</> Code with Papers by Syntology

Papers › VLM: Task-agnostic Video-Language Model Pre-training for Video Understanding

VLM: Task-agnostic Video-Language Model Pre-training for Video Understanding

20 May 2021Findings (ACL) 2021 8arXiv:2105.09996archive 2025-07-28

Hu Xu, Gargi Ghosh, Po-Yao Huang, Prahal Arora, Masoumeh Aminzadeh, Christoph Feichtenhofer, Florian Metze, Luke Zettlemoyer

We present a simplified, task-agnostic multi-modal pre-training approach that can accept either video or text input, or both for a variety of end tasks. Existing pre-training are task-specific by adopting either a single cross-modal encoder that requires both modalities, limiting their use for retrieval-style end tasks or more complex multitask learning with two unimodal encoders, limiting early cross-modal fusion. We instead introduce new pretraining masking schemes that better mix across modalities (e.g. by forcing masks for text to predict the closest video embeddings) while also maintaining separability (e.g. unimodal predictions are sometimes required, without using all the input). Experimental results show strong performance across a wider range of tasks than any previous methods, often outperforming task-specific pre-training. Code is made available at https://github.com/pytorch/fairseq/tree/main/examples/MMPT.

PaperPDFConference PDFCode

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

Code

pytorch/fairseq officialmentioned in paperpytorch 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

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

Action SegmentationLanguage ModelingLanguage ModellingRetrievalTemporal Action LocalizationVideo CaptioningVideo RetrievalVideo Understanding

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Action Segmentation COIN VLM Frame accuracy 68.4 #5 of 9 Archive leaderboard report
Temporal Action Localization CrossTask VLM Recall 46.5 #2 of 7 Archive leaderboard report
Video Captioning YouCook2 VLM BLEU-3 17.78 #5 of 14 Archive leaderboard report
Video Captioning YouCook2 VLM BLEU-4 12.27 #5 of 14 Archive leaderboard report
Video Captioning YouCook2 VLM CIDEr 1.3869 #5 of 14 Archive leaderboard report
Video Captioning YouCook2 VLM METEOR 18.22 #5 of 14 Archive leaderboard report
Video Captioning YouCook2 VLM ROUGE-L 41.51 #5 of 14 Archive leaderboard report
Video Retrieval MSR-VTT-1kA VLM text-to-video Median Rank 4 #52 of 63 Archive leaderboard report
Video Retrieval MSR-VTT-1kA VLM text-to-video R@1 28.10 #52 of 63 Archive leaderboard report
Video Retrieval MSR-VTT-1kA VLM text-to-video R@10 67.40 #52 of 63 Archive leaderboard report
Video Retrieval MSR-VTT-1kA VLM text-to-video R@5 55.50 #52 of 63 Archive leaderboard report
Video Retrieval YouCook2 VLM text-to-video Median Rank 4 #7 of 16 Archive leaderboard report
Video Retrieval YouCook2 VLM text-to-video R@1 27.05 #7 of 16 Archive leaderboard report
Video Retrieval YouCook2 VLM text-to-video R@10 69.38 #7 of 16 Archive leaderboard report
Video Retrieval YouCook2 VLM text-to-video R@5 56.88 #7 of 16 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.

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