SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1411.5752 · 2014

Hypercolumns for Object Segmentation and Fine-grained Localization

arXiv · PDF · Open in the Atlas

Code that ran

10 of 11 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; 0 failures of our instrument; 1 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/1411.5752.json · badge: /api/ran/1411.5752.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
varunagrawal/tiny-faces-pytorch community (classifier label)
classifier basis not recorded on the link
10 ran · 1 no recorded run
FunctionStatusWhere it lives
balance_sampling Ran
ran on a synthesized input
varunagrawal/tiny-faces-pytorch/tinyfaces/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
centralize_bbox Ran
ran on a synthesized input
varunagrawal/tiny-faces-pytorch/tinyfaces/clustering/cluster.py
code served (permissive licence) · get_code("8f54e94fcd216e5b")
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 (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

# Create a small bounding box array
# Format: [x, y, w, h] for each bbox
# The function expects bboxes with shape (N, 4) where:
# - bboxes[:, 0] = x coordinates
# - bboxes[:, 1] = y coordinates
# - bboxes[:, 2] = right/width values
# - bboxes[:, 3] = bottom/height values

bboxes = np.array([
    [10, 20, 50, 80],
    [5, 15, 45, 75]
], dtype=np.float32)

ARGS = {
    "bboxes": bboxes
}
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 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
compute_dense_overlap Ran
ran on a synthesized input
varunagrawal/tiny-faces-pytorch/tinyfaces/datasets/dense_overlap.py
code served (permissive licence) · get_code("6f63cf76e9115cd2")
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 4, dtype family float, shape (4, 5, 2, 2) 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

# Define ARGS dict with valid arguments for compute_dense_overlap
ARGS = {
    'ofx': 10.0,  # offset x
    'ofy': 5.0,   # offset y
    'stx': 2.0,   # stride x
    'sty': 1.5,   # stride y
    'vsx': 5,     # view size x (number of positions in x)
    'vsy': 4,     # view size y (number of positions in y)
    'dx1': np.array([-2.0, -1.5]),  # delta x1 for 2 templates
    'dy1': np.array([-1.0, -0.5]),  # delta y1 for 2 templates
    'dx2': np.array([2.0, 2.5]),    # delta x2 for 2 templates
    'dy2': np.array([1.0, 1.5]),    # delta y2 for 2 templates
    'gx1': np.array([5.0, 15.0]),   # ground truth x1 for 2 boxes
    'gy1': np.array([8.0, 12.0]),   # ground truth y1 for 2 boxes
    'gx2': np.array([20.0, 30.0]),  # ground truth x2 for 2 boxes
    'gy2': np.array([18.0, 25.0]),  # ground truth y2 for 2 boxes
    'zmx': 1,                         # zoom x
    'zmy': 1,                         # zoom y
}
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 4, dtype family float, shape (4, 5, 2, 2) — shape (4, 5, 2, 2) 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_cluster_indices Ran
ran on a synthesized input
varunagrawal/tiny-faces-pytorch/tinyfaces/clustering/k_medoids.py
code served (permissive licence) · get_code("d028f657724b788f")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned ndarray, ndim 1, dtype family int, shape (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

# Construct arguments for get_cluster_indices
# The function uses:
# - distances: a 2D array where distances[medoid_idxs, :] is indexed
# - medoid_idxs: indices to select rows from distances

# Create a small distance matrix (5 samples, 4 features/medoids)
distances = np.array([
    [1.0, 2.0, 3.0, 4.0],
    [2.0, 1.0, 4.0, 3.0],
    [3.0, 4.0, 1.0, 2.0],
    [4.0, 3.0, 2.0, 1.0],
    [1.5, 2.5, 3.5, 0.5],
], dtype=np.float32)

# Create medoid indices (selecting 3 medoids from the 5 samples)
medoid_idxs = np.array([0, 2, 4], dtype=np.int32)

ARGS = {
    "distances": distances,
    "medoid_idxs": medoid_idxs,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 1, dtype family int, shape (4,) — shape (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
jaccard_index Ran
ran on a synthesized input
varunagrawal/tiny-faces-pytorch/tinyfaces/metrics.py
code served (permissive licence) · get_code("7e62d68737fe7c60")
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 torch
import numpy as np

# Set seed for reproducibility
torch.manual_seed(0)

# Create simple box coordinates in (x1, y1, x2, y2) format
# box_a: a bounding box with coordinates (10, 20, 50, 60)
box_a = [10, 20, 50, 60]

# box_b: another bounding box with coordinates (30, 40, 70, 80)
box_b = [30, 40, 70, 80]

# indices: optional list to help with debugging
indices = [0, 1]

ARGS = {
    'box_a': box_a,
    'box_b': box_b,
    'indices': indices
}
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 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
kMedoids Ran
ran on a synthesized input
varunagrawal/tiny-faces-pytorch/tinyfaces/clustering/k_medoids.py
code served (permissive licence) · get_code("412c94801cb77021")
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

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

# Create a small distance matrix
# For k-medoids, we need a symmetric distance matrix
# Let's create a 5x5 distance matrix (5 data points)
n_points = 5
distances = np.array([
    [0.0, 1.2, 2.3, 3.1, 2.8],
    [1.2, 0.0, 1.5, 2.2, 2.5],
    [2.3, 1.5, 0.0, 1.1, 1.8],
    [3.1, 2.2, 1.1, 0.0, 0.9],
    [2.8, 2.5, 1.8, 0.9, 0.0]
], dtype=np.float32)

