SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2303.14404 · CVPR · 2023

Bridging Precision and Confidence: A Train-Time Loss for Calibrating Object Detection

Salman Khan, Muhammad Akhtar Munir, Muhammad Haris Khan, Shahbaz Fahad, Khan

arXiv · PDF · Open in the Atlas

Code that ran

12 of 18 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; 3 honoured the contract we drafted, 1 violated it and 0 have no contract verdict; 8 failures of our instrument; 6 no recorded run), measured 2026-11-22; 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/2303.14404.json · badge: /api/ran/2303.14404.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
akhtarvision/bpc_calibration marked official in Papers with Code
Papers with Code archive
11 ran · 6 no recorded run
copy not recorded community
not in the paper's declared repositories
1 ran · 0 no recorded run
FunctionStatusWhere it lives
_compute_ap_recall Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
akhtarvision/bpc_calibration/models/deformable_detr.py
code served (permissive licence) · get_code("7aa4e421ee805f42")
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 numpy as np

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

# Create small test data
scores = np.array([0.9, 0.7, 0.6, 0.5, 0.3, 0.2], dtype=np.float32)
matched = np.array([True, False, True, True, False, False], dtype=bool)
NP = 3  # Number of positives
recall_thresholds = np.linspace(0.0, 1.0, 11, endpoint=True)

ARGS = {
    'scores': scores,
    'matched': matched,
    'NP': NP,
    'recall_thresholds': recall_thresholds
}
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
_evaluate_image Ran
ran; our contract draft was wrong, not the code
akhtarvision/bpc_calibration/models/deformable_detr.py
code served (permissive licence) · get_code("3311eefa83dfc00b")
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 numpy as np
from enum import Enum
import torch

class BBFormat(Enum):
    """
    Class representing the format of a bounding box.
    """
    XYWH = 1
    XYX2Y2 = 2
    PASCAL_XML = 3
    YOLO = 4


class BoundingBox:
    """Stand-in class for bounding box objects"""
    def __init__(self, x, y, x2, y2, confidence):
        self.x = x
        self.y = y
        self.x2 = x2
        self.y2 = y2
        self.confidence = confidence
    
    def get_absolute_bounding_box(self, format=BBFormat.XYX2Y2):
        if format == BBFormat.XYX2Y2:
            return self.x, self.y, self.x2, self.y2
        raise NotImplementedError()
    
    def get_confidence(self):
        return self.confidence


torch.manual_seed(0)

# Create detection boxes (dt) - 2 detections with varying confidence
dt = [
    BoundingBox(x=10, y=10, x2=50, y2=50, confidence=torch.tensor(0.9)),
    BoundingBox(x=60, y=60, x2=100, y2=100, confidence=torch.tensor(0.7)),
]

# Create ground truth boxes (gt) - 3 ground truth boxes
gt = [
    BoundingBox(x=12, y=12, x2=48, y2=48, confidence=torch.tensor(1.0)),
    BoundingBox(x=15, y=15, x2=45, y2=45, confidence=torch.tensor(1.0)),
    BoundingBox(x=62, y=62, x2=98, y2=98, confidence=torch.tensor(1.0)),
]

# Create IoU matrix (2 detections x 3 ground truths)
ious = np.array([
    [0.8, 0.75, 0.1],
    [0.1, 0.05, 0.85],
], dtype=np.float32)

iou_threshold = 0.5
max_dets = 100
area_range = None

ARGS = {
    'dt': dt,
    'gt': gt,
    'ious': ious,
    'iou_threshold': iou_threshold,
    'max_dets': max_dets,
    'area_range': area_range,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
_get_area Ran
ran, honoured the contract we drafted
akhtarvision/bpc_calibration/models/deformable_detr.py
code served (permissive licence) · get_code("2910190caab7b66c")
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
from enum import Enum

class BBFormat(Enum):
    """
    Class representing the format of a bounding box.
    """
    XYWH = 1
    XYX2Y2 = 2
    PASCAL_XML = 3
    YOLO = 4

