{"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/large-scale-weakly-supervised-pre-training","title":"Large-scale weakly-supervised pre-training for video action recognition","arxiv_id":"1905.00561","date":"2019-05-02","proceeding":"CVPR 2019 6","authors":["Deepti Ghadiyaram","Matt Feiszli","Du Tran","Xueting Yan","Heng Wang","Dhruv Mahajan"],"abstract":"Current fully-supervised video datasets consist of only a few hundred\nthousand videos and fewer than a thousand domain-specific labels. This hinders\nthe progress towards advanced video architectures. This paper presents an\nin-depth study of using large volumes of web videos for pre-training video\nmodels for the task of action recognition. Our primary empirical finding is\nthat pre-training at a very large scale (over 65 million videos), despite on\nnoisy social-media videos and hashtags, substantially improves the\nstate-of-the-art on three challenging public action recognition datasets.\nFurther, we examine three questions in the construction of weakly-supervised\nvideo action datasets. First, given that actions involve interactions with\nobjects, how should one construct a verb-object pre-training label space to\nbenefit transfer learning the most? Second, frame-based models perform quite\nwell on action recognition; is pre-training for good image features sufficient\nor is pre-training for spatio-temporal features valuable for optimal transfer\nlearning? Finally, actions are generally less well-localized in long videos vs.\nshort videos; since action labels are provided at a video level, how should one\nchoose video clips for best performance, given some fixed budget of number or\nminutes of videos?","url_abs":"http://arxiv.org/abs/1905.00561v1","url_pdf":"http://arxiv.org/pdf/1905.00561v1.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":"large-scale-weakly-supervised-pre-training","repo_url":"https://github.com/moabitcoin/ig65m-pytorch","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}},{"paper_slug":"large-scale-weakly-supervised-pre-training","repo_url":"https://github.com/microsoft/computervision-recipes","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}},{"paper_slug":"large-scale-weakly-supervised-pre-training","repo_url":"https://github.com/open-mmlab/mmaction2","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"action-classification","task_name":"Action Classification"},{"task_slug":"action-recognition-in-videos","task_name":"Action Recognition"},{"task_slug":"activity-recognition","task_name":"Activity Recognition"},{"task_slug":"activity-recognition-in-videos","task_name":"Activity Recognition In Videos"},{"task_slug":"egocentric-activity-recognition","task_name":"Egocentric Activity Recognition"},{"task_slug":"action-recognition","task_name":"Temporal Action Localization"},{"task_slug":"transfer-learning","task_name":"Transfer Learning"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/action-classification-on-kinetics-400","task":"Action Classification","dataset":"Kinetics-400","model":"irCSN-152 (IG-Kinetics-65M pretrain)","rank_in_archive_order":72,"of":207,"metrics":{"Acc@1":"82.8"},"uses_additional_data":false},{"leaderboard":"/sota/egocentric-activity-recognition-on-epic-1","task":"Egocentric Activity Recognition","dataset":"EPIC-KITCHENS-55","model":"R(2+1)D-152-SE (ig)","rank_in_archive_order":5,"of":7,"metrics":{"Actions Top-1 (S2)":"25.6"},"uses_additional_data":false},{"leaderboard":"/sota/egocentric-activity-recognition-on-epic-1","task":"Egocentric Activity Recognition","dataset":"EPIC-KITCHENS-55","model":"R(2+1)D-34 (kinetics)","rank_in_archive_order":6,"of":7,"metrics":{"Actions Top-1 (S2)":"16.8"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1905.00561","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"1905.00561"}},"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":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/open-mmlab/mmaction2","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/microsoft/computervision-recipes","reach":{"status":"ok","spdx":"MIT"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/moabitcoin/ig65m-pytorch","reach":{"status":"ok","spdx":"MIT"}}],"summary":{"unverified":21},"by_repo_kind":{"listed":{"samples":21,"ran":0,"repositories":2}},"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":"b79851f2da94bb38","entry":"Compute_PLoss","repo":"microsoft/computervision-recipes","repo_kind":"listed","path":"contrib/document_cleanup/light_weight_document_cleanup_ICDAR2021/loss_function.py","file_url":"https://github.com/microsoft/computervision-recipes/blob/HEAD/contrib/document_cleanup/light_weight_document_cleanup_ICDAR2021/loss_function.