</> Code with Papers by Syntology

Papers › Unmasked Teacher: Towards Training-Efficient Video Foundation Models

Unmasked Teacher: Towards Training-Efficient Video Foundation Models

28 Mar 2023ICCV 2023 1arXiv:2303.16058archive 2025-07-28

Kunchang Li, Yali Wang, Yizhuo Li, Yi Wang, Yinan He, LiMin Wang, Yu Qiao

Video Foundation Models (VFMs) have received limited exploration due to high computational costs and data scarcity. Previous VFMs rely on Image Foundation Models (IFMs), which face challenges in transferring to the video domain. Although VideoMAE has trained a robust ViT from limited data, its low-level reconstruction poses convergence difficulties and conflicts with high-level cross-modal alignment. This paper proposes a training-efficient method for temporal-sensitive VFMs that integrates the benefits of existing methods. To increase data efficiency, we mask out most of the low-semantics video tokens, but selectively align the unmasked tokens with IFM, which serves as the UnMasked Teacher (UMT). By providing semantic guidance, our method enables faster convergence and multimodal friendliness. With a progressive pre-training framework, our model can handle various tasks including scene-related, temporal-related, and complex video-language understanding. Using only public sources for pre-training in 6 days on 32 A100 GPUs, our scratch-built ViT-L/16 achieves state-of-the-art performances on various video tasks. The code and models will be released at https://github.com/OpenGVLab/unmasked_teacher.

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="2303.16058")

Code

Syntology Ran 3 of 8 code samples harvested from 1 repository linked to this paper; 5 have no recorded run. Of those that ran: 1 ran · honoured contract; 1 ran · fixture could not drive it; 1 ran with no contract checked.

By repository: official repository: 8 samples from 1 repository, 3 ran. The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.

opengvlab/unmasked_teacher officialmentioned in papermentioned on GitHubpytorchMIT 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

8 samples harvested; 3 ran; 1 honoured the contract we drafted; 5 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 · honoured contract
1ran · fixture could not drive it
1ran
5unverified

Licence: 0 of the 8 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 opengvlab/unmasked_teacher. “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.

get_sinusoid_encoding_table opengvlab/unmasked_teacher/single_modality/action_detection/modeling_finetune.py official repository ran · honoured contract fingerprinted MIT (permissive) · da651e3979a18f84 · report
interpolate_temporal_pos_embed opengvlab/unmasked_teacher/multi_modality/models/utils.py official repository ran · fixture could not drive it MIT (permissive) · f64c56c357505158 · report
load_temp_embed_with_mismatch opengvlab/unmasked_teacher/multi_modality/models/utils.py official repository ran MIT (permissive) · 5349fd1148712a53 · report
RandomMaskingGenerator opengvlab/unmasked_teacher/multi_modality/models/mask.py official repository unverified MIT (permissive) · 604766e7f8e055b3 · report
TubeMaskingGenerator opengvlab/unmasked_teacher/multi_modality/models/mask.py official repository unverified MIT (permissive) · 4858eefb54696701 · report
get_loss_scale_for_deepspeed opengvlab/unmasked_teacher/single_modality/action_detection/engine_for_finetuning.py official repository unverified MIT (permissive) · 29416043c7035e7c · report
get_sim opengvlab/unmasked_teacher/multi_modality/models/criterions.py official repository unverified MIT (permissive) · 431de8c91d7b558f · report
interpolate_pos_embed opengvlab/unmasked_teacher/multi_modality/models/utils.py official repository unverified MIT (permissive) · dbd049b02a63c3a1 · report

Tasks

Action ClassificationAction RecognitionSpatio-Temporal Action LocalizationVideo Question AnsweringVideo RetrievalVisual Question Answering (VQA)Zero-Shot Video Retrievalcross-modal alignment

