SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1809.08545 · 2018

Bounding Box Regression with Uncertainty for Accurate Object Detection

arXiv · PDF · Open in the Atlas

Code that ran

2 of 4 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 2 have no contract verdict; 0 failures of our instrument; 2 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/1809.08545.json · badge: /api/ran/1809.08545.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
yihui-he/KL-Loss canonical (classifier label)
classifier basis not recorded on the link
2 ran · 1 no recorded run
yihui-he/softer-NMS canonical (classifier label)
classifier basis: census
0 ran · 1 no recorded run
FunctionStatusWhere it lives
add_VGG16_conv5_body Ran
ran on a synthesized input
yihui-he/KL-Loss/detectron/modeling/VGG16.py
code served (permissive licence) · get_code("867590b1664a9388")
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 torch
import torch.nn as nn

torch.manual_seed(0)

class MockModel:
    """Mock model that implements the methods called by add_VGG16_conv5_body"""
    
    def __init__(self):
        self.blobs = {}
    
    def Conv(self, input_blob, output_blob, in_channels, out_channels, kernel_size, pad=0, stride=1):
        """Mock Conv operation"""
        self.blobs[output_blob] = output_blob
        return output_blob
    
    def Relu(self, input_blob, output_blob):
        """Mock Relu operation"""
        self.blobs[output_blob] = output_blob
        return output_blob
    
    def MaxPool(self, input_blob, output_blob, kernel=2, pad=0, stride=2):
        """Mock MaxPool operation"""
        self.blobs[output_blob] = output_blob
        return output_blob
    
    def StopGradient(self, input_blob, output_blob):
        """Mock StopGradient operation"""
        self.blobs[output_blob] = output_blob
        return output_blob

model = MockModel()

ARGS = {
    'model': model
}
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 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
add_VGG_CNN_M_1024_conv5_body Ran
ran on a synthesized input
yihui-he/KL-Loss/detectron/modeling/VGG_CNN_M_1024.py
code served (permissive licence) · get_code("a7195f42969e79d3")
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 torch
import torch.nn as nn


class MockModel(nn.Module):
    """Mock model that implements the methods called by add_VGG_CNN_M_1024_conv5_body"""
    
    def __init__(self):
        super().__init__()
        self.blobs = {}
    
    def Conv(self, input_blob, output_blob, input_channels, output_channels, kernel_size, pad=0, stride=1):
        """Mock Conv operation"""
        self.blobs[output_blob] = f"conv({input_blob})"
        return output_blob
    
    def Relu(self, input_blob, output_blob):
        """Mock Relu operation"""
        self.blobs[output_blob] = f"relu({input_blob})"
        return output_blob
    
    def LRN(self, input_blob, output_blob, size=5, alpha=0.0005, beta=0.75, bias=2.0):
        """Mock Local Response Normalization operation"""
        self.blobs[output_blob] = f"lrn({input_blob})"
        return output_blob
    
    def MaxPool(self, input_blob, output_blob, kernel=3, pad=0, stride=2):
        """Mock MaxPool operation"""
        self.blobs[output_blob] = f"maxpool({input_blob})"
        return output_blob
    
    def StopGradient(self, input_blob, output_blob):
        """Mock StopGradient operation"""
        self.blobs[output_blob] = f"stopgrad({input_blob})"
        return output_blob


torch.manual_seed(0)

model = MockModel()

ARGS = {
    'model': model
}
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 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
generate_anchors Not yet run
unverified (harvested, no recorded run)
yihui-he/KL-Loss/detectron/modeling/generate_anchors.py
code served (permissive licence) · get_code("7b2d56f403fb4468")
How it ran

no invocation recorded.

qb Not yet run
unverified (harvested, no recorded run)
yihui-he/softer-NMS/detectron/utils/py_cpu_nms.py
code served (permissive licence) · get_code("9088711bcb0a4082")
How it ran

no invocation recorded.

Also present in other papers' code

Of 4 samples lifted from this paper's linked repositories, 4 carry a function identity and 0 do not. Of the 4 distinct functions with an identity, 3 are also present in other papers whose code we hold (3 with at least one paper dated earlier, 0 the same month, 3 later) and 1 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
generate_anchors 4 other papers
1 dated earlier, 3 later
1506.01497 (2015-06)
Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
1904.08008 (2019-04)
Clustered Object Detection in Aerial Images
1909.03625 (2019-09)
CBNet: A Novel Composite Backbone Network Architecture for Object Detection
and 1 more not named here
add_VGG16_conv5_body 3 other papers
1 dated earlier, 2 later
1803.03243 (2018-03)
Domain Adaptive Faster R-CNN for Object Detection in the Wild
1904.08008 (2019-04)
Clustered Object Detection in Aerial Images
1909.03625 (2019-09)
CBNet: A Novel Composite Backbone Network Architecture for Object Detection
add_VGG_CNN_M_1024_conv5_body 3 other papers
1 dated earlier, 2 later
1803.03243 (2018-03)
Domain Adaptive Faster R-CNN for Object Detection in the Wild
1904.08008 (2019-04)
Clustered Object Detection in Aerial Images
1909.03625 (2019-09)
CBNet: A Novel Composite Backbone Network Architecture for Object Detection

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

Large-scale object detection datasets (e.g., MS-COCO) try to define the ground truth bounding boxes as clear as possible. However, we observe that ambiguities are still introduced when labeling the bounding boxes. In this paper, we propose a novel bounding box regression loss for learning bounding box transformation and localization variance together. Our loss greatly improves the localization accuracies of various architectures with nearly no additional computation. The learned localization variance allows us to merge neighboring bounding boxes during non-maximum suppression (NMS), which further improves the localization performance. On MS-COCO, we boost the Average Precision (AP) of VGG-16 Faster R-CNN from 23.6% to 29.1%. More importantly, for ResNet-50-FPN Mask R-CNN, our method improves the AP and AP90 by 1.8% and 6.2% respectively, which significantly outperforms previous state-of-the-art bounding box refinement methods. Our code and models are available at: github.com/yihui-he/KL-Loss

For agents

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

get_harvested_code_for_paper("1809.08545")
get_code_for_paper("1809.08545")
have("1809.08545")

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

curl https://syntology.ai/api/ran/1809.08545.json

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

[![Syntology run record](https://syntology.ai/api/ran/1809.08545.svg)](https://syntology.ai/paper/1809.08545)

Connect an agent — have() is free.