SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1811.07628 · 2018

ATOM: Accurate Tracking by Overlap Maximization

arXiv · PDF · Open in the Atlas

Code that ran

8 of 9 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 1 honoured the contract we drafted, 0 violated it and 2 have no contract verdict; 5 failures of our instrument; 1 no recorded run), measured 2026-11-22; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/1811.07628.json · badge: /api/ran/1811.07628.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
xuefeng-zhu5/cdaat community (classifier label)
classifier basis not recorded on the link
8 ran · 1 no recorded run
FunctionStatusWhere it lives
all_gather Ran
ran; our contract draft was wrong, not the code
xuefeng-zhu5/cdaat/lib/utils/misc.py
code served (permissive licence) · get_code("ce98e7db9de27859")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned list, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. The code mentions cuda, and the harness aliases cuda to the CPU for every sample; whether this run needed that is 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 pickle

torch.manual_seed(0)

# The function all_gather takes one parameter: data
# It can be any picklable object since the function pickles it
# Let's create a simple picklable data object - a dictionary
data = {
    "key1": "value1",
    "key2": [1, 2, 3],
    "key3": torch.tensor([1.0, 2.0, 3.0, 4.0]),
}

ARGS = {
    "data": data,
}
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 list, ndim 1, dtype family float — a list output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
box_cxcywh_to_xyxy Ran
ran, honoured the contract we drafted
xuefeng-zhu5/cdaat/lib/utils/box_ops.py
code served (permissive licence) · get_code("e0a06ded5d4f6c3c")
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 Tensor, ndim 2, dtype family float, shape (2, 4) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# The function expects a tensor x with shape (..., 4) where the last dimension
# contains [center_x, center_y, width, height] values
# It unbinds the last dimension and converts to [x_min, y_min, x_max, y_max]

x = torch.tensor([
    [10.0, 20.0, 4.0, 6.0],
    [30.0, 40.0, 8.0, 10.0]
])

ARGS = {
    "x": x
}
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 Tensor, ndim 2, dtype family float, shape (2, 4) — shape (2, 4) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
box_xywh_to_xyxy Ran
ran; our contract draft was wrong, not the code
xuefeng-zhu5/cdaat/lib/utils/box_ops.py
code served (permissive licence) · get_code("40c655a7fe309e49")
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, 4) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# The function box_xywh_to_xyxy expects a tensor x with shape (..., 4)
# where the last dimension contains [x, y, width, height] values.
# It unbinds the last dimension into 4 components and stacks them back
# in the order [x1, y1, x1+w, y1+h] to convert to xyxy format.

x = torch.tensor([
    [10.0, 20.0, 30.0, 40.0],
    [5.0, 15.0, 25.0, 35.0]
])

ARGS = {
    "x": x
}
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 2, dtype family float, shape (2, 4) — shape (2, 4) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
box_xyxy_to_xywh Ran
ran; our contract draft was wrong, not the code
xuefeng-zhu5/cdaat/lib/utils/box_ops.py
code served (permissive licence) · get_code("b48a27604b9574e0")
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, 4) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# The function box_xyxy_to_xywh takes a tensor x, unbinds it along the last dimension
# (expecting 4 elements: x1, y1, x2, y2), and converts to xywh format.
# It returns a stacked tensor with the same shape as input.

# Create a small tensor with shape (2, 4) representing 2 bounding boxes in xyxy format
x = torch.tensor([
    [10.0, 20.0, 30.0, 40.0],
    [5.0, 15.0, 25.0, 35.0]
])

ARGS = {
    "x": x
}
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 2, dtype family float, shape (2, 4) — shape (2, 4) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
collate_fn Ran
ran; our contract draft was wrong, not the code
xuefeng-zhu5/cdaat/lib/utils/misc.py
code served (permissive licence) · get_code("8e3a5839c7c54607")
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 torch
import torch.nn as nn
from typing import List

torch.manual_seed(0)

