Papers › Make Your Training Flexible: Towards Deployment-Efficient Video Models

Make Your Training Flexible: Towards Deployment-Efficient Video Models

18 Mar 2025arXiv:2503.14237archive 2025-07-28

Chenting Wang, Kunchang Li, Tianxiang Jiang, Xiangyu Zeng, Yi Wang, LiMin Wang

Popular video training methods mainly operate on a fixed number of tokens sampled from a predetermined spatiotemporal grid, resulting in sub-optimal accuracy-computation trade-offs due to inherent video redundancy. They also lack adaptability to varying computational budgets for downstream tasks, hindering applications of the most competitive model in real-world scenes. We thus propose a new test setting, Token Optimization, for maximized input information across budgets, which optimizes the size-limited set of input tokens through token selection from more suitably sampled videos. To this end, we propose a novel augmentation tool termed Flux. By making the sampling grid flexible and leveraging token selection, it is easily adopted in most popular video training frameworks, boosting model robustness with nearly no additional cost. We integrate Flux in large-scale video pre-training, and the resulting FluxViT establishes new state-of-the-art results across extensive tasks at standard costs. Notably, with 1/4 tokens only, it can still match the performance of previous state-of-the-art models with Token Optimization, yielding nearly 90% savings. All models and data are available at https://github.com/OpenGVLab/FluxViT.

PaperPDFCodeCode Syntology ran

In Syntology View this paper on Syntology, its page in Syntology's graph. That page adds which of its functions are also present in other papers' code Syntology holds, with dates, and its dated run record as JSON and a README badge. It also lists the repositories linked to the paper, the abstract and the calls for agents.

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

For agents, on Syntology's MCP service (how to connect):

Code

Syntology 17 code samples harvested from 1 repository linked to this paper: 17 ran (of which 0 constructed an object rather than computing a result; 17 with no instrument failure: 3 honoured, 0 violated, 14 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code; here there are none.

By repository: official repository: 17 samples from 1 repository, 17 ran (17 with no instrument failure; 0 where Syntology's instrument failed). The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.

Levels written between 2026-09-24 and 2026-09-25 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed.

opengvlab/fluxvit officialmentioned in papermentioned on GitHubpytorchMITSyntology: 17 samples, 17 ran (0 instrument-failed) · 0 unverified 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

17 samples harvested: 17 ran (of which 0 constructed an object rather than computing a result; 17 with no instrument failure: 3 honoured, 0 violated, 14 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Syntology's record for this page last changed ; when this build read Syntology's graph is in the build record. That date is not when the samples ran.

3ran · honoured contract
14ran

Licence: 6 of the 17 samples are pointer only, meaning Syntology does not serve that copy's text. For 6 of them, this paper's copy is under a permissive licence and cleared, but the stored copy of the same code that Syntology serves is not; the licence label's tooltip gives Syntology's recorded reason. This page shows no code text for any sample; each one links to its file in the repository.

Harvested from opengvlab/fluxvit. “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. A line beginning “fixture:” under a sample that ran gives the arguments Syntology's fixture called the sample with, each as name:type/rank/element kind (? where none was recorded), and the type of what the fixture's run returned, exactly as the harness recorded them; the arguments and the returned type were recorded by two separate runs of the same fixture. No output value is shown, and a returned type says nothing about whether the output was right. “Level recorded” gives the UTC date when Syntology's graph wrote the sample's verification level, not when the code executed; where the level was written before Syntology's graph kept that date, the row says “level recorded, date not recorded”. “Honoured” and “violated” refer to a contract Syntology drafted from the code itself. “Our draft was wrong” and “the contract check could not run” are failures of Syntology's instrument, not of the code: the sample ran on Syntology's synthesized input, and then the drafted contract did not describe it, or the check itself raised an error (RAISES), lacked a dependency (DEP_MISSING) or ran out of time (TIMEOUT); hover the label for which. A metamorphic tier, where shown, is evidence of the output's shape, not that the sample implements the method.

