{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/make-your-training-flexible-towards","title":"Make Your Training Flexible: Towards Deployment-Efficient Video Models","arxiv_id":"2503.14237","date":"2025-03-18","proceeding":null,"authors":["Chenting Wang","Kunchang Li","Tianxiang Jiang","Xiangyu Zeng","Yi Wang","LiMin Wang"],"abstract":"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.","url_abs":"https://arxiv.org/abs/2503.14237v1","url_pdf":"https://arxiv.org/pdf/2503.14237v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"make-your-training-flexible-towards","repo_url":"https://github.com/opengvlab/fluxvit","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[{"task_slug":"action-classification","task_name":"Action Classification"},{"task_slug":"zero-shot-video-retrieval","task_name":"Zero-Shot Video Retrieval"}],"methods":[{"method_slug":"set","method_name":"SET"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/action-classification-on-kinetics-400","task":"Action Classification","dataset":"Kinetics-400","model":"FluxViT-B","rank_in_archive_order":13,"of":207,"metrics":{"Acc@1":"90.0","FLOPs (G) x views":"440x3x4","Parameters (M)":"97"},"uses_additional_data":true},{"leaderboard":"/sota/action-classification-on-kinetics-400","task":"Action Classification","dataset":"Kinetics-400","model":"FluxViT-S","rank_in_archive_order":27,"of":207,"metrics":{"Acc@1":"88.0","FLOPs (G) x views":"154x3x4","Parameters (M)":"24"},"uses_additional_data":true},{"leaderboard":"/sota/zero-shot-video-retrieval-on-msr-vtt","task":"Zero-Shot Video Retrieval","dataset":"MSR-VTT","model":"FluxViT-B","rank_in_archive_order":5,"of":41,"metrics":{"text-to-video R@1":"49.9","text-to-video R@10":"79.6","text-to-video R@5":"71.0","video-to-text R@1":"49.4","video-to-text R@10":"82.4","video-to-text R@5":"73.9"},"uses_additional_data":true},{"leaderboard":"/sota/zero-shot-video-retrieval-on-msr-vtt","task":"Zero-Shot Video Retrieval","dataset":"MSR-VTT","model":"FluxViT-S","rank_in_archive_order":8,"of":41,"metrics":{"text-to-video R@1":"45.0","text-to-video R@10":"75.8","text-to-video R@5":"67.5","video-to-text R@1":"44.9","video-to-text R@10":"76.5","video-to-text R@5":"68.2"},"uses_additional_data":true}],"syntology":{"syntology_url":null,"atlas_url":"https://app.syntology.ai/?focus=2503.14237","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2503.14237"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-24T18:15:14+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"provenance":"deterministic:regex_extraction","url":"https://github.com/OpenGVLab/FluxViT","reach":{"status":"ok","spdx":"MIT"}}],"summary":{"ran":7,"unverified":10},"by_repo_kind":{"official":{"samples":17,"ran":7,"repositories":1}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"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"},"n_pointer_only_for_licence":0,"samples":[{"code_sha256_prefix":"bc39b8ed5fefa7eb","entry":"crop_clip","repo":"OpenGVLab/FluxViT","repo_kind":"official","path":"single_modality/functional.py","file_url":"https://github.com/OpenGVLab/FluxViT/blob/HEAD/single_modality/functional.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"bc39b8ed5fefa7eb"}},{"code_sha256_prefix":"1075afcc0bf619d6","entry":"get_1d_sincos_pos_embed","repo":"OpenGVLab/FluxViT","repo_kind":"official","path":"single_modality/models/pos_embed.py","file_url":"https://github.com/OpenGVLab/FluxViT/blob/HEAD/single_modality/models/pos_embed.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"1075afcc0bf619d6"}},{"code_sha256_prefix":"77e8a3ac46f3afec","entry":"get_2d_sincos_pos_embed","repo":"OpenGVLab/FluxViT","repo_kind":"official","path":"single_modality/models/pos_embed.py","file_url":"https://github.com/OpenGVLab/FluxViT/blob/HEAD/single_modality/models/pos_embed.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"77e8a3ac46f3afec"}},{"code_sha256_prefix":"54f45ddeec14da87","entry":"get_3d_sincos_pos_embed","repo":"OpenGVLab/FluxViT","repo_kind":"official","path":"single_modality/models/pos_embed.py","file_url":"https://github.com/OpenGVLab/FluxViT/blob/HEAD/single_modality/models/pos_embed.