictor Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
SeungyounShin/EXTD/EXTD.py
pointer only (licence: NONE) · get_code("c112e1e0e8ec8d8c")
How it ran

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

conv_1x1_bn Ran
ran on a synthesized input
clovaai/EXTD_Pytorch/mobileFacenet_32_PReLU.py
code served (permissive licence) · get_code("20c32d5668430dcd")
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 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 takes two parameters: inp (input channels) and oup (output channels)
# These are integers representing the number of input and output channels for the Conv2d layer

ARGS = {
    "inp": 3,
    "oup": 16
}
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 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
conv_bn Ran
ran on a synthesized input
clovaai/EXTD_Pytorch/mobileFacenet_32_PReLU.py
code served (permissive licence) · get_code("18aec87cea006da3")
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 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)

ARGS = {
    'inp': 3,
    'oup': 16,
    'stride': 1,
    'k_size': 3
}
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 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
get_gt_boxes Ran
ran; our contract draft was wrong, not the code
clovaai/EXTD_Pytorch/eval_tools/evaluation.py
code served (permissive licence) · get_code("b8864cd681dd62de")
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
from scipy.io import savemat

# Create a temporary directory with mock .mat files
temp_dir = tempfile.mkdtemp()

# Create mock data for the .mat files
facebox_list = np.array([[1, 2, 3, 4], [5, 6, 7, 8]], dtype=object)
event_list = np.array([['event1'], ['event2']], dtype=object)
file_list = np.array([['file1.jpg'], ['file2.jpg']], dtype=object)
hard_gt_list = np.array([[10, 20, 30, 40]], dtype=object)
medium_gt_list = np.array([[50, 60, 70, 80]], dtype=object)
easy_gt_list = np.array([[90, 100, 110, 120]], dtype=object)

# Save the mock data to .mat files
savemat(os.path.join(temp_dir, 'wider_face_val.mat'), {
    'face_bbx_list': facebox_list,
    'event_list': event_list,
    'file_list': file_list
})

savemat(os.path.join(temp_dir, 'wider_hard_val.mat'), {
    'gt_list': hard_gt_list
})

savemat(os.path.join(temp_dir, 'wider_medium_val.mat'), {
    'gt_list': medium_gt_list
})

savemat(os.path.join(temp_dir, 'wider_easy_val.mat'), {
    'gt_list': easy_gt_list
})

ARGS = {
    'gt_dir': temp_dir
}
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 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_gt_boxes_from_txt Ran
ran; our contract draft was wrong, not the code
clovaai/EXTD_Pytorch/eval_tools/evaluation.py
code served (permissive licence) · get_code("fa5b759a22a26bb3")
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 os
import tempfile
import numpy as np

# Create a temporary directory for cache
temp_dir = tempfile.mkdtemp()

# Create a temporary gt_path file with valid ground truth box data
gt_file_path = os.path.join(temp_dir, 'gt_boxes.txt')

# Write sample ground truth data in the expected format
gt_content = """--image_001
info_line
10.5 20.3 50.7 60.2
15.0 25.0 55.0 65.0
--image_002
info_line
30.1 40.2 70.3 80.4
35.5 45.5 75.5 85.5
--image_003
info_line
5.0 10.0 100.0 110.0
"""

with open(gt_file_path, 'w') as f:
    f.write(gt_content)

ARGS = {
    'gt_path': gt_file_path,
    'cache_dir': temp_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 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
intersect Ran
ran; our contract draft was wrong, not the code
clovaai/EXTD_Pytorch/layers/bbox_utils.py
pointer only (licence: MIT); the stored copy get_code serves, from NVlabs/AL-MDN, is pointer only (licence: NOASSERTION) · get_code("f01c5f73515770a5")
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 NVlabs/AL-MDN, and GitHub could not classify the licence of that repository (NOASSERTION). 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.

parse_wider_file Ran
ran on a synthesized input
clovaai/EXTD_Pytorch/prepare_wider_data.py
code served (permissive licence) · get_code("0feb7fb26b64b471")
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

# Create a temporary directory and a test annotation file
temp_dir = tempfile.mkdtemp()

# Create a simple WIDER format annotation file
annotation_content = """event1/img_1.jpg
2
10 20 100 120
15 25 110 130
event1/img_2.jpg
1
50 60 150 160
"""

annotation_file = os.path.join(temp_dir, "wider_annotations.txt")
with open(annotation_file, 'w') as f:
    f.write(annotation_content)