Each sample ends with its code_sha256. A sample's code_sha256 is Syntology's identity for that exact code. It is printed in full (16 hexadecimal characters for every sample in this build), and get_code(code_sha256="…") takes it exactly as printed. An agent calls it on Syntology's MCP service (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 are grouped by repository: a repository marked official first, then by how directly the repository is tied to the paper, then in the archive's order; never by how many ran. 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 six 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; the licence is outside the permissive list Syntology serves text under (MIT, Apache-2.0, BSD and similar); this paper's copy was not located, so its licence is not recorded; or this paper's copy is cleared, but the stored copy of the same code that Syntology serves is not. 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.

A function or class name links to its page when the name occurs in the harvested code of at least 5 papers; the name is a string, not an identity.

opengvlab/fluxvit official repository 17 samples, 17 ran (0 instrument-failed) · 0 unverified
cosine_scheduler in 74 papers single_modality/utils.py ran MIT (permissive) · 3f8a3473642aa49d · report fixture: signature base_value:float/0/float, final_value:float/0/float, epochs:int/0/int, niter_per_ep:int/0/int, warmup_epochs:int/0/int, start_warmup_value:float/0/float, warmup_steps:int/0/int · returned ndarray · level recorded 2026-09-24
create_optimizer in 37 papers single_modality/optim_factory.py ran MIT (permissive) · 5e4e5db6f2ca5c09 · report fixture: signature args:Args/0/?, model:callable/0/?, get_num_layer:callable/0/?, get_layer_scale:callable/0/?, filter_bias_and_bn:bool/0/bool, skip_list:list/1/empty · returned Adam · level recorded 2026-09-25
crop_clip in 14 papers single_modality/functional.py ran MIT (permissive) · bc39b8ed5fefa7eb · report fixture: signature clip:list/1/ndarray, min_h:int/0/int, min_w:int/0/int, h:int/0/int, w:int/0/int · returned list · level recorded 2026-09-24
get_1d_sincos_pos_embed in 16 papers single_modality/models/pos_embed.py ran fingerprinted MIT (permissive) here; the stored copy is pointer only · 1075afcc0bf619d6 · report fixture: signature embed_dim:int/0/int, t_size:int/0/int, cls_token:bool/0/bool · returned ndarray · level recorded 2026-09-24
get_2d_sincos_pos_embed in 170 papers single_modality/models/pos_embed.py ran · honoured contract fingerprinted MIT (permissive) here; the stored copy is pointer only · 77e8a3ac46f3afec · report fixture: signature embed_dim:int/0/int, grid_size:int/0/int, cls_token:bool/0/bool · returned ndarray · level recorded 2026-09-24
get_3d_sincos_pos_embed in 16 papers single_modality/models/pos_embed.py ran · honoured contract MIT (permissive) here; the stored copy is pointer only · 54f45ddeec14da87 · report fixture: signature embed_dim:int/0/int, grid_size:int/0/int, t_size:int/0/int, cls_token:bool/0/bool · returned ndarray · level recorded 2026-09-24
get_ceph_path single_modality/utils.py ran MIT (permissive) · 090a19eec6390c6c · report fixture: signature ckpt_path:str/0/str, ceph_args:dict/0/? · returned tuple · level recorded 2026-09-24
get_grad_norm_ in 22 papers single_modality/utils.py ran · honoured contract MIT (permissive) here; the stored copy is pointer only · ba1356e8ceb654d2 · report fixture: signature parameters:list/1/Tensor, norm_type:float/0/float · returned Tensor · level recorded 2026-09-25
get_num_layer_for_vit in 23 papers single_modality/optim_factory.py ran fingerprinted MIT (permissive) · 009cf18580e4b995 · report fixture: signature var_name:str/0/str, num_max_layer:int/0/int · returned int · level recorded 2026-09-25
get_objective_score single_modality/models/vid_tldr.py ran MIT (permissive) · 4e608b064deca5f5 · report fixture: signature score_attn:Tensor/4/float, eps:float/0/float · returned Tensor · level recorded 2026-09-24
get_objective_score_gini single_modality/models/vid_tldr.py ran fingerprinted MIT (permissive) · 7aae2e8f7142f44d · report fixture: signature score_attn:Tensor/4/float · returned Tensor · level recorded 2026-09-24
get_parameter_groups in 30 papers single_modality/optim_factory.py ran MIT (permissive) here; the stored copy is pointer only · 0ef0019514eaaa09 · report fixture: signature model:callable/0/?, weight_decay:float/0/float, skip_list:tuple/1/empty, get_num_layer:callable/0/?, get_layer_scale:callable/0/? · returned list · level recorded 2026-09-25
get_resize_sizes in 11 papers single_modality/functional.py ran MIT (permissive) · 2b53e457d5632b4e · report fixture: signature im_h:int/0/int, im_w:int/0/int, size:int/0/int · returned tuple · level recorded 2026-09-24
gini_impurity single_modality/models/vid_tldr.py ran fingerprinted MIT (permissive) · 75b43d03703e3dd4 · report fixture: signature probabilities:Tensor/2/float · returned Tensor · level recorded 2026-09-25
resize_clip in 14 papers single_modality/functional.py ran MIT (permissive) · 981936c73d3cdd12 · report fixture: signature clip:list/1/ndarray, size:int/0/int, interpolation:str/0/str · returned list · level recorded 2026-09-24
spatial_sampling in 8 papers single_modality/datasets/kinetics.py ran MIT (permissive) · f188ca6f84532f7f · report fixture: returned Tensor · level recorded 2026-09-24
tensor_normalize in 18 papers single_modality/datasets/kinetics.py ran MIT (permissive) here; the stored copy is pointer only · 14b3b13b2e7ee844 · report fixture: signature tensor:Tensor/2/int, mean:list/1/float, std:list/1/float · returned Tensor · level recorded 2026-09-25

