SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2411.02537 · NeurIPS · 2024

INQUIRE: A Natural World Text-to-Image Retrieval Benchmark

Alexander Shepard, Grant Van Horn, Sara Beery, Edward Vendrow, Omiros Pantazis, G. Brostow, Kate E. Jones, Oisin Mac Aodha

arXiv · PDF · Open in the Atlas

Code that ran

20 of 25 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, 2 violated it and 0 have no contract verdict; 17 failures of our instrument; 5 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/2411.02537.json · badge: /api/ran/2411.02537.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
biubug6/Pytorch_Retinaface canonical (classifier label)
classifier basis: census
11 ran · 2 no recorded run
inquire-benchmark/INQUIRE canonical (classifier label)
classifier basis: census
5 ran · 1 no recorded run
rom1504/img2dataset canonical (classifier label)
classifier basis: census
4 ran · 2 no recorded run
FunctionStatusWhere it lives
ap_at_k Ran
ran; our contract draft was wrong, not the code
inquire-benchmark/INQUIRE/src/metrics.py
code served (permissive licence) · get_code("df88d02117d3396f")
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

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

# Create sample data for the ap_at_k function
# y_true: binary labels (0 or 1) indicating if items are relevant
# y_pred: predicted scores/probabilities for ranking
# count_pos: total number of positive items in the full set
# k: number of top items to consider

y_true = np.array([1, 0, 1, 0, 1, 0], dtype=int)
y_pred = np.array([0.9, 0.8, 0.7, 0.6, 0.5, 0.4], dtype=float)
count_pos = 3  # Total positive items in the dataset
k = 4  # Consider top 4 items

ARGS = {
    "y_true": y_true,
    "y_pred": y_pred,
    "count_pos": count_pos,
    "k": k,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, sklearn — 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
check_keys Ran
ran, violated the contract we drafted
biubug6/Pytorch_Retinaface/convert_to_onnx.py
code served (permissive licence) · get_code("1d429d5224fe49ff")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1) and returned bool, 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 simple model
class SimpleModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear1 = nn.Linear(4, 8)
        self.linear2 = nn.Linear(8, 3)
    
    def forward(self, x):
        x = self.linear1(x)
        x = self.linear2(x)
        return x

# Instantiate the model
model = SimpleModel()

# Create a pretrained state dict with some overlapping and some extra keys
pretrained_state_dict = {
    'linear1.weight': torch.randn(8, 4),
    'linear1.bias': torch.randn(8),
    'linear2.weight': torch.randn(3, 8),
    'linear2.bias': torch.randn(3),
    'extra_layer.weight': torch.randn(2, 2),  # unused key
}

ARGS = {
    'model': model,
    'pretrained_state_dict': pretrained_state_dict,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type bool, 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
conv_bn Ran
ran; our contract draft was wrong, not the code
biubug6/Pytorch_Retinaface/models/net.py
code served (permissive licence) · get_code("f5414e78ff7a815f")
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 conv_bn takes inp, oup, stride, and leaky parameters
# inp: input channels (int)
# oup: output channels (int)
# stride: stride for Conv2d (int, default=1)
# leaky: negative_slope for LeakyReLU (float, default=0)

ARGS = {
    'inp': 3,
    'oup': 16,
    'stride': 1,
    'leaky': 0.1
}
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_bn1X1 Ran
ran; our contract draft was wrong, not the code
biubug6/Pytorch_Retinaface/models/net.py
code served (permissive licence) · get_code("bb07aa213e6aacc9")
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,
    'leaky': 0.1
}
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_no_relu Ran
ran; our contract draft was wrong, not the code
biubug6/Pytorch_Retinaface/models/net.py
code served (permissive licence) · get_code("aacb8f654f54d534")
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
}
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
download_image Ran
ran; our contract draft was wrong, not the code
rom1504/img2dataset/img2dataset/downloader.py
code served (permissive licence) · get_code("152bc0a5197f4792")
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 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 urllib.request
import io
import ssl
import traceback

def is_disallowed(headers, user_agent_token, disallowed_header_directives):
    """Check if HTTP headers contain an X-Robots-Tag directive disallowing usage"""
    for values in headers.get_all('X-Robots-Tag', []):
        try:
            uatoken_directives = values.split(':', 1)
            directives = [x.strip().lower() for x in uatoken_directives[-1].split(',')]
            ua_token = uatoken_directives[0].lower() if len(uatoken_directives) == 2 else None
            if (ua_token is None or ua_token == user_agent_token) and any((x in disallowed_header_directives for x in directives)):
                return True
        except Exception as err:
            traceback.print_exc()
            print(f'Failed to parse X-Robots-Tag: {values}: {err}')
    return False

