SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2405.11862 · IJCAI · 2024

SEMv3: A Fast and Robust Approach to Table Separation Line Detection

Chenyu Liu, Jiefeng Ma, Jun Du, Zhenrong Zhang, Pengfei Hu, Chunxia Qin

arXiv · PDF · Open in the Atlas

Code that ran

7 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, 1 violated it and 0 have no contract verdict; 5 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/2405.11862.json · badge: /api/ran/2405.11862.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
Chunchunwumu/SEMv3 canonical (classifier label)
classifier basis: census
7 ran · 2 no recorded run
FunctionStatusWhere it lives
align_cells_feat Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
Chunchunwumu/SEMv3/libs/model/grid_extractor.py
code served (permissive licence) · get_code("b9f7095b62154c09")
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

torch.manual_seed(0)

# Construct cells_feat: a list of tensors
# Each tensor should have shape (num_cells, feature_dim)
# where num_cells = num_rows * num_cols for that batch item
batch_size = 2
feature_dim = 4

cells_feat = []
for batch_idx in range(batch_size):
    if batch_idx == 0:
        num_rows_val = 3
        num_cols_val = 2
    else:
        num_rows_val = 2
        num_cols_val = 4
    num_cells = num_rows_val * num_cols_val
    # Create a tensor of shape (num_cells, feature_dim)
    feat_tensor = torch.randn(num_cells, feature_dim)
    cells_feat.append(feat_tensor)

num_rows = [3, 2]
num_cols = [2, 4]

ARGS = {
    "cells_feat": cells_feat,
    "num_rows": num_rows,
    "num_cols": num_cols,
}
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
all_gather Ran
ran; our contract draft was wrong, not the code
Chunchunwumu/SEMv3/cal_teds/utils.py
code served (permissive licence) · get_code("d741eb912b24880a")
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 expects a single argument 'data' which can be any picklable object
# Looking at the function body, it pickles the data, converts to tensor, and gathers across ranks
# Since world_size will be 1 (no distributed setup), it will just return [data]
# We can pass any picklable object: dict, list, tensor, string, etc.

# Create a simple picklable data object
data = {
    "key1": "value1",
    "key2": [1, 2, 3],
    "key3": torch.tensor([1.0, 2.0, 3.0, 4.0]),
    "nested": {"inner": "data"}
}

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
build_RoPE Ran
ran; our contract draft was wrong, not the code
Chunchunwumu/SEMv3/libs/model/posemb.py
code served (permissive licence) · get_code("a42189a895e7fbfc")
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 RoPE. 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)

ARGS = {
    'inchannels': 8
}
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 RoPE — a RoPE output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a RoPE 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
convert_to_roi_format Ran
ran, honoured the contract we drafted
Chunchunwumu/SEMv3/libs/model/grid_extractor.py
code served (permissive licence) · get_code("2dd31dba62793af4")
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 (5, 5) 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)

# lines_box should be a list of tensors, where each tensor represents boxes
# The function concatenates them and adds batch IDs
# Each box tensor should have shape (num_boxes, num_coords)
# Using distinct sizes: 2 boxes in first batch, 3 boxes in second batch, 4 coordinates

lines_box_1 = torch.randn(2, 4)  # 2 boxes with 4 coordinates each
lines_box_2 = torch.randn(3, 4)  # 3 boxes with 4 coordinates each

lines_box = [lines_box_1, lines_box_2]

ARGS = {
    "lines_box": lines_box
}
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 (5, 5) — shape (5, 5) 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
dice_loss Ran
ran; our contract draft was wrong, not the code
Chunchunwumu/SEMv3/libs/model/merge_head.py
code served (permissive licence) · get_code("f902a15e913442d6")
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.functional as F

torch.manual_seed(0)

# Create small tensors for pred and target
# pred: logits that will be passed through sigmoid
# target: binary labels
pred = torch.randn(2, 4, requires_grad=True)
target = torch.randint(0, 2, (2, 4), dtype=torch.float32)

# reduction parameter as string
reduction = 'mean'

ARGS = {
    'pred': pred,
    'target': target,
    'reduction': reduction,
}
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
parallel_process Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
Chunchunwumu/SEMv3/cal_teds/utils.py
code served (permissive licence) · get_code("43053af6494bf204")
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.

import torch
import numpy as np
from tqdm import tqdm
from concurrent.futures import ProcessPoolExecutor, as_completed

def simple_function(x):
    """A simple function that works with non-kwargs mode"""
    if isinstance(x, dict):
        return x['value'] * 2
    return x * 2

def kwarg_function(**kwargs):
    """A function that works with kwargs mode"""
    return kwargs.get('value', 0) * 3

