SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2303.06378 · 2023

Learning Grounded Vision-Language Representation for Versatile Understanding in Untrimmed Videos

arXiv · PDF · Open in the Atlas

Code that ran

9 of 12 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 6 have no contract verdict; 2 failures of our instrument; 3 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/2303.06378.json · badge: /api/ran/2303.06378.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
zjr2000/gvl canonical (classifier label)
classifier basis not recorded on the link
9 ran · 3 no recorded run
FunctionStatusWhere it lives
calculate_avg_proposal_num Ran
ran on a synthesized input
zjr2000/gvl/eval_utils.py
code served (permissive licence) · get_code("53e64888999b6cb7")
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 float64, 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 numpy as np
import json
import tempfile
import os

# Create a temporary JSON file with the expected structure
temp_dir = tempfile.mkdtemp()
json_path = os.path.join(temp_dir, "test_data.json")

# Create sample data matching the function's expected structure
data = {
    "results": {
        "item_1": [1, 2, 3],
        "item_2": [1, 2],
        "item_3": [1, 2, 3, 4, 5],
        "item_4": [1]
    }
}

# Write the JSON file
with open(json_path, 'w') as f:
    json.dump(data, f)

ARGS = {
    "json_path": json_path
}
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 float64, ndim 0, dtype family float — no output shape is stated: observe_signature.py recorded float ndim 0, which disagrees with the fixture run
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_iou Ran
ran on a synthesized input
zjr2000/gvl/densevid_eval3/eval_grounding.py
code served (permissive licence) · get_code("fe5aa6f8364bdae4")
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 get_iou takes two segments: pred and gt
# Each segment is a tuple/list of (start, end) coordinates
# The function performs arithmetic operations on these values

# Create valid arguments for pred and gt
pred = (2.5, 8.3)  # (start_pred, end_pred)
gt = (3.0, 7.5)    # (start, end)

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
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
get_miou Ran
ran on a synthesized input
zjr2000/gvl/densevid_eval3/eval_grounding.py
code served (permissive licence) · get_code("931c2ee1f4c9cd81")
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

# Construct predictions dictionary
predictions = {
    'video_1': [{'timestamp': (2.0, 5.0)}],
    'video_2': [{'timestamp': (1.5, 4.5)}],
}

# Construct groundtruths dictionary
groundtruths = {
    'video_1': {'timestamp': (2.0, 6.0)},
    'video_2': {'timestamp': (1.0, 5.0)},
    'video_3': {'timestamp': (0.5, 3.5)},
}

ARGS = {
    'predictions': predictions,
    'groundtruths': groundtruths,
}
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 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
get_recall_at_k Ran
ran on a synthesized input
zjr2000/gvl/densevid_eval3/eval_grounding.py
code served (permissive licence) · get_code("c6a05eda2a29c456")
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 float32, 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 numpy as np

# Construct valid arguments for get_recall_at_k
predictions = {
    'video_1': [
        {'timestamp': (5, 15)},
        {'timestamp': (20, 30)},
        {'timestamp': (40, 50)},
    ],
    'video_2': [
        {'timestamp': (10, 25)},
        {'timestamp': (35, 45)},
    ],
}

groundtruths = {
    'video_1': {'timestamp': (5, 16)},
    'video_2': {'timestamp': (10, 26)},
    'video_3': {'timestamp': (15, 35)},
}

ARGS = {
    'predictions': predictions,
    'groundtruths': groundtruths,
    'iou_threshold': 0.5,
    'max_proposal_num': 5,
}
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 float32, 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
interpolated_prec_rec Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
zjr2000/gvl/densevid_eval3/eval_tal_utils.py
pointer only (licence: MIT); the stored copy get_code serves, from Ontheway361/BSN_pytorch, is pointer only (licence: NONE) · get_code("81aab88321de62c2")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from Ontheway361/BSN_pytorch, and GitHub detected no licence in that repository when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

iou Ran
ran on a synthesized input
zjr2000/gvl/video_dataset.py
code served (permissive licence) · get_code("004b519990491671")
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 ndarray, ndim 2, dtype family float, shape (2, 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 numpy as np

# interval_1 and interval_2 should be array-like objects with shape (N, 2) and (M, 2)
# where each row is [start, end] of an interval
# The function computes IoU between all pairs of intervals

interval_1 = np.array([
    [0.0, 5.0],
    [3.0, 8.0]
])

interval_2 = np.array([
    [1.0, 6.0],
    [4.0, 7.0],
    [9.0, 10.0]
])

ARGS = {
    "interval_1": interval_1,
    "interval_2": interval_2
}
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 ndarray, ndim 2, dtype family float, shape (2, 3) — shape (2, 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
remove_nonascii Ran
ran; our contract draft was wrong, not the code
zjr2000/gvl/densevid_eval3/evaluate2018.py
code served (permissive licence) · get_code("a45f62024eb3d2c5")
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 str. 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 remove_nonascii takes a text parameter (string)
# and filters out non-ASCII characters, replacing them with spaces
# It iterates through each character and checks if ord(i) < 128

ARGS = {
    "text": "Hello World! Café naïve 你好 Москва"
}
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 str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str 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
segment_iou Ran
ran, honoured the contract we drafted
zjr2000/gvl/densevid_eval3/eval_tal_utils.py
code served (permissive licence) · get_code("930f6b6986cf84bf")
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 ndarray, ndim 1, dtype family float, shape (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 numpy as np