class BoundingBox:
    """Stand-in class for bounding box object"""
    def __init__(self):
        self.coords = [10.0, 20.0, 50.0, 80.0]  # x, y, x2, y2
    
    def get_absolute_bounding_box(self, format=BBFormat.XYX2Y2):
        """Return bounding box in requested format"""
        if format == BBFormat.XYX2Y2:
            return tuple(self.coords)
        elif format == BBFormat.XYWH:
            x, y, x2, y2 = self.coords
            return (x, y, x2 - x, y2 - y)
        else:
            return tuple(self.coords)

torch.manual_seed(0)

bbox_obj = BoundingBox()

ARGS = {
    "a": bbox_obj
}
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
_group_detections Ran
ran; our contract draft was wrong, not the code
akhtarvision/bpc_calibration/models/deformable_detr.py
code served (permissive licence) · get_code("b871b728fb3bca2a")
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 defaultdict. 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.

class Detection:
    def __init__(self, image_name, class_id):
        self._image_name = image_name
        self._class_id = class_id
    
    def get_image_name(self):
        return self._image_name
    
    def get_class_id(self):
        return self._class_id


class GroundTruth:
    def __init__(self, image_name, class_id):
        self._image_name = image_name
        self._class_id = class_id
    
    def get_image_name(self):
        return self._image_name
    
    def get_class_id(self):
        return self._class_id


dt = [
    Detection("image_001", 0),
    Detection("image_001", 1),
    Detection("image_002", 0),
]

gt = [
    GroundTruth("image_001", 0),
    GroundTruth("image_001", 0),
    GroundTruth("image_002", 1),
]

ARGS = {
    'dt': dt,
    'gt': gt,
}
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 defaultdict — a defaultdict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a defaultdict 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
_jaccard Ran
ran, honoured the contract we drafted
akhtarvision/bpc_calibration/models/deformable_detr.py
code served (permissive licence) · get_code("48bbfd6a17032332")
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
from enum import Enum

torch.manual_seed(0)

class BBFormat(Enum):
    """
    Class representing the format of a bounding box.
    """
    XYWH = 1
    XYX2Y2 = 2
    PASCAL_XML = 3
    YOLO = 4

class BoundingBox:
    """Stand-in class for a bounding box object."""
    def __init__(self, x, y, x2, y2):
        self.x = x
        self.y = y
        self.x2 = x2
        self.y2 = y2
    
    def get_absolute_bounding_box(self, format=BBFormat.XYX2Y2):
        """Return bounding box in the requested format."""
        if format == BBFormat.XYX2Y2:
            return self.x, self.y, self.x2, self.y2
        return self.x, self.y, self.x2, self.y2

# Create two bounding box objects with overlapping regions
bb_a = BoundingBox(x=10, y=20, x2=50, y2=60)
bb_b = BoundingBox(x=30, y=40, x2=70, y2=80)

ARGS = {
    'a': bb_a,
    'b': bb_b,
}
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
box_iou Ran
ran, honoured the contract we drafted
akhtarvision/bpc_calibration/models/deformable_detr.py
code served (permissive licence) · get_code("ce749f837424dd8c")
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 3, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
from torchvision.ops.boxes import box_area

torch.manual_seed(0)

# Create two sets of bounding boxes in format [x1, y1, x2, y2]
# boxes1: 2 boxes
# boxes2: 3 boxes
boxes1 = torch.tensor([
    [0.0, 0.0, 2.0, 2.0],
    [1.0, 1.0, 3.0, 3.0]
], dtype=torch.float32)

boxes2 = torch.tensor([
    [0.0, 0.0, 1.0, 1.0],
    [1.5, 1.5, 2.5, 2.5],
    [2.0, 2.0, 4.0, 4.0]
], dtype=torch.float32)

