SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2407.09751 · CVPR · 2024

TASeg: Temporal Aggregation Network for LiDAR Semantic Segmentation

Deng Cai, Wanli Ouyang, Xinge Zhu, Boxi Wu, Tong He, Yuexin Ma, Binbin Lin, Haifeng Liu, Xiaoshui Huang, Yuenan Hou, Xiaopei Wu

arXiv · PDF · Open in the Atlas

Code that ran

5 of 6 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 0 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/2407.09751.json · badge: /api/ran/2407.09751.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
LittlePey/TASeg canonical (classifier label)
classifier basis: census
5 ran · 1 no recorded run
FunctionStatusWhere it lives
dice_coefficient Ran
ran; our contract draft was wrong, not the code
LittlePey/TASeg/pcseg/loss/ell_loss.py
code served (permissive licence) · get_code("994d01fd9a5b57bf")
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 target
# pred: predictions (batch_size=2, flattened_features=4)
# target: ground truth binary labels (batch_size=2, flattened_features=4)
pred = torch.rand(2, 4)
target = torch.randint(0, 2, (2, 4), dtype=torch.float32)

# Ensure we have both positive and negative samples for the function logic
target[0, 0] = 1  # positive sample
target[0, 1] = 0  # negative sample
target[1, 0] = 1  # positive sample
target[1, 1] = 0  # negative sample

ARGS = {
    "pred": pred,
    "target": target,
    "smooth": 1,
    "exponent": 2
}
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
get_master Ran
ran; our contract draft was wrong, not the code
LittlePey/TASeg/pcseg/optim/fastai_optim.py
code served (permissive licence) · get_code("a3fa0347652ec832")
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
from torch import nn

torch.manual_seed(0)

# Create a simple model with some layers including BatchNorm
class SimpleBlock(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear = nn.Linear(4, 3)
        self.bn = nn.BatchNorm1d(3)
        self.relu = nn.ReLU()
    
    def forward(self, x):
        x = self.linear(x)
        x = self.bn(x)
        x = self.relu(x)
        return x

class SimpleBlock2(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear = nn.Linear(3, 2)
        self.bn = nn.BatchNorm1d(2)
    
    def forward(self, x):
        x = self.linear(x)
        x = self.bn(x)
        return x

# Create layer groups - each group is a Sequential of layers
group1 = nn.Sequential(SimpleBlock())
group2 = nn.Sequential(SimpleBlock2())

layer_groups = [group1, group2]

# Ensure all parameters require gradients
for group in layer_groups:
    for param in group.parameters():
        param.requires_grad = True

ARGS = {
    "layer_groups": layer_groups,
    "flat_master": False
}
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 — 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
listify Ran
ran; our contract draft was wrong, not the code
LittlePey/TASeg/pcseg/optim/fastai_optim.py
code served (permissive licence) · get_code("03cf34a5c88c655f")
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. 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.

from collections.abc import Iterable

torch_imported = False
try:
    import torch
    torch_imported = True
except ImportError:
    pass

# Test case 1: p is None, q is an integer
p_none = None
q_int = 5

# Test case 2: p is a string, q is a list
p_string = "hello"
q_list = [1, 2, 3]

# Test case 3: p is a non-iterable (int), q is None
p_int = 42
q_none = None

# Test case 4: p is already a list, q is an integer
p_list = ["a", "b"]
q_int_2 = 3

# Test case 5: p is a single-element list, q is a list with different length
p_single_list = [10]
q_list_2 = [1, 2, 3, 4]

# Main test case: p is a string, q is an integer
ARGS = {
    "p": "layer",
    "q": 3
}
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
make_one_hot Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
LittlePey/TASeg/pcseg/loss/dice_loss_v0.py
code served (permissive licence) · get_code("b91852dd7528b877")
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, 5, 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
import numpy as np

torch.manual_seed(0)

# Create a small batch of class indices
# Shape: [N=2, 1, spatial_dim=4] - representing 2 samples with 1 class index each and spatial dimension 4
batch_size = 2
spatial_dim = 4
input_tensor = torch.randint(0, 5, (batch_size, 1, spatial_dim))