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Action Classification Kinetics-400 Unmasked Teacher (ViT-L) Acc@1 90.6 #8 of 207 Archive leaderboard report
Action Classification Kinetics-400 Unmasked Teacher (ViT-L) Acc@5 98.7 #8 of 207 Archive leaderboard report
Action Classification Kinetics-400 Unmasked Teacher (ViT-L) FLOPs (G) x views 1434×3×4 #8 of 207 Archive leaderboard report
Action Classification Kinetics-400 Unmasked Teacher (ViT-L) Parameters (M) 304 #8 of 207 Archive leaderboard report
Action Classification Kinetics-400 UMT-L (ViT-L/16) Acc@1 90.6 #9 of 207 Archive leaderboard report
Action Classification Kinetics-400 UMT-L (ViT-L/16) Acc@5 98.7 #9 of 207 Archive leaderboard report
Action Classification Kinetics-600 UMT-L (ViT-L/16) Top-1 Accuracy 90.5 #8 of 65 Archive leaderboard report
Action Classification Kinetics-600 UMT-L (ViT-L/16) Top-5 Accuracy 98.8 #8 of 65 Archive leaderboard report
Action Classification Kinetics-700 UMT-L (ViT-L/16) Top-1 Accuracy 83.6 #5 of 36 Archive leaderboard report
Action Classification Kinetics-700 UMT-L (ViT-L/16) Top-5 Accuracy 96.7 #5 of 36 Archive leaderboard report
Action Classification MiT UMT-L (ViT-L/16) Top 1 Accuracy 48.7 #3 of 29 Archive leaderboard report
Action Classification MiT UMT-L (ViT-L/16) Top 5 Accuracy 78.2 #3 of 29 Archive leaderboard report
Action Recognition AVA v2.2 UMT-L (ViT-L/16) mAP 39.8 #9 of 38 Archive leaderboard report
Video Question Answering ActivityNet-QA UMT-L (ViT-L/16) Accuracy 47.9 #10 of 36 Archive leaderboard report
Video Retrieval ActivityNet UMT-L (ViT-L/16) text-to-video R@1 66.8 #6 of 31 Archive leaderboard report
Video Retrieval ActivityNet UMT-L (ViT-L/16) text-to-video R@10 94.9 #6 of 31 Archive leaderboard report
Video Retrieval ActivityNet UMT-L (ViT-L/16) text-to-video R@5 89.1 #6 of 31 Archive leaderboard report
Video Retrieval ActivityNet UMT-L (ViT-L/16) video-to-text R@1 64.4 #6 of 31 Archive leaderboard report
Video Retrieval ActivityNet UMT-L (ViT-L/16) video-to-text R@10 94.8 #6 of 31 Archive leaderboard report
Video Retrieval ActivityNet UMT-L (ViT-L/16) video-to-text R@5 89.1 #6 of 31 Archive leaderboard report
Video Retrieval DiDeMo UMT-L (ViT-L/16) text-to-video R@1 70.4 #5 of 40 Archive leaderboard report
Video Retrieval DiDeMo UMT-L (ViT-L/16) text-to-video R@10 93.5 #5 of 40 Archive leaderboard report
Video Retrieval DiDeMo UMT-L (ViT-L/16) text-to-video R@5 90.1 #5 of 40 Archive leaderboard report
Video Retrieval DiDeMo UMT-L (ViT-L/16) video-to-text R@1 65.7 #5 of 40 Archive leaderboard report
Video Retrieval DiDeMo UMT-L (ViT-L/16) video-to-text R@10 93.3 #5 of 40 Archive leaderboard report
Video Retrieval DiDeMo UMT-L (ViT-L/16) video-to-text R@5 89.6 #5 of 40 Archive leaderboard report
Video Retrieval LSMDC UMT-L (ViT-L/16) text-to-video R@1 43.0 #3 of 38 Archive leaderboard report
