SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2303.15616 · CVPR · 2023

Fine-grained Audible Video Description

Dong Li, Zhen Qin, Yiran Zhong, Yuchao Dai, Lingpeng Kong, Yu Qiao, Shen, Xiaodong Han, Bowen He, Meng Wang, Jinxing Zhou, Aixuan Li

arXiv · PDF · Open in the Atlas

Code that ran

7 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; 0 honoured the contract we drafted, 0 violated it and 3 have no contract verdict; 4 failures of our instrument; 5 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/2303.15616.json · badge: /api/ran/2303.15616.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
OpenNLPLab/FAVDBench canonical (classifier label)
classifier basis not recorded on the link
7 ran · 5 no recorded run
FunctionStatusWhere it lives
get_image_binaries Ran
ran; our contract draft was wrong, not the code
OpenNLPLab/FAVDBench/AVLFormer/data_prepro/create_image_frame_tsv.py
code served (permissive licence) · get_code("554b85de7a4d1029")
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 base64
import cv2
import numpy as np
import tempfile
import os

# Create a temporary image file for testing
temp_dir = tempfile.mkdtemp()
temp_img_path = os.path.join(temp_dir, "test_image.jpg")

# Create a simple test image using numpy and cv2
test_image = np.random.randint(0, 255, (100, 100, 3), dtype=np.uint8)
cv2.imwrite(temp_img_path, test_image)

ARGS = {
    "list_of_paths": [temp_img_path, temp_img_path],
    "image_size": 56
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
cv2, 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
load_json_cc_fmt Ran
ran on a synthesized input
OpenNLPLab/FAVDBench/Metrics/EntityScore/score.py
code served (permissive licence) · get_code("a4bc56f41ab8b590")
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 json
import tempfile
import os

# Create a temporary JSON file with the expected format
temp_dir = tempfile.mkdtemp()
temp_file = os.path.join(temp_dir, "test_captions.json")

# Create sample data in the expected format
sample_data = [
    {"image_id": "img_001", "caption": "A cat sitting on a mat"},
    {"image_id": "img_002", "caption": "A dog running in the park"},
    {"image_id": "img_003", "caption": "A bird flying in the sky"}
]

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

# Define ARGS with the file path
ARGS = {
    "file_path": temp_file
}
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
load_tsv_to_mem Ran
ran on a synthesized input
OpenNLPLab/FAVDBench/AVLFormer/data_prepro/create_image_frame_tsv.py
code served (permissive licence) · get_code("a5c0446461afb7a8")
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 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 tempfile
import os

# Create a temporary TSV file for testing
temp_dir = tempfile.mkdtemp()
tsv_file = os.path.join(temp_dir, "test.tsv")

# Write some test data to the TSV file
with open(tsv_file, 'w') as f:
    f.write("col1\tcol2\tcol3\n")
    f.write("value1\tvalue2\tvalue3\n")
    f.write("data1\tdata2\tdata3\n")

ARGS = {
    'tsv_file': tsv_file,
    'sep': '\t'
}
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 list, ndim 2, 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
ordered_unique Ran
ran; our contract draft was wrong, not the code
OpenNLPLab/FAVDBench/AVLFormer/src/datasets/sampler_utils.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from SixCorePeach/MCAM, is pointer only (licence: NONE) · get_code("755c1ce09c3477be")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from SixCorePeach/MCAM, 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.

parse_with_config Ran
ran; our contract draft was wrong, not the code
OpenNLPLab/FAVDBench/AVLFormer/src/configs/config.py
code served (permissive licence) · get_code("9bf9ef86f23d7a76")
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 EasyDict. 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 json
import sys
import tempfile
from argparse import Namespace
from easydict import EasyDict as edict

# Create a temporary config file
config_data = {
    "learning_rate": 0.001,
    "batch_size": 32,
    "num_epochs": 10,
    "model_type": "transformer",
    "hidden_dim": 256
}

with tempfile.NamedTemporaryFile(mode='w', suffix='.json', delete=False) as f:
    json.dump(config_data, f)
    config_file = f.name

# Create parsed_args as a Namespace object (what argparse produces)
parsed_args = Namespace(
    config=config_file,
    learning_rate=0.01,  # This will override config value
    batch_size=16,       # This will override config value
    other_param="test"   # This is not in config, stays as is
)

ARGS = {
    "parsed_args": parsed_args
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
easydict — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type EasyDict — a EasyDict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a EasyDict 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
resize_and_to_binary Ran
ran; our contract draft was wrong, not the code
OpenNLPLab/FAVDBench/AVLFormer/data_prepro/create_image_frame_tsv.py
code served (permissive licence) · get_code("2d81bb25d504d2cc")
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 base64
import cv2
import numpy as np
import tempfile
import os

# Create a temporary image file for testing
def create_test_image():
    # Create a small test image
    img = np.random.randint(0, 256, (100, 120, 3), dtype=np.uint8)
    
    # Write to temporary file
    fd, temp_path = tempfile.mkstemp(suffix='.jpg')
    os.close(fd)
    cv2.imwrite(temp_path, img)
    return temp_path

temp_image_path = create_test_image()

ARGS = {
    "img_path": temp_image_path,
    "target_image_size": 64
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
cv2, numpy — 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
scale Ran
ran on a synthesized input
OpenNLPLab/FAVDBench/Metrics/AudioScore/score.py
code served (permissive licence) · get_code("61b9403f6307a5c0")
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 (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 numpy as np

def scale(x):
    x = (x / 100 - 0.5) * 2
    x = np.exp(-0.693179 * np.exp(-10 * x)) * (1 / np.exp(-0.693179 * np.exp(-10)))
    return x * 100

