Papers › VTimeLLM: Empower LLM to Grasp Video Moments

VTimeLLM: Empower LLM to Grasp Video Moments

30 Nov 2023CVPR 2024 1arXiv:2311.18445archive 2025-07-28

Bin Huang, Xin Wang, Hong Chen, Zihan Song, Wenwu Zhu

Large language models (LLMs) have shown remarkable text understanding capabilities, which have been extended as Video LLMs to handle video data for comprehending visual details. However, existing Video LLMs can only provide a coarse description of the entire video, failing to capture the precise start and end time boundary of specific events. In this paper, we solve this issue via proposing VTimeLLM, a novel Video LLM designed for fine-grained video moment understanding and reasoning with respect to time boundary. Specifically, our VTimeLLM adopts a boundary-aware three-stage training strategy, which respectively utilizes image-text pairs for feature alignment, multiple-event videos to increase temporal-boundary awareness, and high-quality video-instruction tuning to further improve temporal understanding ability as well as align with human intents. Extensive experiments demonstrate that in fine-grained time-related comprehension tasks for videos such as Temporal Video Grounding and Dense Video Captioning, VTimeLLM significantly outperforms existing Video LLMs. Besides, benefits from the fine-grained temporal understanding of the videos further enable VTimeLLM to beat existing Video LLMs in video dialogue benchmark, showing its superior cross-modal understanding and reasoning abilities.

PaperPDFConference PDFCodeCode Syntology ran

In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

For agents, Syntology's MCP tool lists every function and class Syntology harvested from this paper and whether it ran (how to connect): get_harvested_code_for_paper(arxiv_id="2311.18445")

Code

Syntology Ran 5 of 11 code samples harvested from 1 repository linked to this paper; 6 have no recorded run. Of those that ran: 1 ran · our draft was wrong; 1 ran · fixture could not drive it; 3 ran with no contract checked.

By repository: official repository: 11 samples from 1 repository, 5 ran. The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.

huangb23/vtimellm officialmentioned in papermentioned on GitHubpytorchNOASSERTION report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

11 samples harvested; 5 ran; 0 honoured the contract we drafted; 6 have no recorded run. Read from Syntology's graph 2026-09-24; that is when this build read the record, not when the samples ran.

1ran · our draft was wrong
1ran · fixture could not drive it
3ran
6unverified

Licence: 11 of the 11 samples are pointer only, meaning Syntology does not serve that copy's text. This page shows no code text for any sample; each one links to its file in the repository.

Harvested from huangb23/vtimellm. “Ran” means the sample executed on a synthesized input. It does not mean the output is correct, and nothing here reproduces the paper's results. “Honoured” and “violated” refer to a contract Syntology drafted from the code itself; “our draft was wrong” and “fixture could not drive it” are failures of Syntology's instrument, not of the code.

Each sample ends with its code_sha256, Syntology's identity for that exact code. An agent fetches the stored sample with Syntology's MCP tool get_code(code_sha256="…") (how to connect); click an identity to copy that call.

Repository labels, per sample. official repository: The archive marks this repository official for the paper. named in the paper: The archive records that the paper mentions this repository; it is not marked official. community (archive-listed): In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper. found in paper text by Syntology: Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted. community: Not in the archive's code links for this paper; a community repository Syntology harvested. Samples from a repository marked official are listed first. Licence labels name the repository's licence as recorded at harvest. “Pointer only” means Syntology does not serve that copy's text, for one of four reasons: no licence file was found; the licence was not identified; the licence is recorded as permissive but that copy's record is not marked cleared; or the licence is outside the permissive list Syntology serves text under (MIT, Apache-2.0, BSD and similar). Some licences outside that list permit redistribution, such as WTFPL, and GPL-3.0 under its conditions; they are simply not on the list. Hover a licence label for the reason. File links open the file on GitHub at the default branch, which may have changed since the harvest.

apply_rotary_pos_emb huangb23/vtimellm/vtimellm/model/chatglm/modeling_chatglm.py official repository ran · fixture could not drive it no licence file found · pointer only · 5cfeaee89a75c5ff · report
grounding_metrics huangb23/vtimellm/vtimellm/eval/metric.py official repository ran licence not identified · pointer only · f6eec537f45dfb96 · report
load_image_from_base64 huangb23/vtimellm/vtimellm/mm_utils.py official repository ran no licence file found · pointer only · c3ee9d07c900dd55 · report
process_images huangb23/vtimellm/vtimellm/mm_utils.py official repository ran no licence file found · pointer only · 1df990c375318896 · report
split_tensor_along_last_dim huangb23/vtimellm/vtimellm/model/chatglm/modeling_chatglm.py official repository ran · our draft was wrong no licence file found · pointer only · d7e378ee7899fbde · report
default_init huangb23/vtimellm/vtimellm/model/chatglm/modeling_chatglm.py official repository unverified no licence file found · pointer only · 237019e36b9481d9 · report
load_lora huangb23/vtimellm/vtimellm/model/builder.py official repository unverified no licence file found · pointer only · d6b600df5a8550bc · report
load_pretrained_model huangb23/vtimellm/vtimellm/model/builder.py official repository unverified licence not identified · pointer only · 3d00e5b35fde051d · report
merge_similar_sentences huangb23/vtimellm/vtimellm/eval/metric.py official repository unverified no licence file found · pointer only · 362a9306f09e0d58 · report
pretty_print_semaphore huangb23/vtimellm/vtimellm/utils.py official repository unverified no licence file found · pointer only · 37899f22fb191b37 · report
violates_moderation huangb23/vtimellm/vtimellm/utils.py official repository unverified no licence file found · pointer only · f9939a84b9a65279 · report