ARGS = {
    "boxes1": boxes1,
    "boxes2": boxes2
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch, torchvision — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 3, 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
dice_loss Ran
ran, violated the contract we drafted
akhtarvision/bpc_calibration/models/deformable_detr.py
code served (permissive licence) · get_code("ac8fe530cdad4d8c")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 0, dtype family float, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn

torch.manual_seed(0)

# Create sample inputs and targets tensors
# The function expects tensors that can be flattened and have sigmoid applied
# Shape: batch_size=2, spatial_dims flattened to create predictions
batch_size = 2
num_spatial = 4

# inputs: predictions before sigmoid, shape (batch_size, num_spatial)
inputs = torch.randn(batch_size, num_spatial)

# targets: binary labels matching inputs shape, values 0 or 1
targets = torch.randint(0, 2, (batch_size, num_spatial), dtype=torch.float32)

# num_boxes: number of boxes for normalization
num_boxes = 2

ARGS = {
    "inputs": inputs,
    "targets": targets,
    "num_boxes": num_boxes,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 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
generalized_box_iou Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
akhtarvision/bpc_calibration/models/deformable_detr.py
code served (permissive licence) · get_code("6097d6378e7d3046")
How it ran

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

torch.manual_seed(0)

# Create two sets of bounding boxes in [x0, y0, x1, y1] format
# boxes1: 2 boxes
boxes1 = torch.tensor([
    [10.0, 20.0, 30.0, 40.0],
    [50.0, 60.0, 70.0, 80.0]
], dtype=torch.float32)

# boxes2: 3 boxes
boxes2 = torch.tensor([
    [15.0, 25.0, 35.0, 45.0],
    [55.0, 65.0, 75.0, 85.0],
    [100.0, 110.0, 120.0, 130.0]
], dtype=torch.float32)

ARGS = {
    "boxes1": boxes1,
    "boxes2": boxes2,
    "return_iou": False
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch, torchvision — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 2, dtype family float, shape (2, 3) — shape (2, 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
get_coco_metrics Ran
ran; our contract draft was wrong, not the code
akhtarvision/bpc_calibration/models/deformable_detr.py
code served (permissive licence) · get_code("9f4690f675176640")
How it ran

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

import torch
import numpy as np
from collections import defaultdict, deque
from enum import Enum

class BBFormat(Enum):
    XYWH = 1
    XYX2Y2 = 2
    PASCAL_XML = 3
    YOLO = 4

class BoundingBox:
    """Mock BoundingBox class with required methods"""
    def __init__(self, image_name, class_id, confidence, x, y, x2, y2):
        self.image_name = image_name
        self.class_id = class_id
        self.confidence = torch.tensor(confidence)
        self.coords = (x, y, x2, y2)
    
    def get_image_name(self):
        return self.image_name
    
    def get_class_id(self):
        return self.class_id
    
    def get_confidence(self):
        return self.confidence
    
    def get_absolute_bounding_box(self, format=BBFormat.XYX2Y2):
        if format == BBFormat.XYX2Y2:
            return self.coords
        return self.coords

# Create ground truth bounding boxes
gt_bbs = [
    BoundingBox("img1", 0, 0.9, 10.0, 20.0, 50.0, 60.0),
    BoundingBox("img1", 1, 0.8, 100.0, 150.0, 200.0, 250.0),
    BoundingBox("img2", 0, 0.95, 5.0, 5.0, 40.0, 40.0),
]

# Create detected bounding boxes
det_bbs = [
    BoundingBox("img1", 0, 0.85, 12.0, 22.0, 48.0, 58.0),
    BoundingBox("img1", 1, 0.75, 102.0, 152.0, 198.0, 248.0),
    BoundingBox("img2", 0, 0.90, 6.0, 6.0, 38.0, 38.0),
    BoundingBox("img2", 1, 0.70, 60.0, 60.0, 100.0, 100.0),
]

ARGS = {
    "groundtruth_bbs": gt_bbs,
    "detected_bbs": det_bbs,
    "iou_threshold": 0.5,
    "area_range": (0, np.inf),
    "max_dets": 100
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
interpolate Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
akhtarvision/bpc_calibration/models/deformable_detr.py
code served (permissive licence) · get_code("9705264816e566d9")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 4, dtype family float, shape (2, 3, 8, 10) 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 torchvision

torch.manual_seed(0)

# Create a small input tensor with shape (batch=2, channels=3, height=4, width=5)
# This is a 4D tensor suitable for spatial interpolation
input_tensor = torch.randn(2, 3, 4, 5)

# Define the arguments for the interpolate function
ARGS = {
    'input': input_tensor,
    'size': (8, 10),  # Target spatial dimensions
    'scale_factor': None,  # Mutually exclusive with size
    'mode': 'nearest',  # Interpolation mode
    'align_corners': None  # Not used with 'nearest' mode
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch, torchvision — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 4, dtype family float, shape (2, 3, 8, 10) — shape (2, 3, 8, 10) 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
nested_tensor_from_tensor_list Ran
ran; our contract draft was wrong, not the code
akhtarvision/bpc_calibration/models/deformable_detr.py
code served (permissive licence) · get_code("07a415e6bef6b5cf")
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 NestedTensor. 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
from typing import List

torch.manual_seed(0)

# Create a list of 3D tensors with different sizes
# Each tensor has shape (C, H, W) where C=channels, H=height, W=width
tensor1 = torch.randn(3, 4, 5)  # 3 channels, 4 height, 5 width
tensor2 = torch.randn(3, 6, 7)  # 3 channels, 6 height, 7 width

tensor_list = [tensor1, tensor2]

ARGS = {
    "tensor_list": tensor_list
}
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 NestedTensor — a NestedTensor output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a NestedTensor 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
sigmoid_focal_loss Ran
ran; our contract draft was wrong, not the code
this paper's copy was not recorded; identical code first harvested from IDEA-Research/DAB-DETR
pointer only · get_code("5c0711aada67957e")
How it ran

The body of this sample is withheld: this paper's copy of the code was not recorded (its link names no repository or source path), so no licence was read for it. 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.

BBFormat Not yet run
unverified (harvested, no recorded run)
akhtarvision/bpc_calibration/models/deformable_detr.py
code served (permissive licence) · get_code("72d9dc8fd5fb3ce4")
How it ran

no invocation recorded.

BBType Not yet run
unverified (harvested, no recorded run)
akhtarvision/bpc_calibration/models/deformable_detr.py
code served (permissive licence) · get_code("c7b3326fb084b72d")
How it ran

no invocation recorded.

PostProcess_grad Not yet run
unverified (harvested, no recorded run)
akhtarvision/bpc_calibration/models/deformable_detr.py
code served (permissive licence) · get_code("29cf19c6e7a43150")
How it ran

no invocation recorded.

SetCriterion Not yet run
unverified (harvested, no recorded run)
akhtarvision/bpc_calibration/models/deformable_detr.py
code served (permissive licence) · get_code("62b1926bc1824609")
How it ran

no invocation recorded.

_compute_ious Not yet run
unverified (harvested, no recorded run)
akhtarvision/bpc_calibration/models/deformable_detr.py
code served (permissive licence) · get_code("ada48ab6a8e8dd76")
How it ran

no invocation recorded.

accuracy Not yet run
unverified (harvested, no recorded run)
akhtarvision/bpc_calibration/models/deformable_detr.py
code served (permissive licence) · get_code("9724d55577f6ca2c")
How it ran

no invocation recorded.