def download_image(row, timeout, user_agent_token, disallowed_header_directives, ignore_ssl_certificate):
    """Download an image with urllib"""
    key, url = row
    img_stream = None
    user_agent_string = 'Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:72.0) Gecko/20100101 Firefox/72.0'
    if user_agent_token:
        user_agent_string += f' (compatible; {user_agent_token}; +https://github.com/rom1504/img2dataset)'
    try:
        request = urllib.request.Request(url, data=None, headers={'User-Agent': user_agent_string})
        ctx = ssl.create_default_context()
        if ignore_ssl_certificate:
            ctx.check_hostname = False
            ctx.verify_mode = ssl.CERT_NONE
        with urllib.request.urlopen(request, context=ctx, timeout=timeout) as r:
            if disallowed_header_directives and is_disallowed(r.headers, user_agent_token, disallowed_header_directives):
                return (key, None, 'Use of image disallowed by X-Robots-Tag directive')
            img_stream = io.BytesIO(r.read())
        return (key, img_stream, None)
    except Exception as err:
        if img_stream is not None:
            img_stream.close()
        return (key, None, str(err))


class MockHeaders:
    def get_all(self, header_name, default=None):
        return default if default is not None else []


ARGS = {
    'row': ('image_key_001', 'https://example.com/image.jpg'),
    'timeout': 30,
    'user_agent_token': 'img2dataset-bot',
    'disallowed_header_directives': ['noai', 'noimageai'],
    'ignore_ssl_certificate': False,
}
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, ndim 1, dtype family float — 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
download_image_with_retry Ran
ran; our contract draft was wrong, not the code
rom1504/img2dataset/img2dataset/downloader.py
code served (permissive licence) · get_code("361d700a09bc78e0")
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 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 urllib.request
import io
import ssl
import traceback

def is_disallowed(headers, user_agent_token, disallowed_header_directives):
    """Check if HTTP headers contain an X-Robots-Tag directive disallowing usage"""
    for values in headers.get_all('X-Robots-Tag', []):
        try:
            uatoken_directives = values.split(':', 1)
            directives = [x.strip().lower() for x in uatoken_directives[-1].split(',')]
            ua_token = uatoken_directives[0].lower() if len(uatoken_directives) == 2 else None
            if (ua_token is None or ua_token == user_agent_token) and any((x in disallowed_header_directives for x in directives)):
                return True
        except Exception as err:
            traceback.print_exc()
            print(f'Failed to parse X-Robots-Tag: {values}: {err}')
    return False

def download_image(row, timeout, user_agent_token, disallowed_header_directives, ignore_ssl_certificate):
    """Download an image with urllib"""
    key, url = row
    img_stream = None
    user_agent_string = 'Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:72.0) Gecko/20100101 Firefox/72.0'
    if user_agent_token:
        user_agent_string += f' (compatible; {user_agent_token}; +https://github.com/rom1504/img2dataset)'
    try:
        request = urllib.request.Request(url, data=None, headers={'User-Agent': user_agent_string})
        ctx = ssl.create_default_context()
        if ignore_ssl_certificate:
            ctx.check_hostname = False
            ctx.verify_mode = ssl.CERT_NONE
        with urllib.request.urlopen(request, context=ctx, timeout=timeout) as r:
            if disallowed_header_directives and is_disallowed(r.headers, user_agent_token, disallowed_header_directives):
                return (key, None, 'Use of image disallowed by X-Robots-Tag directive')
            img_stream = io.BytesIO(r.read())
        return (key, img_stream, None)
    except Exception as err:
        if img_stream is not None:
            img_stream.close()
        return (key, None, str(err))

def download_image_with_retry(row, timeout, retries, user_agent_token, disallowed_header_directives, ignore_ssl_certificate):
    """Download an image with urllib, retrying if it fails."""
    for _ in range(retries + 1):
        key, img_stream, err = download_image(row, timeout, user_agent_token, disallowed_header_directives, ignore_ssl_certificate)
        if img_stream is not None:
            return (key, img_stream, err)
    return (key, None, err)


class MockHeaders:
    """Mock HTTP headers object with get_all method"""
    def __init__(self):
        self._headers = {}
    
    def get_all(self, key, default=None):
        return self._headers.get(key, default if default is not None else [])


