Papers › Video Mamba Suite: State Space Model as a Versatile Alternative for Video Understanding

Video Mamba Suite: State Space Model as a Versatile Alternative for Video Understanding

14 Mar 2024arXiv:2403.09626archive 2025-07-28

Guo Chen, Yifei HUANG, Jilan Xu, Baoqi Pei, Zhe Chen, Zhiqi Li, Jiahao Wang, Kunchang Li, Tong Lu, LiMin Wang

Understanding videos is one of the fundamental directions in computer vision research, with extensive efforts dedicated to exploring various architectures such as RNN, 3D CNN, and Transformers. The newly proposed architecture of state space model, e.g., Mamba, shows promising traits to extend its success in long sequence modeling to video modeling. To assess whether Mamba can be a viable alternative to Transformers in the video understanding domain, in this work, we conduct a comprehensive set of studies, probing different roles Mamba can play in modeling videos, while investigating diverse tasks where Mamba could exhibit superiority. We categorize Mamba into four roles for modeling videos, deriving a Video Mamba Suite composed of 14 models/modules, and evaluating them on 12 video understanding tasks. Our extensive experiments reveal the strong potential of Mamba on both video-only and video-language tasks while showing promising efficiency-performance trade-offs. We hope this work could provide valuable data points and insights for future research on video understanding. Code is public: https://github.com/OpenGVLab/video-mamba-suite.

PaperPDFCodeCode Syntology ran

In Syntology View this paper on Syntology: its repositories, every harvested function with whether it ran, its licence and the call to fetch it.

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="2403.09626")

Code

Syntology Ran 8 of 13 code samples harvested from 1 repository linked to this paper; 5 have no recorded run. Of those that ran: 8 ran with no contract checked.

By repository: official repository: 13 samples from 1 repository, 8 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/video-mamba-suite 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

13 samples harvested; 8 ran; 0 honoured the contract we drafted; 5 have no recorded run. Read from Syntology's graph 2026-09-25; that is when this build read the record, not when the samples ran.

8ran
5unverified

Licence: 0 of the 13 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/video-mamba-suite. “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.

crop_clip opengvlab/video-mamba-suite/video-mamba-suite/action-recognition/functional.py official repository ran MIT (permissive) · bc39b8ed5fefa7eb · report
exponential_descrease opengvlab/video-mamba-suite/video-mamba-suite/temporal-action-segmentation/model.py official repository ran fingerprinted MIT (permissive) · 8467c56ed6949a06 · report
get_loss_scale_for_deepspeed opengvlab/video-mamba-suite/video-mamba-suite/action-recognition/engines/engine_for_finetuning.py official repository ran MIT (permissive) · 6aef977b44c0feaf · report
get_resize_sizes opengvlab/video-mamba-suite/video-mamba-suite/action-recognition/functional.py official repository ran MIT (permissive) · 2b53e457d5632b4e · report
get_sinusoid_encoding_table opengvlab/video-mamba-suite/video-mamba-suite/action-recognition/models/modeling_finetune.py official repository ran MIT (permissive) · f9c6fc8c1ddfbfac · report
resize_clip opengvlab/video-mamba-suite/video-mamba-suite/action-recognition/functional.py official repository ran MIT (permissive) · 981936c73d3cdd12 · report
spatial_sampling opengvlab/video-mamba-suite/video-mamba-suite/action-recognition/datasets/kinetics.py official repository ran MIT (permissive) · f188ca6f84532f7f · report
train_class_batch opengvlab/video-mamba-suite/video-mamba-suite/action-recognition/engines/engine_for_finetuning_regression.py official repository ran MIT (permissive) · fb98703650f02bf6 · report
clip_b16 opengvlab/video-mamba-suite/video-mamba-suite/action-recognition/models/clip.py official repository unverified MIT (permissive) · e72f71197ef8b97e · report
clip_l14 opengvlab/video-mamba-suite/video-mamba-suite/action-recognition/models/clip.py official repository unverified MIT (permissive) · 055a054d676fb93c · report
inflate_weight opengvlab/video-mamba-suite/video-mamba-suite/action-recognition/models/clip.py official repository unverified MIT (permissive) · bdec7f112532de68 · report
tensor_normalize opengvlab/video-mamba-suite/video-mamba-suite/action-recognition/datasets/kinetics.py official repository unverified MIT (permissive) · 14b3b13b2e7ee844 · report
train_class_batch opengvlab/video-mamba-suite/video-mamba-suite/action-recognition/engines/engine_for_finetuning.py official repository unverified MIT (permissive) · c7976ea27edc377a · report