Video Retrieval LSMDC UMT-L (ViT-L/16) text-to-video R@10 73.0 #3 of 38 Archive leaderboard report
Video Retrieval LSMDC UMT-L (ViT-L/16) text-to-video R@5 65.5 #3 of 38 Archive leaderboard report
Video Retrieval LSMDC UMT-L (ViT-L/16) video-to-text R@1 41.4 #3 of 38 Archive leaderboard report
Video Retrieval LSMDC UMT-L (ViT-L/16) video-to-text R@10 71.5 #3 of 38 Archive leaderboard report
Video Retrieval LSMDC UMT-L (ViT-L/16) video-to-text R@5 64.3 #3 of 38 Archive leaderboard report
Video Retrieval MSR-VTT UMT-L (ViT-L/16) text-to-video R@1 58.8 #5 of 40 Archive leaderboard report
Video Retrieval MSR-VTT UMT-L (ViT-L/16) text-to-video R@10 87.1 #5 of 40 Archive leaderboard report
Video Retrieval MSR-VTT UMT-L (ViT-L/16) text-to-video R@5 81.0 #5 of 40 Archive leaderboard report
Video Retrieval MSR-VTT UMT-L (ViT-L/16) video-to-text R@1 58.6 #5 of 40 Archive leaderboard report
Video Retrieval MSR-VTT UMT-L (ViT-L/16) video-to-text R@10 86.5 #5 of 40 Archive leaderboard report
Video Retrieval MSR-VTT UMT-L (ViT-L/16) video-to-text R@5 81.6 #5 of 40 Archive leaderboard report
Video Retrieval SSv2-label retrieval UMT-L (ViT-L/16) text-to-video R@1 73.3 #1 of 5 Archive leaderboard report
Video Retrieval SSv2-label retrieval UMT-L (ViT-L/16) text-to-video R@10 96.6 #1 of 5 Archive leaderboard report
Video Retrieval SSv2-label retrieval UMT-L (ViT-L/16) text-to-video R@5 92.7 #1 of 5 Archive leaderboard report
Video Retrieval SSv2-template retrieval UMT-L (ViT-L/16) text-to-video R@1 90.8 #1 of 5 Archive leaderboard report
Video Retrieval SSv2-template retrieval UMT-L (ViT-L/16) text-to-video R@10 100.0 #1 of 5 Archive leaderboard report
Video Retrieval SSv2-template retrieval UMT-L (ViT-L/16) text-to-video R@5 100.0 #1 of 5 Archive leaderboard report
Video Retrieval VATEX Unmasked Teacher text-to-video R@1 72 #5 of 13 Archive leaderboard report
Video Retrieval VATEX Unmasked Teacher text-to-video R@10 97.8 #5 of 13 Archive leaderboard report
Video Retrieval VATEX Unmasked Teacher text-to-video R@5 95.1 #5 of 13 Archive leaderboard report
Video Retrieval VATEX Unmasked Teacher video-to-text R@1 86.0 #5 of 13 Archive leaderboard report
Video Retrieval VATEX Unmasked Teacher video-to-text R@10 99.6 #5 of 13 Archive leaderboard report
Visual Question Answering (VQA) MSRVTT-QA UMT-L (ViT-L/16) Accuracy 0.471 #7 of 34 Archive leaderboard report
Visual Question Answering (VQA) MSVD-QA UMT-L (ViT-L/16) Accuracy 0.552 #13 of 36 Archive leaderboard report
Zero-Shot Video Retrieval ActivityNet UMT-L (ViT-L/16) text-to-video R@1 42.8 #4 of 12 Archive leaderboard report
Zero-Shot Video Retrieval ActivityNet UMT-L (ViT-L/16) text-to-video R@10 79.8 #4 of 12 Archive leaderboard report
Zero-Shot Video Retrieval ActivityNet UMT-L (ViT-L/16) text-to-video R@5 69.6 #4 of 12 Archive leaderboard report
Zero-Shot Video Retrieval ActivityNet UMT-L (ViT-L/16) video-to-text R@1 40.7 #4 of 12 Archive leaderboard report