ARGS = {
    'row': ('image_key_1', 'https://invalid-url-that-will-fail.example.com/image.jpg'),
    'timeout': 5,
    'retries': 2,
    'user_agent_token': 'img2dataset-bot',
    'disallowed_header_directives': ['noai', 'noimageai'],
    'ignore_ssl_certificate': False,
}
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, ndim 1, dtype family float — 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
get_gt_boxes Ran
ran; our contract draft was wrong, not the code
biubug6/Pytorch_Retinaface/widerface_evaluate/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
biubug6/Pytorch_Retinaface/widerface_evaluate/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
get_lmm_prompt Ran
ran; our contract draft was wrong, not the code
inquire-benchmark/INQUIRE/src/lmm_utils.py
code served (permissive licence) · get_code("6e1cc85a886de59f")
How it ran

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

import torch
import numpy as np

# Analyze the function: it takes two parameters
# - query: a string describing what to look for in a picture
# - model_name: a string identifying which model/prompt format to use
# The function uses these parameters with string operations (startswith, ==, f-strings)
# and returns a prompt string

ARGS = {
    "query": "a cat sitting on a table",
    "model_name": "Salesforce/instructblip-vicuna-7b"
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
inter_str_to_cv2 Ran
ran, honoured the contract we drafted
rom1504/img2dataset/img2dataset/resizer.py
code served (permissive licence) · get_code("b1a9d0a36153d492")
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 cv2

ARGS = {
    'inter_str': 'linear'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
cv2 — 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
intersect Ran
ran; our contract draft was wrong, not the code
biubug6/Pytorch_Retinaface/utils/box_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.

is_disallowed Ran
ran, violated the contract we drafted
rom1504/img2dataset/img2dataset/downloader.py
code served (permissive licence) · get_code("82c902b8810910be")
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 bool, 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 traceback

class MockHeaders:
    """Mock HTTP headers object that supports get_all method"""
    def __init__(self, headers_dict):
        self.headers_dict = headers_dict
    
    def get_all(self, key, default=None):
        """Get all values for a header key"""
        return self.headers_dict.get(key, default)

ARGS = {
    "headers": MockHeaders({
        "X-Robots-Tag": [
            "googlebot: noindex, nofollow",
            "bingbot: noarchive",
            "noai, noimageai"
        ]
    }),
    "user_agent_token": "googlebot",
    "disallowed_header_directives": ["noindex", "nofollow", "noai"]
}
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 bool, 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
load_metdata Ran
ran; our contract draft was wrong, not the code
inquire-benchmark/INQUIRE/src/utils.py
code served (permissive licence) · get_code("ab9f7256c24c0c27")
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 ParquetMetadataProvider. 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 pandas as pd
from pathlib import Path
import tempfile
import os

# Create a temporary directory with parquet files
temp_dir = tempfile.mkdtemp()
metadata_dir = os.path.join(temp_dir, 'metadata')
os.makedirs(metadata_dir, exist_ok=True)

# Create sample parquet files
df1 = pd.DataFrame({
    'id': [0, 1, 2],
    'name': ['item_0', 'item_1', 'item_2'],
    'value': [10, 20, 30]
})

df2 = pd.DataFrame({
    'id': [3, 4, 5],
    'name': ['item_3', 'item_4', 'item_5'],
    'value': [40, 50, 60]
})

# Save as parquet files in sorted order
df1.to_parquet(os.path.join(metadata_dir, 'data_0.parquet'), index=False)
df2.to_parquet(os.path.join(metadata_dir, 'data_1.parquet'), index=False)

# Define ARGS
ARGS = {
    'index_folder': temp_dir
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
pandas — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ParquetMetadataProvider — a ParquetMetadataProvider output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a ParquetMetadataProvider 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
mrr Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
inquire-benchmark/INQUIRE/src/metrics.py
code served (permissive licence) · get_code("b5d31ecf05735aa7")
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 numpy as np

# Construct arguments for mrr function
# y_true: binary array indicating relevant items (1 = relevant, 0 = not relevant)
# y_pred: prediction scores for ranking
# count_pos: count of positive items (appears in function signature but not used in body)

y_true = np.array([0, 1, 0, 1])
y_pred = np.array([0.1, 0.9, 0.3, 0.8])
count_pos = 2

ARGS = {
    "y_true": y_true,
    "y_pred": y_pred,
    "count_pos": count_pos
}
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 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
ndcg_at_k Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
inquire-benchmark/INQUIRE/src/metrics.py
code served (permissive licence) · get_code("8b7cbc085fe9dfdf")
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

torch.manual_seed(0)

# Analyze the function:
# - y_true: needs .tolist() method, so numpy array or tensor
# - y_pred: needs .tolist() method, so numpy array or tensor
# - count_pos: integer, used in arithmetic
# - k: integer, passed to sklearn.metrics.ndcg_score