Tasks

Dense Video CaptioningTemporal Relation ExtractionVCGBench-DiverseVideo CaptioningVideo GroundingVideo Question AnsweringVideo-based Generative Performance BenchmarkingVideo-based Generative Performance Benchmarking (Consistency)Video-based Generative Performance Benchmarking (Contextual Understanding)Video-based Generative Performance Benchmarking (Correctness of Information)Video-based Generative Performance Benchmarking (Detail Orientation))Video-based Generative Performance Benchmarking (Temporal Understanding)

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Dense Video Captioning ActivityNet Captions VTimeLLM CIDEr 27.6 #12 of 12 Archive leaderboard report
Dense Video Captioning ActivityNet Captions VTimeLLM SODA 5.8 #12 of 12 Archive leaderboard report
Temporal Relation Extraction Vinoground VTimeLLM Group Score 5.2 #21 of 24 Archive leaderboard report
Temporal Relation Extraction Vinoground VTimeLLM Text Score 19.4 #21 of 24 Archive leaderboard report
Temporal Relation Extraction Vinoground VTimeLLM Video Score 27 #21 of 24 Archive leaderboard report
VCGBench-Diverse VideoInstruct VTimeLLM Consistency 2.35 #5 of 6 Archive leaderboard report
VCGBench-Diverse VideoInstruct VTimeLLM Contextual Understanding 2.48 #5 of 6 Archive leaderboard report
VCGBench-Diverse VideoInstruct VTimeLLM Correctness of Information 2.16 #5 of 6 Archive leaderboard report
VCGBench-Diverse VideoInstruct VTimeLLM Dense Captioning 1.13 #5 of 6 Archive leaderboard report
VCGBench-Diverse VideoInstruct VTimeLLM Detail Orientation 2.41 #5 of 6 Archive leaderboard report
VCGBench-Diverse VideoInstruct VTimeLLM Reasoning 3.45 #5 of 6 Archive leaderboard report
VCGBench-Diverse VideoInstruct VTimeLLM Spatial Understanding 2.29 #5 of 6 Archive leaderboard report
VCGBench-Diverse VideoInstruct VTimeLLM Temporal Understanding 1.46 #5 of 6 Archive leaderboard report
VCGBench-Diverse VideoInstruct VTimeLLM mean 2.17 #5 of 6 Archive leaderboard report
Video Question Answering OVBench VTimeLLM (7B) AVG 33.1 #11 of 16 Archive leaderboard report
Video-based Generative Performance Benchmarking VideoInstruct VTimeLLM Consistency 2.47 #17 of 23 Archive leaderboard report
Video-based Generative Performance Benchmarking VideoInstruct VTimeLLM Contextual Understanding 3.40 #17 of 23 Archive leaderboard report
Video-based Generative Performance Benchmarking VideoInstruct VTimeLLM Correctness of Information 2.78 #17 of 23 Archive leaderboard report
Video-based Generative Performance Benchmarking VideoInstruct VTimeLLM Detail Orientation 3.10 #17 of 23 Archive leaderboard report
Video-based Generative Performance Benchmarking VideoInstruct VTimeLLM Temporal Understanding 2.49 #17 of 23 Archive leaderboard report
Video-based Generative Performance Benchmarking VideoInstruct VTimeLLM mean 2.85 #17 of 23 Archive leaderboard report
Video-based Generative Performance Benchmarking (Consistency) VideoInstruct VTimeLLM gpt-score 2.47 #11 of 18 Archive leaderboard report
Video-based Generative Performance Benchmarking (Contextual Understanding) VideoInstruct VTimeLLM gpt-score 3.40 #11 of 18 Archive leaderboard report
Video-based Generative Performance Benchmarking (Correctness of Information) VideoInstruct VTimeLLM gpt-score 2.78 #11 of 18 Archive leaderboard report
Video-based Generative Performance Benchmarking (Detail Orientation)) VideoInstruct VTimeLLM gpt-score 3.10 #4 of 18 Archive leaderboard report
Video-based Generative Performance Benchmarking (Temporal Understanding) VideoInstruct VTimeLLM gpt-score 2.49 #10 of 18 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

ALIGN

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections