SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2103.08808 · CVPR · 2021

Track to Detect and Segment: An Online Multi-Object Tracker

Yu Wang, Liangchen Song, Junsong Yuan, Ming Yang, Jialian Wu, Jiale Cao, Suny Buffalo, Tju

arXiv · PDF · Open in the Atlas

Code that ran

4 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 1 has no contract verdict; 3 failures of our instrument; 0 no recorded run), measured 2026-11-21; 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/2103.08808.json · badge: /api/ran/2103.08808.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
JialianW/TraDeS marked official in Papers with Code
Papers with Code archive
4 ran · 0 no recorded run
FunctionStatusWhere it lives
CostVolumeLoss1D Ran
ran on a synthesized input
JialianW/TraDeS/src/lib/model/losses.py
code served (permissive licence) · get_code("73050dc5b6e34509")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, class construct-then-call prompt, model not recorded) and returned Tensor, 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 torch
import torch.nn as nn
import torch.nn.functional as F

def _gather_feat(feat, ind):
    dim = feat.size(2)
    ind = ind.unsqueeze(2).expand(ind.size(0), ind.size(1), dim)
    feat = feat.gather(1, ind)
    return feat

def _only_neg_loss(pred, gt):
    gt = torch.pow(1 - gt, 4)
    neg_loss = torch.log(1 - pred) * torch.pow(pred, 2) * gt
    return neg_loss.sum()

def _tranpose_and_gather_feat_1d(feat, ind):
    feat = feat.permute(0, 2, 1).contiguous()
    feat = _gather_feat(feat, ind)
    return feat

class CostVolumeLoss1D(nn.Module):
    """
    L_CVA: Please refer to TraDeS for details
    """

    def __init__(self):
        super(CostVolumeLoss1D, self).__init__()
        self.only_neg_loss = _only_neg_loss

    def forward(self, out, target, ind, mask, cat):
        """
        Arguments:
          out, target: B x C x H
          ind, mask: B x M
          cat (category id for peaks): B x M
        """
        out = F.softmax(out * (1 - target), dim=2)
        pos_pred_pix = _tranpose_and_gather_feat_1d(out, ind)
        pos_pred = pos_pred_pix.gather(2, cat.unsqueeze(2))
        num_pos = mask.sum()
        pos_loss = torch.log(pos_pred) * torch.pow(1 - pos_pred, 2) * mask.unsqueeze(2)
        pos_loss = pos_loss.sum()
        if num_pos == 0:
            return -pos_loss
        return -pos_loss / num_pos


INIT_ARGS = {}

ARGS = {
    'out': torch.randn(2, 8, 16),
    'target': torch.randn(2, 8, 16),
    'ind': torch.randint(0, 16, (2, 5)),
    'mask': torch.ones(2, 5),
    'cat': torch.randint(0, 8, (2, 5)),
}

if __name__ == '__main__':
    model = CostVolumeLoss1D(**INIT_ARGS)
    output = model(**ARGS)
    print("Output:", output)
    print("Output shape:", output.shape if hasattr(output, 'shape') else type(output))
Setup block drafted by
a language model, class construct-then-call prompt, model not recorded
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 — no output shape is stated: observe_signature.py observed this class with its constructor sizes canonicalised where it could, which the node does not record, so its shape is not stated as this setup block's
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: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
_gather_feat Ran
ran; our contract draft was wrong, not the code
JialianW/TraDeS/src/lib/model/losses.py
code served (permissive licence) · get_code("e1ccb8546b47ae19")
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 3, dtype family float, shape (2, 3, 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)

# feat: tensor of shape (batch_size, num_elements, dim)
# The function calls feat.size(2) to get dim, and feat.gather(1, ind)
# gather operates on dimension 1, so ind must have same shape as feat after unsqueeze
feat = torch.randn(2, 5, 4)  # batch=2, num_elements=5, dim=4

# ind: tensor of shape (batch_size, num_elements)
# The function calls ind.unsqueeze(2) and ind.expand(ind.size(0), ind.size(1), dim)
# Then uses it with gather on dimension 1
# Indices for gather must be valid (< size of dimension 1 which is 5)
ind = torch.randint(0, 5, (2, 3))  # batch=2, num_elements=3, values in [0, 5)

ARGS = {
    "feat": feat,
    "ind": ind,
}
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 3, dtype family float, shape (2, 3, 4) — shape (2, 3, 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
_only_neg_loss Ran
ran; our contract draft was wrong, not the code
JialianW/TraDeS/src/lib/model/losses.py
code served (permissive licence) · get_code("81d2bca3ba7f29f1")
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

torch.manual_seed(0)

# Create small tensors for pred and gt
# pred should be probabilities (values between 0 and 1)
# gt should be ground truth labels (values between 0 and 1)
pred = torch.rand(2, 4)  # batch size 2, dim 4
gt = torch.rand(2, 4)    # batch size 2, dim 4

ARGS = {
    "pred": pred,
    "gt": gt
}
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
_tranpose_and_gather_feat_1d Ran
ran; our contract draft was wrong, not the code
JialianW/TraDeS/src/lib/model/losses.py
code served (permissive licence) · get_code("065423f71a6e4e6d")
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 3, dtype family float, shape (2, 3, 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
import torch.nn as nn

torch.manual_seed(0)

# Construct arguments for _tranpose_and_gather_feat_1d(feat, ind)
# 
# From the function body:
# - feat: needs .permute(0, 2, 1) and .contiguous() methods, then passed to _gather_feat
#   In _gather_feat: feat.size(2), feat.gather(1, ind)
#   So feat should be shape (batch, seq_len, dim) before permute, then (batch, dim, seq_len) after
# - ind: needs .unsqueeze(2), .expand(), .size(0), .size(1) methods
#   Used in gather(1, ind), so should be shape (batch, seq_len)

batch_size = 2
seq_len = 3
dim = 4

# feat: shape (batch, seq_len, dim) = (2, 3, 4)
feat = torch.randn(batch_size, seq_len, dim)

# ind: shape (batch, seq_len) = (2, 3) with valid indices for gathering along dim 1
# After permute, feat will be (2, 4, 3), so we gather along dim 1 which has size 4
# ind values should be in range [0, 4)
ind = torch.randint(0, dim, (batch_size, seq_len))

ARGS = {
    "feat": feat,
    "ind": ind,
}
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 3, dtype family float, shape (2, 3, 3) — shape (2, 3, 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

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, 0 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 0 later) and 4 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.

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 2026-11-21.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Most online multi-object trackers perform object detection stand-alone in a neural net without any input from tracking. In this paper, we present a new online joint detection and tracking model, TraDeS (TRAck to DEtect and Segment), exploiting tracking clues to assist detection end-to-end. TraDeS infers object tracking offset by a cost volume, which is used to propagate previous object features for improving current object detection and segmentation. Effectiveness and superiority of TraDeS are shown on 4 datasets, including MOT (2D tracking), nuScenes (3D tracking), MOTS and Youtube-VIS (instance segmentation tracking). Project page: https://jialianwu.com/ projects/TraDeS.html.

For agents

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

get_harvested_code_for_paper("2103.08808")
get_code_for_paper("2103.08808")
have("2103.08808")

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

curl https://syntology.ai/api/ran/2103.08808.json

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

[![Syntology run record](https://syntology.ai/api/ran/2103.08808.svg)](https://syntology.ai/paper/2103.08808)

Connect an agent — have() is free.