SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2007.06866 · 2020

Alleviating Over-segmentation Errors by Detecting Action Boundaries

arXiv · PDF · Open in the Atlas

Code that ran

10 of 11 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 10 have no contract verdict; 0 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/2007.06866.json · badge: /api/ran/2007.06866.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
yiskw713/asrf canonical (classifier label)
classifier basis: census
10 ran · 1 no recorded run
FunctionStatusWhere it lives
collate_fn Ran
ran on a synthesized input
yiskw713/asrf/libs/dataset.py
code served (permissive licence) · get_code("29fa6ba076fa9873")
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.functional as F

torch.manual_seed(0)

# Create sample data matching the expected structure
sample = [
    {
        'feature': torch.randn(4, 10),  # shape (feature_dim=4, time=10)
        'label': torch.randint(0, 10, (1, 10)),  # shape (1, time=10)
        'boundary': torch.randn(1, 10),  # shape (1, time=10)
        'feature_path': '/path/to/feature1.npy'
    },
    {
        'feature': torch.randn(4, 7),  # shape (feature_dim=4, time=7)
        'label': torch.randint(0, 10, (1, 7)),  # shape (1, time=7)
        'boundary': torch.randn(1, 7),  # shape (1, time=7)
        'feature_path': '/path/to/feature2.npy'
    }
]

ARGS = {
    'sample': sample
}
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
convert_list2tuple Ran
ran on a synthesized input
yiskw713/asrf/libs/config.py
code served (permissive licence) · get_code("ac451dddcf59b322")
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
from typing import Any, Dict

torch.manual_seed(0)

# The function convert_list2tuple takes a dict and converts any list values to tuples
# It iterates through items and checks if the value is a list, then converts it

ARGS = {
    "_dict": {
        "key1": [1, 2, 3],
        "key2": "string_value",
        "key3": [4.5, 6.7],
        "key4": torch.randn(2, 4),
        "key5": {"nested": "dict"},
        "key6": [torch.randn(2, 4), torch.randn(3, 4)],
    }
}
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
get_class2id_map Ran
ran on a synthesized input
yiskw713/asrf/libs/class_id_map.py
code served (permissive licence) · get_code("dfd17ef36e26d208")
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 os
import tempfile
import shutil

# Create a temporary directory structure with mapping files
temp_dir = tempfile.mkdtemp()

# Create dataset subdirectories and mapping files
for dataset_name in ['50salads', 'breakfast', 'gtea']:
    dataset_path = os.path.join(temp_dir, dataset_name)
    os.makedirs(dataset_path, exist_ok=True)
    
    # Create mapping.txt with sample action mappings
    mapping_file = os.path.join(dataset_path, 'mapping.txt')
    with open(mapping_file, 'w') as f:
        f.write('0 action1\n')
        f.write('1 action2\n')
        f.write('2 action3\n')

ARGS = {
    'dataset': '50salads',
    'dataset_dir': temp_dir
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — 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
get_config Ran
ran on a synthesized input
yiskw713/asrf/libs/config.py
code served (permissive licence) · get_code("e9c60bbf1182ce79")
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 Config. 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 tempfile
import yaml
from pathlib import Path

# Create a temporary YAML config file
config_content = {
    'model': 'ActionSegmentRefinementNetwork',
    'n_layers': 10,
    'n_stages': 4,
    'n_features': 64,
    'n_stages_asb': 4,
    'n_stages_brb': 4,
    'ce': True,
    'ce_weight': 1.0,
    'focal': False,
    'focal_weight': 1.0,
    'tmse': False,
    'tmse_weight': 0.15,
    'gstmse': True,
    'gstmse_weight': 1.0,
    'gstmse_index': 'feature',
    'class_weight': True,
    'batch_size': 1,
    'in_channel': 2048,
    'num_workers': 4,
    'max_epoch': 50,
    'optimizer': 'Adam',
    'learning_rate': 0.0005,
    'momentum': 0.9,
    'dampening': 0.0,
    'weight_decay': 0.0001,
    'nesterov': True,
    'param_search': False,
    'iou_thresholds': [0.1, 0.25, 0.5],
    'tolerance': 5,
    'boundary_th': 0.5,
    'lambda_b': 0.1,
    'dataset': 'breakfast',
    'dataset_dir': './dataset',
    'csv_dir': './csv',
    'split': 1
}