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"54f45ddeec14da87"}},{"code_sha256_prefix":"2b53e457d5632b4e","entry":"get_resize_sizes","repo":"OpenGVLab/FluxViT","repo_kind":"official","path":"single_modality/functional.py","file_url":"https://github.com/OpenGVLab/FluxViT/blob/HEAD/single_modality/functional.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"2b53e457d5632b4e"}},{"code_sha256_prefix":"981936c73d3cdd12","entry":"resize_clip","repo":"OpenGVLab/FluxViT","repo_kind":"official","path":"single_modality/functional.py","file_url":"https://github.com/OpenGVLab/FluxViT/blob/HEAD/single_modality/functional.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"981936c73d3cdd12"}},{"code_sha256_prefix":"f188ca6f84532f7f","entry":"spatial_sampling","repo":"OpenGVLab/FluxViT","repo_kind":"official","path":"single_modality/datasets/kinetics.py","file_url":"https://github.com/OpenGVLab/FluxViT/blob/HEAD/single_modality/datasets/kinetics.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"f188ca6f84532f7f"}},{"code_sha256_prefix":"3f8a3473642aa49d","entry":"cosine_scheduler","repo":"OpenGVLab/FluxViT","repo_kind":"official","path":"single_modality/utils.py","file_url":"https://github.com/OpenGVLab/FluxViT/blob/HEAD/single_modality/utils.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"3f8a3473642aa49d"}},{"code_sha256_prefix":"5e4e5db6f2ca5c09","entry":"create_optimizer","repo":"OpenGVLab/FluxViT","repo_kind":"official","path":"single_modality/optim_factory.py","file_url":"https://github.com/OpenGVLab/FluxViT/blob/HEAD/single_modality/optim_factory.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"5e4e5db6f2ca5c09"}},{"code_sha256_prefix":"090a19eec6390c6c","entry":"get_ceph_path","repo":"OpenGVLab/FluxViT","repo_kind":"official","path":"single_modality/utils.py","file_url":"https://github.com/OpenGVLab/FluxViT/blob/HEAD/single_modality/utils.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"090a19eec6390c6c"}},{"code_sha256_prefix":"ba1356e8ceb654d2","entry":"get_grad_norm_","repo":"OpenGVLab/FluxViT","repo_kind":"official","path":"single_modality/utils.py","file_url":"https://github.com/OpenGVLab/FluxViT/blob/HEAD/single_modality/utils.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"ba1356e8ceb654d2"}},{"code_sha256_prefix":"009cf18580e4b995","entry":"get_num_layer_for_vit","repo":"OpenGVLab/FluxViT","repo_kind":"official","path":"single_modality/optim_factory.py","file_url":"https://github.com/OpenGVLab/FluxViT/blob/HEAD/single_modality/optim_factory.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"009cf18580e4b995"}},{"code_sha256_prefix":"4e608b064deca5f5","entry":"get_objective_score","repo":"OpenGVLab/FluxViT","repo_kind":"official","path":"single_modality/models/vid_tldr.py","file_url":"https://github.com/OpenGVLab/FluxViT/blob/HEAD/single_modality/models/vid_tldr.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"4e608b064deca5f5"}},{"code_sha256_prefix":"7aae2e8f7142f44d","entry":"get_objective_score_gini","repo":"OpenGVLab/FluxViT","repo_kind":"official","path":"single_modality/models/vid_tldr.py","file_url":"https://github.com/OpenGVLab/FluxViT/blob/HEAD/single_modality/models/vid_tldr.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"7aae2e8f7142f44d"}},{"code_sha256_prefix":"0ef0019514eaaa09","entry":"get_parameter_groups","repo":"OpenGVLab/FluxViT","repo_kind":"official","path":"single_modality/optim_factory.py","file_url":"https://github.com/OpenGVLab/FluxViT/blob/HEAD/single_modality/optim_factory.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"0ef0019514eaaa09"}},{"code_sha256_prefix":"75b43d03703e3dd4","entry":"gini_impurity","repo":"OpenGVLab/FluxViT","repo_kind":"official","path":"single_modality/models/vid_tldr.py","file_url":"https://github.com/OpenGVLab/FluxViT/blob/HEAD/single_modality/models/vid_tldr.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"75b43d03703e3dd4"}},{"code_sha256_prefix":"14b3b13b2e7ee844","entry":"tensor_normalize","repo":"OpenGVLab/FluxViT","repo_kind":"official","path":"single_modality/datasets/kinetics.py","file_url":"https://github.com/OpenGVLab/FluxViT/blob/HEAD/single_modality/datasets/kinetics.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"14b3b13b2e7ee844"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}