# Test with simple list of integers
simple_array = [1, 2, 3, 4, 5]

# Test with list of dicts for kwargs mode
dict_array = [
    {'value': 1},
    {'value': 2},
    {'value': 3},
    {'value': 4},
    {'value': 5}
]

ARGS = {
    'array': simple_array,
    'function': simple_function,
    'n_jobs': 2,
    'use_kwargs': False,
    'front_num': 1
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch, tqdm — 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
reduce_dict Ran
ran, violated the contract we drafted
Chunchunwumu/SEMv3/cal_teds/utils.py
code served (permissive licence) · get_code("2a3280757ce5a198")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Sonnet 4.5, prompt python_fixture_v1) 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.distributed as dist

torch.manual_seed(0)

# Create a simple input dictionary with tensor values
# The function expects all values to be tensors that can be stacked
input_dict = {
    'loss': torch.tensor(2.5),
    'accuracy': torch.tensor(0.85),
    'metric_a': torch.tensor(1.2),
}

# The average parameter is a boolean
average = True

ARGS = {
    'input_dict': input_dict,
    'average': average
}
Setup block drafted by
a language model, Claude Sonnet 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 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
sigmoid_focal_loss_mutil Not yet run
unverified (harvested, no recorded run)
Chunchunwumu/SEMv3/libs/model/merge_head.py
code served (permissive licence) · get_code("847f28d7a0936194")
How it ran

no invocation recorded.

sigmoid_focal_loss_mutil_2 Not yet run
unverified (harvested, no recorded run)
Chunchunwumu/SEMv3/libs/model/merge_head.py
code served (permissive licence) · get_code("d1b9855fa7d2bfeb")
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, 3 are also present in other papers whose code we hold (3 with at least one paper dated earlier, 0 the same month, 2 later) and 6 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
reduce_dict 15 other papers
12 dated earlier, 3 later
2003.04030 (2020-03)
Learning Delicate Local Representations for Multi-Person Pose Estimation
2004.01946 (2020-04)
Weakly-Supervised Mesh-Convolutional Hand Reconstruction in the Wild
2108.02833 (2021-08)
Elaborative Rehearsal for Zero-shot Action Recognition
and 12 more not named here
all_gather 7 other papers
7 dated earlier
2004.01946 (2020-04)
Weakly-Supervised Mesh-Convolutional Hand Reconstruction in the Wild
2007.10985 (2020-07)
PointContrast: Unsupervised Pre-training for 3D Point Cloud Understanding
2104.10011 (2021-04)
Perceptual Loss for Robust Unsupervised Homography Estimation
and 4 more not named here
parallel_process 3 other papers
1 dated earlier, 2 later
2311.05565 (2023-11)
High-Performance Transformers for Table Structure Recognition Need Early Convolutions
2501.00321 (2025-01)
OCRBench v2: An Improved Benchmark for Evaluating Large Multimodal Models on Visual Text Localization and Reasoning
2607.02956 (2026-07)
MORE: A Multilingual Document Parsing Benchmark and Evaluation

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

Table structure recognition (TSR) aims to parse the inherent structure of a table from its input image. The "split-and-merge" paradigm is a pivotal approach to parse table structure, where the table separation line detection is crucial. However, challenges such as wireless and deformed tables make it demanding. In this paper, we adhere to the "split-and-merge" paradigm and propose SEMv3 (SEM: Split, Embed and Merge), a method that is both fast and robust for detecting table separation lines. During the split stage, we introduce a Keypoint Offset Regression (KOR) module, which effectively detects table separation lines by directly regressing the offset of each line relative to its keypoint proposals. Moreover, in the merge stage, we define a series of merge actions to efficiently describe the table structure based on table grids. Extensive ablation studies demonstrate that our proposed KOR module can detect table separation lines quickly and accurately. Furthermore, on public datasets (e.g. WTW, ICDAR-2019 cTDaR Historical and iFLYTAB), SEMv3 achieves state-ofthe-art (SOTA) performance. The code is available at https://github.com/Chunchunwumu/SEMv3.

For agents

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

get_harvested_code_for_paper("2405.11862")
get_code_for_paper("2405.11862")
have("2405.11862")

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

curl https://syntology.ai/api/ran/2405.11862.json

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

[![Syntology run record](https://syntology.ai/api/ran/2405.11862.svg)](https://syntology.ai/paper/2405.11862)

Connect an agent — have() is free.