Zero-Shot Video Retrieval ActivityNet UMT-L (ViT-L/16) video-to-text R@10 78.6 #4 of 12 Archive leaderboard report
Zero-Shot Video Retrieval ActivityNet UMT-L (ViT-L/16) video-to-text R@5 67.6 #4 of 12 Archive leaderboard report
Zero-Shot Video Retrieval DiDeMo UMT-L (ViT-L/16) text-to-video R@1 48.6 #6 of 26 Archive leaderboard report
Zero-Shot Video Retrieval DiDeMo UMT-L (ViT-L/16) text-to-video R@10 79.0 #6 of 26 Archive leaderboard report
Zero-Shot Video Retrieval DiDeMo UMT-L (ViT-L/16) text-to-video R@5 72.9 #6 of 26 Archive leaderboard report
Zero-Shot Video Retrieval DiDeMo UMT-L (ViT-L/16) video-to-text R@1 49.9 #6 of 26 Archive leaderboard report
Zero-Shot Video Retrieval DiDeMo UMT-L (ViT-L/16) video-to-text R@10 81.4 #6 of 26 Archive leaderboard report
Zero-Shot Video Retrieval DiDeMo UMT-L (ViT-L/16) video-to-text R@5 74.8 #6 of 26 Archive leaderboard report
Zero-Shot Video Retrieval LSMDC UMT-L (ViT-L/16) text-to-video R@1 25.2 #4 of 16 Archive leaderboard report
Zero-Shot Video Retrieval LSMDC UMT-L (ViT-L/16) text-to-video R@10 50.5 #4 of 16 Archive leaderboard report
Zero-Shot Video Retrieval LSMDC UMT-L (ViT-L/16) text-to-video R@5 43.0 #4 of 16 Archive leaderboard report
Zero-Shot Video Retrieval LSMDC UMT-L (ViT-L/16) video-to-text R@1 23.2 #4 of 16 Archive leaderboard report
Zero-Shot Video Retrieval LSMDC UMT-L (ViT-L/16) video-to-text R@10 44.2 #4 of 16 Archive leaderboard report
Zero-Shot Video Retrieval LSMDC UMT-L (ViT-L/16) video-to-text R@5 37.7 #4 of 16 Archive leaderboard report
Zero-Shot Video Retrieval MSR-VTT UMT-L (ViT-L/16) text-to-video R@1 42.6 #11 of 41 Archive leaderboard report
Zero-Shot Video Retrieval MSR-VTT UMT-L (ViT-L/16) text-to-video R@10 73.1 #11 of 41 Archive leaderboard report
Zero-Shot Video Retrieval MSR-VTT UMT-L (ViT-L/16) text-to-video R@5 64.4 #11 of 41 Archive leaderboard report
Zero-Shot Video Retrieval MSR-VTT UMT-L (ViT-L/16) video-to-text R@1 38.6 #11 of 41 Archive leaderboard report
Zero-Shot Video Retrieval MSR-VTT UMT-L (ViT-L/16) video-to-text R@10 69.6 #11 of 41 Archive leaderboard report
Zero-Shot Video Retrieval MSR-VTT UMT-L (ViT-L/16) video-to-text R@5 59.8 #11 of 41 Archive leaderboard report
Zero-Shot Video Retrieval MSVD UMT-L (ViT-L/16) text-to-video R@1 49.0 #7 of 14 Archive leaderboard report
Zero-Shot Video Retrieval MSVD UMT-L (ViT-L/16) text-to-video R@10 84.7 #7 of 14 Archive leaderboard report
Zero-Shot Video Retrieval MSVD UMT-L (ViT-L/16) text-to-video R@5 76.9 #7 of 14 Archive leaderboard report
Zero-Shot Video Retrieval MSVD UMT-L (ViT-L/16) video-to-text R@1 74.5 #7 of 14 Archive leaderboard report
Zero-Shot Video Retrieval MSVD UMT-L (ViT-L/16) video-to-text R@10 92.8 #7 of 14 Archive leaderboard report
Zero-Shot Video Retrieval MSVD UMT-L (ViT-L/16) video-to-text R@5 89.7 #7 of 14 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