# Create dummy image files
os.makedirs(os.path.join(temp_dir, "event1"), exist_ok=True)
open(os.path.join(temp_dir, "event1", "img_1.jpg"), 'w').close()
open(os.path.join(temp_dir, "event1", "img_2.jpg"), 'w').close()

ARGS = {
    "root": temp_dir,
    "file": annotation_file,
}
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
point_form Ran
ran; our contract draft was wrong, not the code
clovaai/EXTD_Pytorch/layers/bbox_utils.py
pointer only (licence: MIT); the stored copy get_code serves, from NVlabs/AL-MDN, is pointer only (licence: NOASSERTION) · get_code("3bea1addc1e14162")
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 NVlabs/AL-MDN, and GitHub could not classify the licence of that repository (NOASSERTION). 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.

read_pred_file Ran
ran on a synthesized input
clovaai/EXTD_Pytorch/eval_tools/evaluation.py
code served (permissive licence) · get_code("524d5a85cd29ba7c")
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
import tempfile
import os

# Create a temporary file with the expected format
temp_dir = tempfile.mkdtemp()
filepath = os.path.join(temp_dir, "pred_file.txt")

# Write content in the expected format:
# Line 0: image file path
# Line 1: (empty/skipped)
# Lines 2+: boxes with space-separated floats
with open(filepath, 'w') as f:
    f.write("path/to/image.jpg\n")
    f.write("\n")
    f.write("0.1 0.2 0.3 0.4\n")
    f.write("0.5 0.6 0.7 0.8\n")
    f.write("0.9 1.0 1.1 1.2\n")

ARGS = {
    "filepath": filepath
}
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
upsample Ran
ran on a synthesized input
clovaai/EXTD_Pytorch/EXTD_32.py
code served (permissive licence) · get_code("40d252cf22e9e1ea")
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 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)

ARGS = {
    "in_channels": 3,
    "out_channels": 16,
}
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 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

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, 6 are also present in other papers whose code we hold (6 with at least one paper dated earlier, 0 the same month, 4 later) and 7 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
intersect 14 other papers
7 dated earlier, 7 later
1409.1556 (2014-09)
Very Deep Convolutional Networks for Large-Scale Image Recognition
1512.02325 (2015-12)
SSD: Single Shot MultiBox Detector
and 11 more not named here
point_form 12 other papers
6 dated earlier, 6 later
1409.1556 (2014-09)
Very Deep Convolutional Networks for Large-Scale Image Recognition
1512.02325 (2015-12)
SSD: Single Shot MultiBox Detector
and 9 more not named here
get_gt_boxes 2 other papers
1 dated earlier, 1 later
1905.00641 (2019-05)
RetinaFace: Single-stage Dense Face Localisation in the Wild
2411.02537 (2024-11)
INQUIRE: A Natural World Text-to-Image Retrieval Benchmark
get_gt_boxes_from_txt 2 other papers
1 dated earlier, 1 later
1905.00641 (2019-05)
RetinaFace: Single-stage Dense Face Localisation in the Wild
2411.02537 (2024-11)
INQUIRE: A Natural World Text-to-Image Retrieval Benchmark
center_size 1 other paper
1 dated earlier
1409.1556 (2014-09)
Very Deep Convolutional Networks for Large-Scale Image Recognition
read_pred_file 1 other paper
1 dated earlier
1905.00641 (2019-05)
RetinaFace: Single-stage Dense Face Localisation in the Wild

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

In this paper, we propose a new multi-scale face detector having an extremely tiny number of parameters (EXTD),less than 0.1 million, as well as achieving comparable performance to deep heavy detectors. While existing multi-scale face detectors extract feature maps with different scales from a single backbone network, our method generates the feature maps by iteratively reusing a shared lightweight and shallow backbone network. This iterative sharing of the backbone network significantly reduces the number of parameters, and also provides the abstract image semantics captured from the higher stage of the network layers to the lower-level feature map. The proposed idea is employed by various model architectures and evaluated by extensive experiments. From the experiments from WIDER FACE dataset, we show that the proposed face detector can handle faces with various scale and conditions, and achieved comparable performance to the more massive face detectors that few hundreds and tens times heavier in model size and floating point operations.

For agents

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

get_harvested_code_for_paper("1906.06579")
get_code_for_paper("1906.06579")
have("1906.06579")

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

curl https://syntology.ai/api/ran/1906.06579.json

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

[![Syntology run record](https://syntology.ai/api/ran/1906.06579.svg)](https://syntology.ai/paper/1906.06579)

Connect an agent — have() is free.