# Set random seed for reproducibility
np.random.seed(0)

# Define target segment as 1d array with [start, end] times
target_segment = np.array([2.0, 8.0])

# Define candidate segments as 2d array with N x [start, end] times
# Using 5 candidate segments with distinct dimensions
candidate_segments = np.array([
    [1.0, 5.0],
    [3.0, 7.0],
    [6.0, 10.0],
    [0.5, 3.5],
    [7.5, 9.0]
])

ARGS = {
    "target_segment": target_segment,
    "candidate_segments": candidate_segments
}
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 ndarray, ndim 1, dtype family float, shape (5,) — shape (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
sort_events Ran
ran on a synthesized input
zjr2000/gvl/video_dataset.py
code served (permissive licence) · get_code("956db474f5613b75")
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 numpy as np

torch.manual_seed(0)

# Construct proposal_data with the structure expected by sort_events
# The function expects a dict where:
# - keys are video IDs
# - values are lists of proposal dictionaries
# - each proposal dict has 'score' and 'segment' keys

proposal_data = {
    'video_1': [
        {'score': 0.5, 'segment': [2.0, 5.0]},
        {'score': 0.3, 'segment': [1.0, 3.0]},
        {'score': -0.1, 'segment': [0.5, 1.5]},  # Will be filtered out
        {'score': 0.8, 'segment': [3.0, 6.0]},
    ],
    'video_2': [
        {'score': 0.9, 'segment': [5.0, 8.0]},
        {'score': 0.2, 'segment': [1.0, 2.5]},
        {'score': -0.5, 'segment': [0.0, 1.0]},  # Will be filtered out
    ],
}

ARGS = {
    'proposal_data': proposal_data
}
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 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
collate_fn Not yet run
unverified (harvested, no recorded run)
zjr2000/gvl/video_dataset.py
code served (permissive licence) · get_code("cc29ecc7589efdff")
How it ran

no invocation recorded.

collate_fn Not yet run
unverified (harvested, no recorded run)
zjr2000/gvl/video_dataset_with_data_aug.py
code served (permissive licence) · get_code("85e25c640a683e5a")
How it ran

no invocation recorded.

get_blocked_videos Not yet run
unverified (harvested, no recorded run)
zjr2000/gvl/densevid_eval3/eval_tal_utils.py
code served (permissive licence) · get_code("0f585e754688f3eb")
How it ran

no invocation recorded.

Also present in other papers' code

Of 12 samples lifted from this paper's linked repositories, 12 carry a function identity and 0 do not. Of the 12 distinct functions with an identity, 4 are also present in other papers whose code we hold (4 with at least one paper dated earlier, 0 the same month, 3 later) and 8 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
interpolated_prec_rec 13 other papers
9 dated earlier, 4 later
1804.05113 (2018-04)
Multilevel Language and Vision Integration for Text-to-Clip Retrieval
1806.02964 (2018-06)
BSN: Boundary Sensitive Network for Temporal Action Proposal Generation
1907.09702 (2019-07)
BMN: Boundary-Matching Network for Temporal Action Proposal Generation
and 10 more not named here
segment_iou 13 other papers
9 dated earlier, 4 later
1804.05113 (2018-04)
Multilevel Language and Vision Integration for Text-to-Clip Retrieval
1806.02964 (2018-06)
BSN: Boundary Sensitive Network for Temporal Action Proposal Generation
1907.09702 (2019-07)
BMN: Boundary-Matching Network for Temporal Action Proposal Generation
and 10 more not named here
remove_nonascii 4 other papers
1 dated earlier, 3 later
2001.09099 (2020-01)
TVR: A Large-Scale Dataset for Video-Subtitle Moment Retrieval
2403.17935 (2024-03)
OmniViD: A Generative Framework for Universal Video Understanding
2405.13382 (2024-05)
VTG-LLM: Integrating Timestamp Knowledge into Video LLMs for Enhanced Video Temporal Grounding
and 1 more not named here
calculate_avg_proposal_num 1 other paper
1 dated earlier
2108.07781 (2021-08)
End-to-End Dense Video Captioning with Parallel Decoding

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

Joint video-language learning has received increasing attention in recent years. However, existing works mainly focus on single or multiple trimmed video clips (events), which makes human-annotated event boundaries necessary during inference. To break away from the ties, we propose a grounded vision-language learning framework for untrimmed videos, which automatically detects informative events and effectively excavates the alignments between multi-sentence descriptions and corresponding event segments. Instead of coarse-level video-language alignments, we present two dual pretext tasks to encourage fine-grained segment-level alignments, i.e., text-to-event grounding (TEG) and event-to-text generation (ETG). TEG learns to adaptively ground the possible event proposals given a set of sentences by estimating the cross-modal distance in a joint semantic space. Meanwhile, ETG aims to reconstruct (generate) the matched texts given event proposals, encouraging the event representation to retain meaningful semantic information. To encourage accurate label assignment between the event set and the text set, we propose a novel semantic-aware cost to mitigate the sub-optimal matching results caused by ambiguous boundary annotations. Our framework is easily extensible to tasks covering visually-grounded language understanding and generation. We achieve state-of-the-art dense video captioning performance on ActivityNet Captions, YouCook2 and YouMakeup, and competitive performance on several other language generation and understanding tasks. Our method also achieved 1st place in both the MTVG and MDVC tasks of the PIC 4th Challenge. Our code is publicly available at https://github.com/zjr2000/GVL.

For agents

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

get_harvested_code_for_paper("2303.06378")
get_code_for_paper("2303.06378")
have("2303.06378")

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

curl https://syntology.ai/api/ran/2303.06378.json

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

[![Syntology run record](https://syntology.ai/api/ran/2303.06378.svg)](https://syntology.ai/paper/2303.06378)

Connect an agent — have() is free.