# Create sample tensors with 3 dimensions (as expected by the function)
# Using distinct sizes: batch=2, channels=3, height=4, width=5
tensor1 = torch.randn(3, 4, 5)  # 3D tensor: (C, H, W)
tensor2 = torch.randn(3, 6, 7)  # Different H, W to test padding

# Create sample batch items (tuples of tensor and some metadata)
batch_item1 = (tensor1, "label1")
batch_item2 = (tensor2, "label2")

# Create the batch as a list of tuples
batch = [batch_item1, batch_item2]

ARGS = {
    "batch": batch
}
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 — 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_optimizer_scheduler Ran
ran on a synthesized input
xuefeng-zhu5/cdaat/lib/train/base_functions.py
code served (permissive licence) · get_code("e857124b392e7d82")
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)

# Build a minimal neural network with 'backbone' and non-backbone parameters
class SimpleNet(nn.Module):
    def __init__(self):
        super().__init__()
        # Backbone parameters
        self.backbone = nn.Linear(4, 3)
        # Non-backbone parameters
        self.head = nn.Linear(3, 2)
    
    def forward(self, x):
        x = self.backbone(x)
        x = self.head(x)
        return x

# Build config object with required attributes
class Config:
    class TRAIN:
        LR = 1e-4
        BACKBONE_MULTIPLIER = 0.1
        WEIGHT_DECAY = 1e-5
        LR_DROP_EPOCH = 10

# Create instances
net = SimpleNet()
cfg = Config()

ARGS = {
    'net': net,
    'cfg': cfg,
}
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
merge_feature_sequence Ran
ran on a synthesized input
xuefeng-zhu5/cdaat/lib/utils/merge.py
code served (permissive licence) · get_code("604be06d5f714be8")
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

torch.manual_seed(0)

# The function expects inp_list to be a list of dicts,
# where each dict has keys: 'feat', 'mask', 'pos', 'inr'
# Each dict's tensors are concatenated along specific dimensions:
# - 'feat': dim=0
# - 'mask': dim=1
# - 'pos': dim=0
# - 'inr': dim=0

# Create sample dicts with compatible tensor shapes
dict1 = {
    'feat': torch.randn(2, 4),      # shape (2, 4), will concat on dim=0
    'mask': torch.randn(3, 2),      # shape (3, 2), will concat on dim=1
    'pos': torch.randn(2, 5),       # shape (2, 5), will concat on dim=0
    'inr': torch.randn(2, 6),       # shape (2, 6), will concat on dim=0
}

dict2 = {
    'feat': torch.randn(2, 4),      # shape (2, 4), compatible with dict1 for dim=0 concat
    'mask': torch.randn(3, 2),      # shape (3, 2), compatible with dict1 for dim=1 concat
    'pos': torch.randn(2, 5),       # shape (2, 5), compatible with dict1 for dim=0 concat
    'inr': torch.randn(2, 6),       # shape (2, 6), compatible with dict1 for dim=0 concat
}

inp_list = [dict1, dict2]

ARGS = {
    'inp_list': inp_list
}
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 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
reduce_dict Ran
ran; our contract draft was wrong, not the code
xuefeng-zhu5/cdaat/lib/utils/misc.py
code served (permissive licence) · get_code("5c8f207cb5f19197")
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 torch.nn as nn

torch.manual_seed(0)

# Create a simple input_dict with tensor values
# The function expects a dict where values are tensors that can be stacked
input_dict = {
    'loss': torch.tensor(1.5),
    'accuracy': torch.tensor(0.95),
    'metric_3': torch.tensor(2.3)
}

ARGS = {
    'input_dict': input_dict,
    'average': True
}
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 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
tensor_operation Not yet run
unverified (harvested, no recorded run)
xuefeng-zhu5/cdaat/lib/utils/tensor.py
code served (permissive licence) · get_code("2c47f808b8359c8f")
How it ran

no invocation recorded.