# Write to temporary file
temp_dir = tempfile.mkdtemp()
config_path = Path(temp_dir) / 'config.yaml'
with open(config_path, 'w') as f:
    yaml.dump(config_content, f)

ARGS = {
    'config_path': str(config_path)
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
yaml — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Config — a Config output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Config 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
get_id2class_map Ran
ran on a synthesized input
yiskw713/asrf/libs/class_id_map.py
code served (permissive licence) · get_code("9b75604e2a93f75c")
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 os
import tempfile
from typing import Dict

# Create a temporary directory with mock dataset files
temp_dir = tempfile.mkdtemp()

# Create dataset subdirectories and mapping files
for dataset_name in ['50salads', 'breakfast', 'gtea']:
    dataset_path = os.path.join(temp_dir, dataset_name)
    os.makedirs(dataset_path, exist_ok=True)
    
    # Create a mock mapping.txt file with action mappings
    mapping_content = "0 SomeAction\n1 AnotherAction\n2 ThirdAction\n"
    with open(os.path.join(dataset_path, 'mapping.txt'), 'w') as f:
        f.write(mapping_content)

ARGS = {
    'dataset': '50salads',
    'dataset_dir': temp_dir
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — 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
get_n_classes Ran
ran on a synthesized input
yiskw713/asrf/libs/class_id_map.py
code served (permissive licence) · get_code("b71c0028125d46e6")
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 int, ndim 0, dtype family int, 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 os
import tempfile
from typing import Dict

# Create a temporary directory with mock dataset files
temp_dir = tempfile.mkdtemp()

# Create subdirectories for each dataset
for dataset_name in ['50salads', 'gtea', 'breakfast']:
    dataset_path = os.path.join(temp_dir, dataset_name)
    os.makedirs(dataset_path, exist_ok=True)
    
    # Create a mock mapping.txt file with action mappings
    mapping_content = "0 background\n1 action1\n2 action2\n3 action3\n"
    mapping_file = os.path.join(dataset_path, 'mapping.txt')
    with open(mapping_file, 'w') as f:
        f.write(mapping_content)

ARGS = {
    'dataset': '50salads',
    'dataset_dir': temp_dir
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type int, ndim 0, dtype family int, 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_n_samples Ran
ran on a synthesized input
yiskw713/asrf/libs/metric.py
code served (permissive licence) · get_code("e6db58c545a1ebc4")
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 numpy as np

ARGS = {
    "p_label": [1, 2],
    "p_start": [0, 10],
    "p_end": [5, 15],
    "g_label": [1, 2, 3],
    "g_start": [0, 8, 20],
    "g_end": [6, 12, 25],
    "iou_threshold": 0.5,
    "bg_class": ["background"]
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — 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
get_pos_weight Ran
ran on a synthesized input
yiskw713/asrf/libs/class_weight.py
code served (permissive licence) · get_code("be42afa58596127e")
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 os
import tempfile
import numpy as np
import pandas as pd
import torch

torch.manual_seed(0)

# Create a temporary directory structure with CSV files and boundary numpy files
temp_dir = tempfile.mkdtemp()
csv_base_dir = os.path.join(temp_dir, 'csv')

# Create dataset directories
for dataset in ['50salads', 'breakfast', 'gtea']:
    dataset_dir = os.path.join(csv_base_dir, dataset)
    os.makedirs(dataset_dir, exist_ok=True)
    
    # Create temporary boundary files
    boundary_dir = os.path.join(temp_dir, 'boundaries', dataset)
    os.makedirs(boundary_dir, exist_ok=True)
    
    # Create sample boundary numpy files
    boundary_file_1 = os.path.join(boundary_dir, 'boundary_1.npy')
    boundary_file_2 = os.path.join(boundary_dir, 'boundary_2.npy')
    