Tasks

MambaMoment RetrievalTemporal Action LocalizationVideo Understanding

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Moment Retrieval Charades-STA video-mamba-suite R@1 IoU=0.5 57.18 #19 of 25 Archive leaderboard report
Moment Retrieval Charades-STA video-mamba-suite R@1 IoU=0.7 36.05 #19 of 25 Archive leaderboard report
Moment Retrieval QVHighlights video-mamba-suite R@1 IoU=0.5 66.65 #14 of 32 Archive leaderboard report
Moment Retrieval QVHighlights video-mamba-suite R@1 IoU=0.7 52.19 #14 of 32 Archive leaderboard report
Moment Retrieval QVHighlights video-mamba-suite mAP 45.18 #14 of 32 Archive leaderboard report
Moment Retrieval QVHighlights video-mamba-suite mAP@0.5 64.37 #14 of 32 Archive leaderboard report
Moment Retrieval QVHighlights video-mamba-suite mAP@0.75 46.68 #14 of 32 Archive leaderboard report
Temporal Action Localization ActivityNet-1.3 ActionMamba (InternVideo2-6B) mAP 42.02 #2 of 33 Archive leaderboard report
Temporal Action Localization ActivityNet-1.3 ActionMamba (InternVideo2-6B) mAP IOU@0.5 62.43 #2 of 33 Archive leaderboard report
Temporal Action Localization ActivityNet-1.3 ActionMamba (InternVideo2-6B) mAP IOU@0.75 43.49 #2 of 33 Archive leaderboard report
Temporal Action Localization ActivityNet-1.3 ActionMamba (InternVideo2-6B) mAP IOU@0.95 10.23 #2 of 33 Archive leaderboard report
Temporal Action Localization FineAction ActionMamba(InternVideo2-6B) mAP 29.04 #2 of 9 Archive leaderboard report
Temporal Action Localization FineAction ActionMamba(InternVideo2-6B) mAP IOU@0.5 45.44 #2 of 9 Archive leaderboard report
Temporal Action Localization FineAction ActionMamba(InternVideo2-6B) mAP IOU@0.75 28.82 #2 of 9 Archive leaderboard report
Temporal Action Localization FineAction ActionMamba(InternVideo2-6B) mAP IOU@0.95 6.79 #2 of 9 Archive leaderboard report
Temporal Action Localization HACS ActionMamba(InternVideo2-6B) Average-mAP 44.56 #2 of 12 Archive leaderboard report
Temporal Action Localization HACS ActionMamba(InternVideo2-6B) mAP@0.5 64.02 #2 of 12 Archive leaderboard report
Temporal Action Localization HACS ActionMamba(InternVideo2-6B) mAP@0.75 45.71 #2 of 12 Archive leaderboard report
Temporal Action Localization HACS ActionMamba(InternVideo2-6B) mAP@0.95 13.34 #2 of 12 Archive leaderboard report
Temporal Action Localization THUMOS’14 ActionMamba(InternVideo2-6B) Avg mAP (0.3:0.7) 72.72 #3 of 42 Archive leaderboard report
Temporal Action Localization THUMOS’14 ActionMamba(InternVideo2-6B) mAP IOU@0.3 86.89 #3 of 42 Archive leaderboard report
Temporal Action Localization THUMOS’14 ActionMamba(InternVideo2-6B) mAP IOU@0.4 83.09 #3 of 42 Archive leaderboard report
Temporal Action Localization THUMOS’14 ActionMamba(InternVideo2-6B) mAP IOU@0.5 76.90 #3 of 42 Archive leaderboard report
Temporal Action Localization THUMOS’14 ActionMamba(InternVideo2-6B) mAP IOU@0.6 65.91 #3 of 42 Archive leaderboard report
Temporal Action Localization THUMOS’14 ActionMamba(InternVideo2-6B) mAP IOU@0.7 50.82 #3 of 42 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

3D CNNSET

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