Also present in other papers' code

Of 18 samples lifted from this paper's linked repositories, 18 carry a function identity and 0 do not. Of the 18 distinct functions with an identity, 6 are also present in other papers whose code we hold (5 with at least one paper dated earlier, 2 the same month, 6 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
sigmoid_focal_loss
brick of this paper's assembly: helper
56 other papers
24 dated earlier, 29 later, 3 same year, order not determinable · 1 only as an assembly brick
2005.12872 (2020-05)
End-to-End Object Detection with Transformers
2010.04159 (2020-10)
Deformable DETR: Deformable Transformers for End-to-End Object Detection
2103.11731 (2021-03)
Meta-DETR: Image-Level Few-Shot Object Detection with Inter-Class Correlation Exploitation
and 53 more not named here
dice_loss
brick of this paper's assembly: helper
40 other papers
20 dated earlier, 18 later, 2 same year, order not determinable · 1 only as an assembly brick
2005.12872 (2020-05)
End-to-End Object Detection with Transformers
2010.04159 (2020-10)
Deformable DETR: Deformable Transformers for End-to-End Object Detection
2103.12115 (2021-03)
End-to-End Trainable Multi-Instance Pose Estimation with Transformers
and 37 more not named here
box_iou
brick of this paper's assembly: helper
24 other papers
7 dated earlier, 1 the same month, 15 later, 1 undated · 5 only as an assembly brick
2103.12115 (2021-03)
End-to-End Trainable Multi-Instance Pose Estimation with Transformers
2104.00969 (2021-04)
TubeR: Tubelet Transformer for Video Action Detection
2112.05375 (2021-12)
Rethinking the Two-Stage Framework for Grounded Situation Recognition
and 21 more not named here
nested_tensor_from_tensor_list
brick of this paper's assembly: helper
19 other papers
14 dated earlier, 1 the same month, 4 later · 11 only as an assembly brick
2005.12872 (2020-05)
End-to-End Object Detection with Transformers
2010.04159 (2020-10)
Deformable DETR: Deformable Transformers for End-to-End Object Detection
2103.05399 (2021-03)
QPIC: Query-Based Pairwise Human-Object Interaction Detection with Image-Wide Contextual Information
and 16 more not named here
accuracy
brick of this paper's assembly: helper
3 other papers
1 dated earlier, 2 later · 2 only as an assembly brick
2208.01582 (2022-08)
ViP3D: End-to-end Visual Trajectory Prediction via 3D Agent Queries
2304.04742 (2023-04)
Detection Transformer with Stable Matching
2403.06213 (2024-03)
V k D : Improving Knowledge Distillation using Orthogonal Projections
interpolate
brick of this paper's assembly: helper
1 other paper
1 later · 1 only as an assembly brick
2304.04742 (2023-04)
Detection Transformer with Stable Matching

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

Deep neural networks (DNNs) have enabled astounding progress in several vision-based problems. Despite showing high predictive accuracy, recently, several works have revealed that they tend to provide overconfident predictions and thus are poorly calibrated. The majority of the works addressing the miscalibration of DNNs fall under the scope of classification and consider only in-domain predictions. However, there is little to no progress in studying the calibration of DNN-based object detection models, which are central to many vision-based safety-critical applications. In this paper, inspired by the train-time calibration methods, we propose a novel auxiliary loss formulation that explicitly aims to align the class confidence of bounding boxes with the accurateness of predictions (i.e. precision). Since the original formulation of our loss depends on the counts of true positives and false positives in a minibatch, we develop a differentiable proxy of our loss that can be used during training with other application-specific loss functions. We perform extensive experiments on challenging in-domain and out-domain scenarios with six benchmark datasets including MS-COCO, Cityscapes, Sim10k, and BDD100k. Our results reveal that our train-time loss surpasses strong calibration baselines in reducing calibration error for both in and out-domain scenarios. Our source code and pre-trained models are available at https:// github.com/akhtarvision/bpc_calibration

For agents

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

get_harvested_code_for_paper("2303.14404")
get_code_for_paper("2303.14404")
have("2303.14404")

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

curl https://syntology.ai/api/ran/2303.14404.json

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

[![Syntology run record](https://syntology.ai/api/ran/2303.14404.svg)](https://syntology.ai/paper/2303.14404)

Connect an agent — have() is free.