SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1612.04402 · 2016

Finding Tiny Faces

arXiv · PDF · Open in the Atlas

Code that ran

8 of 14 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; 1 honoured the contract we drafted, 0 violated it and 7 have no contract verdict; 0 failures of our instrument; 6 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/1612.04402.json · badge: /api/ran/1612.04402.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
GuoQuanhao/Tiny-Face-Paddle community (classifier label)
classifier basis not recorded on the link
3 ran · 1 no recorded run
varununleashed/tiny_faces_temp community (classifier label)
classifier basis not recorded on the link
2 ran · 3 no recorded run
gpspelle/Crowd-Counting community (classifier label)
classifier basis not recorded on the link
2 ran · 2 no recorded run
Eltomad/Tiny_Faces community (classifier label)
classifier basis not recorded on the link
1 ran · 0 no recorded run
FunctionStatusWhere it lives
Average Ran
ran on a synthesized input
varununleashed/tiny_faces_temp/tiny_face_regions_of_interest.py
code served (permissive licence) · get_code("905d2ecd38193e5d")
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 int, 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

torch.manual_seed(0)

# The function Average takes a list and computes the average
# It checks len(lst) and uses sum() on it
# It returns an int

# Construct a simple list of numbers
lst = [10, 20, 30, 40]

ARGS = {
    "lst": lst
}
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 int, 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
balance_sampling Ran
ran on a synthesized input
GuoQuanhao/Tiny-Face-Paddle/models/utils.py
code served (permissive licence) · get_code("249c084ce92ceb64")
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 2, dtype family int, shape (4, 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

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

# Create a small label array with shape (4, 3)
# Contains values: 1 (positive), -1 (negative), 0 (ignore)
label_cls = np.array([
    [1, -1, 1],
    [-1, 1, -1],
    [1, -1, 1],
    [-1, 1, 0]
], dtype=np.int32)

# Define parameters for balance_sampling
pos_fraction = 0.5  # 50% positive samples
sample_size = 256   # Total samples to aim for

ARGS = {
    "label_cls": label_cls,
    "pos_fraction": pos_fraction,
    "sample_size": sample_size
}
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 2, dtype family int, shape (4, 3) — shape (4, 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
box_size_calculate Ran
ran on a synthesized input
varununleashed/tiny_faces_temp/tiny_face_regions_of_interest.py
code served (permissive licence) · get_code("280de894d1161d94")
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 1, 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
import numpy as np

torch.manual_seed(0)

# The function box_size_calculate takes a parameter 'r' which should be
# a list/array containing 4 coordinates: [x1, y1, x2, y2]
# It calculates length = x2 - x1 and width = y2 - y1

# Create a simple bounding box representation
# Using a list with 4 elements: [x1, y1, x2, y2]
r = [10, 20, 50, 80]

ARGS = {
    'r': r
}
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, ndim 1, dtype family int — 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
calc_area Ran
ran on a synthesized input
gpspelle/Crowd-Counting/inside.py
code served (permissive licence) · get_code("f848974ea09f6c76")
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

# This function calculates the area of a triangle given three points (A, B, C)
# Each point is defined by x and y coordinates
# The function uses these 6 scalar parameters in arithmetic operations

ARGS = {
    'Ax': 0.0,
    'Ay': 0.0,
    'Bx': 4.0,
    'By': 0.0,
    'Cx': 2.0,
    'Cy': 3.0,
}
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 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
calculate_gain Ran
ran, honoured the contract we drafted
GuoQuanhao/Tiny-Face-Paddle/models/initializer.py
code served (permissive licence) · get_code("d32692be0e110bc9")
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 float, 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 math
import torch
import torch.nn as nn

torch.manual_seed(0)

# The function calculate_gain takes two parameters:
# 1. nonlinearity: a string indicating the type of nonlinearity
# 2. param: optional parameter (used for 'leaky_relu' to specify negative_slope)

# Test with 'relu' (common case)
ARGS = {
    'nonlinearity': 'relu',
    'param': None
}
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 float, 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
get_image Ran
ran on a synthesized input
gpspelle/Crowd-Counting/find_people.py
code served (permissive licence) · get_code("6126b300e57ac811")
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 2, dtype family float, shape (3, 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 tempfile
import numpy as np
from PIL import Image

# Create a temporary PNG file with test image data
temp_dir = tempfile.gettempdir()
temp_path = f"{temp_dir}/test_image.png"

# Create a small test image (3x4 RGB)
test_image_array = np.random.randint(0, 256, (3, 4, 3), dtype=np.uint8)
test_image = Image.fromarray(test_image_array, mode='RGB')
test_image.save(temp_path)

ARGS = {
    "im_path": temp_path
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
PIL, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 2, dtype family float, shape (3, 4) — shape (3, 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
nms Ran
ran on a synthesized input
Eltomad/Tiny_Faces/util.py
code served (permissive licence) · get_code("d6a73bba4327bf04")
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 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 numpy as np

# Create sample detection data: [x1, y1, x2, y2, score]
# Each row is a bounding box with coordinates and confidence score
dets = np.array([
    [10.0, 20.0, 50.0, 60.0, 0.9],
    [12.0, 22.0, 52.0, 62.0, 0.8],
    [100.0, 100.0, 150.0, 150.0, 0.85],
    [15.0, 25.0, 55.0, 65.0, 0.7],
], dtype=np.float32)

prob_thresh = 0.3

ARGS = {
    "dets": dets,
    "prob_thresh": prob_thresh,
}
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 list, ndim 1, 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
regression_refinement Ran
ran on a synthesized input
GuoQuanhao/Tiny-Face-Paddle/models/utils.py
code served (permissive licence) · get_code("20e5dfe30b779157")
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 3, dtype family float, shape (1, 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 numpy as np