# Number of clusters
k = 2

ARGS = {
    "distances": distances,
    "k": k
}
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
rect_dist Ran
ran on a synthesized input
varunagrawal/tiny-faces-pytorch/tinyfaces/metrics.py
code served (permissive licence) · get_code("7365247cee10c002")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned ndarray, ndim 1, dtype family float, shape (2,) 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

# rect_dist expects two arguments: I and J
# Both should be numpy arrays representing rectangles
# Each rectangle is defined by [x1, y1, x2, y2] coordinates
# The function handles both 1D (single rectangle) and 2D (batch of rectangles) inputs

# Create sample rectangle batches
# Using distinct sizes: 2 rectangles with 4 coordinates each
I = np.array([
    [10, 20, 50, 80],      # rectangle 1: x1=10, y1=20, x2=50, y2=80
    [30, 40, 100, 120]     # rectangle 2: x1=30, y1=40, x2=100, y2=120
], dtype=np.float32)

J = np.array([
    [15, 25, 60, 90],      # rectangle 1: x1=15, y1=25, x2=60, y2=90
    [35, 45, 95, 125]      # rectangle 2: x1=35, y1=45, x2=95, y2=125
], dtype=np.float32)

ARGS = {
    "I": I,
    "J": J
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 1, dtype family float, shape (2,) — shape (2,) 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
regression_refinement Ran
ran on a synthesized input
varunagrawal/tiny-faces-pytorch/tinyfaces/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
update_medoids Ran
ran on a synthesized input
varunagrawal/tiny-faces-pytorch/tinyfaces/clustering/k_medoids.py
code served (permissive licence) · get_code("526545ae72e11869")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned ndarray, ndim 1, dtype family int, shape (3,) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np
import warnings

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

# Create a small distance matrix (pairwise distances between 5 points)
# Shape: (5, 5) - symmetric distance matrix
distances = np.array([
    [0.0, 1.2, 2.3, 3.1, 1.5],
    [1.2, 0.0, 1.8, 2.5, 2.1],
    [2.3, 1.8, 0.0, 1.9, 2.8],
    [3.1, 2.5, 1.9, 0.0, 3.2],
    [1.5, 2.1, 2.8, 3.2, 0.0]
], dtype=np.float32)

# Cluster assignments for each point (5 points assigned to 3 clusters)
# Shape: (5,) - each element is a cluster index
cluster_idxs = np.array([0, 0, 1, 1, 2], dtype=np.int32)

# Current medoid indices for each cluster (3 clusters)
# Shape: (3,) - each element is the point index that is the medoid for that cluster
medoid_idxs = np.array([0, 2, 4], dtype=np.int32)

ARGS = {
    'distances': distances,
    'cluster_idxs': cluster_idxs,
    'medoid_idxs': medoid_idxs
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 1, dtype family int, shape (3,) — shape (3,) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
voc_ap Ran
ran on a synthesized input
varunagrawal/tiny-faces-pytorch/tinyfaces/metrics.py
code served (permissive licence) · get_code("1583e930fade515a")
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

# Construct valid arguments for voc_ap function
# rec: recall array (should be monotonically increasing values between 0 and 1)
# prec: precision array (should be between 0 and 1)
# use_07_metric: boolean flag (not used in the function body, but included as parameter)

rec = np.array([0.1, 0.3, 0.5, 0.7, 0.9])
prec = np.array([0.95, 0.85, 0.75, 0.65, 0.55])
use_07_metric = False

ARGS = {
    'rec': rec,
    'prec': prec,
    'use_07_metric': use_07_metric
}
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
get_bboxes Not yet run
unverified (harvested, no recorded run)
varunagrawal/tiny-faces-pytorch/tinyfaces/models/utils.py
code served (permissive licence) · get_code("504c00b3ebee7453")
How it ran

no invocation recorded.

Also present in other papers' code

Of 11 samples lifted from this paper's linked repositories, 11 carry a function identity and 0 do not. Of the 11 distinct functions with an identity, 3 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 3 later) and 8 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
balance_sampling 1 other paper
1 later
1612.04402 (2016-12)
Finding Tiny Faces
get_bboxes 1 other paper
1 later
1612.04402 (2016-12)
Finding Tiny Faces
regression_refinement 1 other paper
1 later
1612.04402 (2016-12)
Finding Tiny Faces

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

Recognition algorithms based on convolutional networks (CNNs) typically use the output of the last layer as feature representation. However, the information in this layer may be too coarse to allow precise localization. On the contrary, earlier layers may be precise in localization but will not capture semantics. To get the best of both worlds, we define the hypercolumn at a pixel as the vector of activations of all CNN units above that pixel. Using hypercolumns as pixel descriptors, we show results on three fine-grained localization tasks: simultaneous detection and segmentation[22], where we improve state-of-the-art from 49.7[22] mean AP^r to 60.0, keypoint localization, where we get a 3.3 point boost over[20] and part labeling, where we show a 6.6 point gain over a strong baseline.

For agents

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

get_harvested_code_for_paper("1411.5752")
get_code_for_paper("1411.5752")
have("1411.5752")

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

curl https://syntology.ai/api/ran/1411.5752.json

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

[![Syntology run record](https://syntology.ai/api/ran/1411.5752.svg)](https://syntology.ai/paper/1411.5752)

Connect an agent — have() is free.