Also present in other papers' code

Of 9 samples lifted from this paper's linked repositories, 9 carry a function identity and 0 do not. Of the 9 distinct functions with an identity, 6 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 2 the same month, 6 later) and 3 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
box_cxcywh_to_xyxy 49 other papers
49 later · 11 only as an assembly brick
2011.10881 (2020-11)
Rethinking Transformer-based Set Prediction for Object Detection
2103.12115 (2021-03)
End-to-End Trainable Multi-Instance Pose Estimation with Transformers
2103.14259 (2021-03)
OTA: Optimal Transport Assignment for Object Detection
and 46 more not named here
reduce_dict 33 other papers
1 the same month, 32 later
1811.10200 (2018-11)
IDD: A Dataset for Exploring Problems of Autonomous Navigation in Unconstrained Environments
2007.02419 (2020-07)
Attention-based Joint Detection of Object and Semantic Part
2104.00969 (2021-04)
TubeR: Tubelet Transformer for Video Action Detection
and 30 more not named here
all_gather 30 other papers
1 the same month, 29 later
1811.10200 (2018-11)
IDD: A Dataset for Exploring Problems of Autonomous Navigation in Unconstrained Environments
2007.02419 (2020-07)
Attention-based Joint Detection of Object and Semantic Part
2104.00969 (2021-04)
TubeR: Tubelet Transformer for Video Action Detection
and 27 more not named here
collate_fn 16 other papers
16 later
2104.00969 (2021-04)
TubeR: Tubelet Transformer for Video Action Detection
2104.08541 (2021-04)
TransVG: End-to-End Visual Grounding with Transformers
2111.10135 (2021-11)
Grounded Situation Recognition with Transformers
and 13 more not named here
box_xywh_to_xyxy 9 other papers
9 later
2305.15896 (2023-05)
MixFormerV2: Efficient Fully Transformer Tracking
2308.11322 (2023-08)
CiteTracker: Correlating Image and Text for Visual Tracking
2309.14611 (2023-09)
Event Stream-based Visual Object Tracking: A High-Resolution Benchmark Dataset and A Novel Baseline
and 6 more not named here
box_xyxy_to_xywh 8 other papers
8 later
2305.15896 (2023-05)
MixFormerV2: Efficient Fully Transformer Tracking
2308.11322 (2023-08)
CiteTracker: Correlating Image and Text for Visual Tracking
2309.14611 (2023-09)
Event Stream-based Visual Object Tracking: A High-Resolution Benchmark Dataset and A Novel Baseline
and 5 more not named here

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

While recent years have witnessed astonishing improvements in visual tracking robustness, the advancements in tracking accuracy have been limited. As the focus has been directed towards the development of powerful classifiers, the problem of accurate target state estimation has been largely overlooked. In fact, most trackers resort to a simple multi-scale search in order to estimate the target bounding box. We argue that this approach is fundamentally limited since target estimation is a complex task, requiring high-level knowledge about the object. We address this problem by proposing a novel tracking architecture, consisting of dedicated target estimation and classification components. High level knowledge is incorporated into the target estimation through extensive offline learning. Our target estimation component is trained to predict the overlap between the target object and an estimated bounding box. By carefully integrating target-specific information, our approach achieves previously unseen bounding box accuracy. We further introduce a classification component that is trained online to guarantee high discriminative power in the presence of distractors. Our final tracking framework sets a new state-of-the-art on five challenging benchmarks. On the new large-scale TrackingNet dataset, our tracker ATOM achieves a relative gain of 15% over the previous best approach, while running at over 30 FPS. Code and models are available at https://github.com/visionml/pytracking.

For agents

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

get_harvested_code_for_paper("1811.07628")
get_code_for_paper("1811.07628")
have("1811.07628")

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

curl https://syntology.ai/api/ran/1811.07628.json

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

[![Syntology run record](https://syntology.ai/api/ran/1811.07628.svg)](https://syntology.ai/paper/1811.07628)

Connect an agent — have() is free.