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

# Create small test data
batch_size = 2
num_predictions = 4

# tx, ty, tw, th are prediction deltas/offsets - shape (num_predictions,)
tx = np.random.randn(num_predictions).astype(np.float32)
ty = np.random.randn(num_predictions).astype(np.float32)
tw = np.random.randn(num_predictions).astype(np.float32)
th = np.random.randn(num_predictions).astype(np.float32)

# cx, cy, cw, ch are anchor/prior box parameters - shape (batch_size,)
cx = np.random.rand(batch_size).astype(np.float32) * 100
cy = np.random.rand(batch_size).astype(np.float32) * 100
cw = np.random.rand(batch_size).astype(np.float32) * 50 + 10
ch = np.random.rand(batch_size).astype(np.float32) * 50 + 10

# indices to select which predictions apply to which anchors
indices = np.array([0, 1], dtype=np.int32)

ARGS = {
    "tx": tx,
    "ty": ty,
    "tw": tw,
    "th": th,
    "cx": cx,
    "cy": cy,
    "cw": cw,
    "ch": ch,
    "indices": indices,
}
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 3, dtype family float, shape (1, 2, 4) — shape (1, 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
crop_faces_save Not yet run
unverified (harvested, no recorded run)
varununleashed/tiny_faces_temp/tiny_face_regions_of_interest.py
code served (permissive licence) · get_code("08e89c9ecd0e1150")
How it ran

no invocation recorded.

download Not yet run
unverified (harvested, no recorded run)
varununleashed/tiny_faces_temp/download_WIDERFACE.py
code served (permissive licence) · get_code("c1c3c4d881e7b0ae")
How it ran

no invocation recorded.

get_bboxes Not yet run
unverified (harvested, no recorded run)
GuoQuanhao/Tiny-Face-Paddle/models/utils.py
code served (permissive licence) · get_code("504c00b3ebee7453")
How it ran

no invocation recorded.

get_confirm_token Not yet run
unverified (harvested, no recorded run)
varununleashed/tiny_faces_temp/download_WIDERFACE.py
pointer only (licence: MIT); the stored copy get_code serves, from med-air/FedBN, is pointer only (licence: NONE) · get_code("135b3dc835ffe6ad")
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 med-air/FedBN, 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. No invocation recorded.

maaae Not yet run
unverified (harvested, no recorded run)
gpspelle/Crowd-Counting/keras_train.py
code served (permissive licence) · get_code("2e372208416e2c16")
How it ran

no invocation recorded.

mssse Not yet run
unverified (harvested, no recorded run)
gpspelle/Crowd-Counting/keras_train.py
code served (permissive licence) · get_code("fd023123fe830ac0")
How it ran

no invocation recorded.

Also present in other papers' code

Of 14 samples lifted from this paper's linked repositories, 14 carry a function identity and 0 do not. Of the 14 distinct functions with an identity, 5 are also present in other papers whose code we hold (4 with at least one paper dated earlier, 0 the same month, 2 later) and 9 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
get_confirm_token 38 other papers
5 dated earlier, 33 later
1409.8230 (2014-09)
RENOIR - A Dataset for Real Low-Light Image Noise Reduction
1503.03832 (2015-03)
FaceNet: A Unified Embedding for Face Recognition and Clustering
1506.03134 (2015-06)
Pointer Networks
and 35 more not named here
balance_sampling 1 other paper
1 dated earlier
1411.5752 (2014-11)
Hypercolumns for Object Segmentation and Fine-grained Localization
calculate_gain 1 other paper
1 later
1910.02562 (2019-10)
MASTER: Multi-Aspect Non-local Network for Scene Text Recognition
get_bboxes 1 other paper
1 dated earlier
1411.5752 (2014-11)
Hypercolumns for Object Segmentation and Fine-grained Localization
regression_refinement 1 other paper
1 dated earlier
1411.5752 (2014-11)
Hypercolumns for Object Segmentation and Fine-grained Localization

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

Though tremendous strides have been made in object recognition, one of the remaining open challenges is detecting small objects. We explore three aspects of the problem in the context of finding small faces: the role of scale invariance, image resolution, and contextual reasoning. While most recognition approaches aim to be scale-invariant, the cues for recognizing a 3px tall face are fundamentally different than those for recognizing a 300px tall face. We take a different approach and train separate detectors for different scales. To maintain efficiency, detectors are trained in a multi-task fashion: they make use of features extracted from multiple layers of single (deep) feature hierarchy. While training detectors for large objects is straightforward, the crucial challenge remains training detectors for small objects. We show that context is crucial, and define templates that make use of massively-large receptive fields (where 99% of the template extends beyond the object of interest). Finally, we explore the role of scale in pre-trained deep networks, providing ways to extrapolate networks tuned for limited scales to rather extreme ranges. We demonstrate state-of-the-art results on massively-benchmarked face datasets (FDDB and WIDER FACE). In particular, when compared to prior art on WIDER FACE, our results reduce error by a factor of 2 (our models produce an AP of 82% while prior art ranges from 29-64%).

For agents

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

get_harvested_code_for_paper("1612.04402")
get_code_for_paper("1612.04402")
have("1612.04402")

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

curl https://syntology.ai/api/ran/1612.04402.json

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

[![Syntology run record](https://syntology.ai/api/ran/1612.04402.svg)](https://syntology.ai/paper/1612.04402)

Connect an agent — have() is free.