# The function expects y_true to be binary labels (0/1)
# y_pred to be scores/predictions
# count_pos is the total number of positives in the full dataset
# k is the cutoff for nDCG@k

# Create sample data:
# Let's say we have 5 items in top-k results, with 3 total positives in full dataset
y_true = np.array([1, 0, 1, 0, 0])  # binary labels for top-k items
y_pred = np.array([0.9, 0.7, 0.8, 0.3, 0.2])  # prediction scores
count_pos = 3  # total positives in full dataset (some unknown/not in top-k)
k = 5  # compute nDCG@5

ARGS = {
    "y_true": y_true,
    "y_pred": y_pred,
    "count_pos": count_pos,
    "k": k
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, sklearn, 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
point_form Ran
ran; our contract draft was wrong, not the code
biubug6/Pytorch_Retinaface/utils/box_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.

py_cpu_nms Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
biubug6/Pytorch_Retinaface/utils/nms/py_cpu_nms.py
pointer only (licence: MIT); the stored copy get_code serves, from Anymake/DRN_CVPR2020, is pointer only (licence: GPL-2.0) · get_code("5502a789c2da0e5e")
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 Anymake/DRN_CVPR2020, and GPL-2.0, as GitHub detects it on that repository, is not in the permissive set Syntology serves inline. 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; our contract draft was wrong, not the code
biubug6/Pytorch_Retinaface/widerface_evaluate/evaluation.py
code served (permissive licence) · get_code("5e211291c72bb0bd")
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 prediction file
temp_dir = tempfile.mkdtemp()
pred_filepath = os.path.join(temp_dir, "pred.txt")

# Write a sample prediction file with the expected format
with open(pred_filepath, 'w') as f:
    f.write("path/to/image.jpg\n")
    f.write("\n")
    f.write("0.1 0.2 0.3 0.4 0.9\n")
    f.write("0.5 0.6 0.7 0.8 0.95\n")
    f.write("0.15 0.25 0.35 0.45 0.85\n")

ARGS = {
    'filepath': pred_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
remove_prefix Ran
ran; our contract draft was wrong, not the code
biubug6/Pytorch_Retinaface/convert_to_onnx.py
code served (permissive licence) · get_code("10cf4647f778342c")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1) and returned 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 torch.nn as nn

torch.manual_seed(0)

# The function removes a prefix from dictionary keys
# state_dict: a dictionary with string keys and tensor values
# prefix: a string prefix to remove from keys

# Create a sample state_dict with some keys that have a common prefix
state_dict = {
    "module.layer1.weight": torch.randn(4, 3),
    "module.layer1.bias": torch.randn(4),
    "module.layer2.weight": torch.randn(5, 4),
    "layer3.weight": torch.randn(3, 5),  # This one doesn't have the prefix
}

prefix = "module."

ARGS = {
    "state_dict": state_dict,
    "prefix": prefix,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type 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
center_size Not yet run
unverified (harvested, no recorded run)
biubug6/Pytorch_Retinaface/utils/box_utils.py
code served (permissive licence) · get_code("aa1d78402838820a")
How it ran

no invocation recorded.

convert_gpt_response_to_preds Not yet run
unverified (harvested, no recorded run)
inquire-benchmark/INQUIRE/src/lmm_utils.py
code served (permissive licence) · get_code("ac0cedb8ef574cbd")
How it ran

no invocation recorded.

load_model Not yet run
unverified (harvested, no recorded run)
biubug6/Pytorch_Retinaface/convert_to_onnx.py
code served (permissive licence) · get_code("ec9dcc3bf7178b3c")
How it ran

no invocation recorded.

multiprocessing_distributor Not yet run
unverified (harvested, no recorded run)
rom1504/img2dataset/img2dataset/distributor.py
code served (permissive licence) · get_code("2a880ff5a658af6a")
How it ran

no invocation recorded.

pyspark_distributor Not yet run
unverified (harvested, no recorded run)
rom1504/img2dataset/img2dataset/distributor.py
code served (permissive licence) · get_code("9ef9ab679387fbf6")
How it ran

no invocation recorded.