    # Save boundary arrays (binary labels: 0 and 1)
    # Make them have different distributions to test pos_weight calculation
    np.save(boundary_file_1, np.array([0, 0, 0, 1], dtype=np.int64))
    np.save(boundary_file_2, np.array([1, 1, 0, 0], dtype=np.int64))
    
    # Create train CSV files
    train_df = pd.DataFrame({
        'boundary': [boundary_file_1, boundary_file_2]
    })
    train_csv_path = os.path.join(dataset_dir, 'train1.csv')
    train_df.to_csv(train_csv_path, index=False)
    
    # Create val CSV files
    val_df = pd.DataFrame({
        'boundary': [boundary_file_1]
    })
    val_csv_path = os.path.join(dataset_dir, 'val1.csv')
    val_df.to_csv(val_csv_path, index=False)

ARGS = {
    'dataset': '50salads',
    'split': 1,
    'csv_dir': csv_base_dir,
    'mode': 'trainval',
    'norm': None
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, pandas, 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_segments Ran
ran on a synthesized input
yiskw713/asrf/libs/metric.py
code served (permissive licence) · get_code("f9168dfe61b67c0d")
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 2, 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 numpy as np

# Create a small frame-wise label array (1D numpy array of integers)
frame_wise_label = np.array([0, 0, 1, 1, 2, 0, 0, 1, 1, 1, 0], dtype=int)

# Create a mapping from integer IDs to class names
id2class_map = {
    0: 'background',
    1: 'action_a',
    2: 'action_b'
}

# Define the background class name
bg_class = 'background'

ARGS = {
    'frame_wise_label': frame_wise_label,
    'id2class_map': id2class_map,
    'bg_class': bg_class
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 2, 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
levenshtein Ran
ran on a synthesized input
yiskw713/asrf/libs/metric.py
code served (permissive licence) · get_code("3b9ef968d52cf9c1")
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 float, 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 numpy as np

torch.manual_seed(0)

# The function takes two lists of integers (pred and gt) and an optional bool (norm)
# It computes Levenshtein distance between two sequences

# Create small example lists of integers representing segments
pred_segments = [1, 2, 3, 4, 1]
gt_segments = [1, 2, 4, 1, 5]

ARGS = {
    "pred": pred_segments,
    "gt": gt_segments,
    "norm": True
}
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 float, 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
resume Not yet run
unverified (harvested, no recorded run)
yiskw713/asrf/libs/checkpoint.py
code served (permissive licence) · get_code("f2001c39f931d663")
How it ran

no invocation recorded.

Also present in other papers' code

Of 11 samples lifted from this paper's linked repositories, 11 carry a function identity and 0 do not. Of the 11 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 11 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 2026-11-22.

Repositories linked to this paper

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

Abstract

We propose an effective framework for the temporal action segmentation task, namely an Action Segment Refinement Framework (ASRF). Our model architecture consists of a long-term feature extractor and two branches: the Action Segmentation Branch (ASB) and the Boundary Regression Branch (BRB). The long-term feature extractor provides shared features for the two branches with a wide temporal receptive field. The ASB classifies video frames with action classes, while the BRB regresses the action boundary probabilities. The action boundaries predicted by the BRB refine the output from the ASB, which results in a significant performance improvement. Our contributions are three-fold: (i) We propose a framework for temporal action segmentation, the ASRF, which divides temporal action segmentation into frame-wise action classification and action boundary regression. Our framework refines frame-level hypotheses of action classes using predicted action boundaries. (ii) We propose a loss function for smoothing the transition of action probabilities, and analyze combinations of various loss functions for temporal action segmentation. (iii) Our framework outperforms state-of-the-art methods on three challenging datasets, offering an improvement of up to 13.7% in terms of segmental edit distance and up to 16.1% in terms of segmental F1 score. Our code will be publicly available soon.

For agents

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

get_harvested_code_for_paper("2007.06866")
get_code_for_paper("2007.06866")
have("2007.06866")

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

curl https://syntology.ai/api/ran/2007.06866.json

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

[![Syntology run record](https://syntology.ai/api/ran/2007.06866.svg)](https://syntology.ai/paper/2007.06866)

Connect an agent — have() is free.