# Number of classes
num_classes = 5

ARGS = {
    "input": input_tensor,
    "num_classes": num_classes,
}
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 Tensor, ndim 3, dtype family float, shape (2, 5, 4) — shape (2, 5, 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
split_bn_bias Ran
ran; our contract draft was wrong, not the code
LittlePey/TASeg/pcseg/optim/fastai_optim.py
code served (permissive licence) · get_code("422ccaf97ebcd18f")
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. 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 layer groups with a mix of batchnorm and non-batchnorm layers
# Each group is an nn.Sequential containing various layer types

group1 = nn.Sequential(
    nn.Linear(4, 8),
    nn.BatchNorm1d(8),
    nn.ReLU(),
    nn.BatchNorm1d(8)
)

group2 = nn.Sequential(
    nn.Conv2d(3, 16, kernel_size=3),
    nn.BatchNorm2d(16),
    nn.ReLU()
)

group3 = nn.Sequential(
    nn.Conv3d(2, 8, kernel_size=3),
    nn.BatchNorm3d(8),
    nn.Linear(4, 5),
    nn.SyncBatchNorm(5)
)

layer_groups = [group1, group2, group3]

ARGS = {
    "layer_groups": layer_groups
}
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 — a list 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
dice_loss Not yet run
unverified (harvested, no recorded run)
LittlePey/TASeg/pcseg/loss/ell_loss.py
code served (permissive licence) · get_code("ea9c7bcb55c7fd48")
How it ran

no invocation recorded.

Also present in other papers' code

Of 6 samples lifted from this paper's linked repositories, 6 carry a function identity and 0 do not. Of the 6 distinct functions with an identity, 6 are also present in other papers whose code we hold (6 with at least one paper dated earlier, 0 the same month, 0 later) and 0 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
make_one_hot 2 other papers
2 dated earlier
2307.05898 (2023-07)
Rectifying Noisy Labels with Sequential Prior: Multi-Scale Temporal Feature Affinity Learning for Robust Video Segmentation
2309.05573 (2023-09)
UniSeg: A Unified Multi-Modal LiDAR Segmentation Network and the OpenPCSeg Codebase
dice_coefficient 1 other paper
1 dated earlier
2309.05573 (2023-09)
UniSeg: A Unified Multi-Modal LiDAR Segmentation Network and the OpenPCSeg Codebase
dice_loss 1 other paper
1 dated earlier
2309.05573 (2023-09)
UniSeg: A Unified Multi-Modal LiDAR Segmentation Network and the OpenPCSeg Codebase
get_master 1 other paper
1 dated earlier
2309.05573 (2023-09)
UniSeg: A Unified Multi-Modal LiDAR Segmentation Network and the OpenPCSeg Codebase
listify 1 other paper
1 dated earlier
2309.05573 (2023-09)
UniSeg: A Unified Multi-Modal LiDAR Segmentation Network and the OpenPCSeg Codebase
split_bn_bias 1 other paper
1 dated earlier
2309.05573 (2023-09)
UniSeg: A Unified Multi-Modal LiDAR Segmentation Network and the OpenPCSeg Codebase

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

Training deep models for LiDAR semantic segmentation is challenging due to the inherent sparsity of point clouds. Utilizing temporal data is a natural remedy against the sparsity problem as it makes the input signal denser. However, previous multi-frame fusion algorithms fall short in utilizing sufficient temporal information due to the memory constraint, and they also ignore the informative temporal images. To fully exploit rich information hidden in long-term temporal point clouds and images, we present the Temporal Aggregation Network, termed TASeg. Specifically, we propose a Temporal LiDAR Aggregation and Distillation (TLAD) algorithm, which leverages historical priors to assign different aggregation steps for different classes. It can largely reduce memory and time overhead while achieving higher accuracy. Besides, TLAD trains a teacher injected with gt priors to distill the model, further boosting the performance. To make full use of temporal images, we design a Temporal Image Aggregation and Fusion (TIAF) module, which can greatly expand the camera FOV and enhance the present features. Temporal LiDAR points in the camera FOV are used as mediums to transform temporal image features to the present coordinate for temporal multi-modal fusion. Moreover, we develop a Static-Moving Switch Augmentation (SMSA) algorithm, which utilizes sufficient temporal information to enable objects to switch their motion states freely, thus greatly increasing static and moving training samples. Our TASeg ranks 1 st † on three challenging tracks, i.e., SemanticKITTI single-scan track, multi-scan track and nuScenes LiDAR segmentation track, strongly demonstrating the superiority of our method. Codes are available at https://github.com/LittlePey/TASeg.

For agents

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

get_harvested_code_for_paper("2407.09751")
get_code_for_paper("2407.09751")
have("2407.09751")

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

curl https://syntology.ai/api/ran/2407.09751.json

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

[![Syntology run record](https://syntology.ai/api/ran/2407.09751.svg)](https://syntology.ai/paper/2407.09751)

Connect an agent — have() is free.