Tasks

Action ClassificationZero-Shot Video Retrieval

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Action Classification Kinetics-400 FluxViT-B Acc@1 90.0 #13 of 207 Archive leaderboard report
Action Classification Kinetics-400 FluxViT-B FLOPs (G) x views 440x3x4 #13 of 207 Archive leaderboard report
Action Classification Kinetics-400 FluxViT-B Parameters (M) 97 #13 of 207 Archive leaderboard report
Action Classification Kinetics-400 FluxViT-S Acc@1 88.0 #27 of 207 Archive leaderboard report
Action Classification Kinetics-400 FluxViT-S FLOPs (G) x views 154x3x4 #27 of 207 Archive leaderboard report
Action Classification Kinetics-400 FluxViT-S Parameters (M) 24 #27 of 207 Archive leaderboard report
Zero-Shot Video Retrieval MSR-VTT FluxViT-B text-to-video R@1 49.9 #5 of 41 Archive leaderboard report
Zero-Shot Video Retrieval MSR-VTT FluxViT-B text-to-video R@10 79.6 #5 of 41 Archive leaderboard report
Zero-Shot Video Retrieval MSR-VTT FluxViT-B text-to-video R@5 71.0 #5 of 41 Archive leaderboard report
Zero-Shot Video Retrieval MSR-VTT FluxViT-B video-to-text R@1 49.4 #5 of 41 Archive leaderboard report
Zero-Shot Video Retrieval MSR-VTT FluxViT-B video-to-text R@10 82.4 #5 of 41 Archive leaderboard report
Zero-Shot Video Retrieval MSR-VTT FluxViT-B video-to-text R@5 73.9 #5 of 41 Archive leaderboard report
Zero-Shot Video Retrieval MSR-VTT FluxViT-S text-to-video R@1 45.0 #8 of 41 Archive leaderboard report
Zero-Shot Video Retrieval MSR-VTT FluxViT-S text-to-video R@10 75.8 #8 of 41 Archive leaderboard report
Zero-Shot Video Retrieval MSR-VTT FluxViT-S text-to-video R@5 67.5 #8 of 41 Archive leaderboard report
Zero-Shot Video Retrieval MSR-VTT FluxViT-S video-to-text R@1 44.9 #8 of 41 Archive leaderboard report
Zero-Shot Video Retrieval MSR-VTT FluxViT-S video-to-text R@10 76.5 #8 of 41 Archive leaderboard report
Zero-Shot Video Retrieval MSR-VTT FluxViT-S video-to-text R@5 68.2 #8 of 41 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

SET

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