Also present in other papers' code

Of 25 samples lifted from this paper's linked repositories, 25 carry a function identity and 0 do not. Of the 25 distinct functions with an identity, 13 are also present in other papers whose code we hold (13 with at least one paper dated earlier, 0 the same month, 3 later) and 12 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
13 dated earlier, 1 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
11 dated earlier, 1 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
center_size 9 other papers
8 dated earlier, 1 later
1409.1556 (2014-09)
Very Deep Convolutional Networks for Large-Scale Image Recognition
1512.02325 (2015-12)
SSD: Single Shot MultiBox Detector
and 6 more not named here
py_cpu_nms 9 other papers
8 dated earlier, 1 same year, order not determinable
1506.01497 (2015-06)
Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
1812.00155 (2018-12)
Learning RoI Transformer for Detecting Oriented Objects in Aerial Images
1905.00641 (2019-05)
RetinaFace: Single-stage Dense Face Localisation in the Wild
and 6 more not named here
conv_bn 4 other papers
4 dated earlier
1703.07332 (2017-03)
How far are we from solving the 2D & 3D Face Alignment problem? (and a dataset of 230,000 3D facial landmarks)
1905.00641 (2019-05)
RetinaFace: Single-stage Dense Face Localisation in the Wild
2203.11593 (2022-03)
Unified Negative Pair Generation toward Well-discriminative Feature Space for Face Recognition
and 1 more not named here
conv_bn_no_relu 4 other papers
4 dated earlier
1703.07332 (2017-03)
How far are we from solving the 2D & 3D Face Alignment problem? (and a dataset of 230,000 3D facial landmarks)
1905.00641 (2019-05)
RetinaFace: Single-stage Dense Face Localisation in the Wild
2203.11593 (2022-03)
Unified Negative Pair Generation toward Well-discriminative Feature Space for Face Recognition
and 1 more not named here
check_keys 3 other papers
3 dated earlier
1905.00641 (2019-05)
RetinaFace: Single-stage Dense Face Localisation in the Wild
2006.14749 (2020-06)
Deepfake Detection using Spatiotemporal Convolutional Networks
2303.17176 (2023-03)
A View From Somewhere: Human-Centric Face Representations
conv_bn1X1 3 other papers
3 dated earlier
1905.00641 (2019-05)
RetinaFace: Single-stage Dense Face Localisation in the Wild
2203.11593 (2022-03)
Unified Negative Pair Generation toward Well-discriminative Feature Space for Face Recognition
2303.17176 (2023-03)
A View From Somewhere: Human-Centric Face Representations
load_model 3 other papers
3 dated earlier
1905.00641 (2019-05)
RetinaFace: Single-stage Dense Face Localisation in the Wild
2006.14749 (2020-06)
Deepfake Detection using Spatiotemporal Convolutional Networks
2303.17176 (2023-03)
A View From Somewhere: Human-Centric Face Representations
get_gt_boxes 2 other papers
2 dated earlier
1905.00641 (2019-05)
RetinaFace: Single-stage Dense Face Localisation in the Wild
1906.06579 (2019-06)
EXTD: Extremely Tiny Face Detector via Iterative Filter Reuse
get_gt_boxes_from_txt 2 other papers
2 dated earlier
1905.00641 (2019-05)
RetinaFace: Single-stage Dense Face Localisation in the Wild
1906.06579 (2019-06)
EXTD: Extremely Tiny Face Detector via Iterative Filter Reuse
remove_prefix 2 other papers
2 dated earlier
1905.00641 (2019-05)
RetinaFace: Single-stage Dense Face Localisation in the Wild
2303.17176 (2023-03)
A View From Somewhere: Human-Centric Face Representations
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

We introduce INQUIRE, a text-to-image retrieval benchmark designed to challenge multimodal vision-language models on expert-level queries. INQUIRE includes iNaturalist 2024 (iNat24), a new dataset of five million natural world images, along with 250 expert-level retrieval queries. These queries are paired with all relevant images comprehensively labeled within iNat24, comprising 33,000 total matches. Queries span categories such as species identification, context, behavior, and appearance, emphasizing tasks that require nuanced image understanding and domain expertise. Our benchmark evaluates two core retrieval tasks: (1) INQUIRE-FULLRANK, a full dataset ranking task, and (2) INQUIRE-RERANK, a reranking task for refining top-100 retrievals. Detailed evaluation of a range of recent multimodal models demonstrates that INQUIRE poses a significant challenge, with the best models failing to achieve an mAP@50 above 50%. In addition, we show that reranking with more powerful multimodal models can enhance retrieval performance, yet there remains a significant margin for improvement. By focusing on scientifically-motivated ecological challenges, INQUIRE aims to bridge the gap between AI capabilities and the needs of real-world scientific inquiry, encouraging the development of retrieval systems that can assist with accelerating ecological and biodiversity research.

For agents

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

get_harvested_code_for_paper("2411.02537")
get_code_for_paper("2411.02537")
have("2411.02537")

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

curl https://syntology.ai/api/ran/2411.02537.json

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

[![Syntology run record](https://syntology.ai/api/ran/2411.02537.svg)](https://syntology.ai/paper/2411.02537)

Connect an agent — have() is free.