# Create a small numpy array with values suitable for the scaling function
x = np.array([25.0, 50.0, 75.0, 100.0])

ARGS = {
    "x": x
}
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 (4,) — shape (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
build_tensorizer Not yet run
unverified (harvested, no recorded run)
OpenNLPLab/FAVDBench/AVLFormer/src/datasets/caption_tensorizer.py
code served (permissive licence) · get_code("cb62e564cc1ccd5b")
How it ran

no invocation recorded.

get_video_duration Not yet run
unverified (harvested, no recorded run)
OpenNLPLab/FAVDBench/AVLFormer/data_prepro/extract_frames.py
code served (permissive licence) · get_code("bb782d82002f6dd0")
How it ran

no invocation recorded.

get_video_fps Not yet run
unverified (harvested, no recorded run)
OpenNLPLab/FAVDBench/AVLFormer/data_prepro/extract_frames.py
code served (permissive licence) · get_code("f012c075d57a2112")
How it ran

no invocation recorded.

noun_extracter Not yet run
unverified (harvested, no recorded run)
OpenNLPLab/FAVDBench/Metrics/EntityScore/score.py
code served (permissive licence) · get_code("ab1e5ed3de5dc955")
How it ran

no invocation recorded.

preprocess Not yet run
unverified (harvested, no recorded run)
OpenNLPLab/FAVDBench/Metrics/EntityScore/score.py
code served (permissive licence) · get_code("815aeecf457a2c0b")
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, 5 are also present in other papers whose code we hold (5 with at least one paper dated earlier, 1 the same month, 2 later) and 7 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
parse_with_config 6 other papers
3 dated earlier, 1 the same month, 2 later
Li_Align_and_Prompt_Video-and-Language_Pre-Training_With_Entity_Prompts_CVPR_2022_paper (2022 (year only))
Align and Prompt: Video-and-Language Pre-training with Entity Prompts
Lin_SwinBERT_End-to-End_Transformers_With_Sparse_Attention_for_Video_Captioning_CVPR_2022_paper (2022 (year only))
SWINBERT: End-to-End Transformers with Sparse Attention for Video Captioning
2206.03428 (2022-06)
Revealing Single Frame Bias for Video-and-Language Learning
and 3 more not named here
ordered_unique 2 other papers
1 dated earlier, 1 later
Lin_SwinBERT_End-to-End_Transformers_With_Sparse_Attention_for_Video_Captioning_CVPR_2022_paper (2022 (year only))
SWINBERT: End-to-End Transformers with Sparse Attention for Video Captioning
2507.06072 (2025-07)
MCAM: Multimodal Causal Analysis Model for Ego-Vehicle-Level Driving Video Understanding
get_image_binaries 1 other paper
1 dated earlier
Lin_SwinBERT_End-to-End_Transformers_With_Sparse_Attention_for_Video_Captioning_CVPR_2022_paper (2022 (year only))
SWINBERT: End-to-End Transformers with Sparse Attention for Video Captioning
load_tsv_to_mem 1 other paper
1 dated earlier
Lin_SwinBERT_End-to-End_Transformers_With_Sparse_Attention_for_Video_Captioning_CVPR_2022_paper (2022 (year only))
SWINBERT: End-to-End Transformers with Sparse Attention for Video Captioning
resize_and_to_binary 1 other paper
1 dated earlier
Lin_SwinBERT_End-to-End_Transformers_With_Sparse_Attention_for_Video_Captioning_CVPR_2022_paper (2022 (year only))
SWINBERT: End-to-End Transformers with Sparse Attention for Video Captioning

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 explore a new task for audio-visual-language modeling called fine-grained audible video description (FAVD). It aims to provide detailed textual descriptions for the given audible videos, including the appearance and spatial locations of each object, the actions of moving objects, and the sounds in videos. Existing visual-language modeling tasks often concentrate on visual cues in videos while undervaluing the language and audio modalities. On the other hand, FAVD requires not only audio-visual-language modeling skills but also paragraph-level language generation abilities. We construct the first fine-grained audible video description benchmark (FAVDBench) to facilitate this research. For each video clip, we first provide a one-sentence summary of the video, i.e., the caption, followed by 4-6 sentences describing the visual details and 1-2 audio-related descriptions at the end. The descriptions are provided in both English and Chinese. We create two new metrics for this task: an EntityScore to gauge the completeness of entities in the visual descriptions, and an Au-dioScore to assess the audio descriptions. As a preliminary approach to this task, we propose an audio-visuallanguage transformer that extends existing video captioning model with an additional audio branch. We combine the masked language modeling and auto-regressive language modeling losses to optimize our model so that it can produce paragraph-level descriptions. We illustrate the efficiency of our model in audio-visual-language modeling by evaluating it against the proposed benchmark using both conventional captioning metrics and our proposed metrics. We further put our benchmark to the test in video generation models, demonstrating that employing finegrained video descriptions can create more intricate videos than using captions. Code and dataset are available at https://github.com/OpenNLPLab/FAVDBench. Our online benchmark is available at www.avlbench.opennlplab.cn.

For agents

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

get_harvested_code_for_paper("2303.15616")
get_code_for_paper("2303.15616")
have("2303.15616")

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

curl https://syntology.ai/api/ran/2303.15616.json

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

[![Syntology run record](https://syntology.ai/api/ran/2303.15616.svg)](https://syntology.ai/paper/2303.15616)

Connect an agent — have() is free.