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":"b79851f2da94bb38"}},{"code_sha256_prefix":"d2e1797dd9a3e283","entry":"CreateModel_M16_binary","repo":"microsoft/computervision-recipes","repo_kind":"listed","path":"contrib/document_cleanup/light_weight_document_cleanup_ICDAR2021/model.py","file_url":"https://github.com/microsoft/computervision-recipes/blob/HEAD/contrib/document_cleanup/light_weight_document_cleanup_ICDAR2021/model.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":"d2e1797dd9a3e283"}},{"code_sha256_prefix":"8439277e085dfc38","entry":"blobs_from_pkl","repo":"moabitcoin/ig65m-pytorch","repo_kind":"listed","path":"ig65m/cli/convert.py","file_url":"https://github.com/moabitcoin/ig65m-pytorch/blob/HEAD/ig65m/cli/convert.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":"8439277e085dfc38"}},{"code_sha256_prefix":"2ae3a0bf00128d4b","entry":"convert2gray","repo":"microsoft/computervision-recipes","repo_kind":"listed","path":"contrib/document_cleanup/light_weight_document_cleanup_ICDAR2021/model.py","file_url":"https://github.com/microsoft/computervision-recipes/blob/HEAD/contrib/document_cleanup/light_weight_document_cleanup_ICDAR2021/model.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":"2ae3a0bf00128d4b"}},{"code_sha256_prefix":"ca5129f92643bc3e","entry":"evaluate_mot","repo":"microsoft/computervision-recipes","repo_kind":"listed","path":"utils_cv/tracking/model.py","file_url":"https://github.com/microsoft/computervision-recipes/blob/HEAD/utils_cv/tracking/model.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":"ca5129f92643bc3e"}},{"code_sha256_prefix":"e057f034f658dbfe","entry":"get_optimal_threshold","repo":"microsoft/computervision-recipes","repo_kind":"listed","path":"utils_cv/classification/model.py","file_url":"https://github.com/microsoft/computervision-recipes/blob/HEAD/utils_cv/classification/model.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":"e057f034f658dbfe"}},{"code_sha256_prefix":"6f8cd1e7f213cb4a","entry":"get_pretrained_fasterrcnn","repo":"microsoft/computervision-recipes","repo_kind":"listed","path":"utils_cv/detection/model.py","file_url":"https://github.com/microsoft/computervision-recipes/blob/HEAD/utils_cv/detection/model.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":"6f8cd1e7f213cb4a"}},{"code_sha256_prefix":"f763516cdbd12500","entry":"get_pretrained_keypointrcnn","repo":"microsoft/computervision-recipes","repo_kind":"listed","path":"utils_cv/detection/model.py","file_url":"https://github.com/microsoft/computervision-recipes/blob/HEAD/utils_cv/detection/model.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":"f763516cdbd12500"}},{"code_sha256_prefix":"3769b815f2bf3b26","entry":"get_pretrained_maskrcnn","repo":"microsoft/computervision-recipes","repo_kind":"listed","path":"utils_cv/detection/model.py","file_url":"https://github.com/microsoft/computervision-recipes/blob/HEAD/utils_cv/detection/model.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":"3769b815f2bf3b26"}},{"code_sha256_prefix":"f809ba52b7fe6bf7","entry":"get_ratio_correct_metric","repo":"microsoft/computervision-recipes","repo_kind":"listed","path":"utils_cv/segmentation/model.py","file_url":"https://github.com/microsoft/computervision-recipes/blob/HEAD/utils_cv/segmentation/model.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":"f809ba52b7fe6bf7"}},{"code_sha256_prefix":"24caeadc560dd7b1","entry":"gram_matrix","repo":"microsoft/computervision-recipes","repo_kind":"listed","path":"contrib/document_cleanup/light_weight_document_cleanup_ICDAR2021/loss_function.py","file_url":"https://github.com/microsoft/computervision-recipes/blob/HEAD/contrib/document_cleanup/light_weight_document_cleanup_ICDAR2021/loss_function.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":"24caeadc560dd7b1"}},{"code_sha256_prefix":"6e3c07a22fc403cd","entry":"hamming_accuracy","repo":"microsoft/computervision-recipes","repo_kind":"listed","path":"utils_cv/classification/model.py","file_url":"https://github.com/microsoft/computervision-recipes/blob/HEAD/utils_cv/classification/model.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":"6e3c07a22fc403cd"}},{"code_sha256_prefix":"8baa827e7a093444","entry":"is_installed","repo":"microsoft/computervision-recipes","repo_kind":"listed","path":"contrib/vm_builder/vm_builder.py","file_url":"https://github.com/microsoft/computervision-recipes/blob/HEAD/contrib/vm_builder/vm_builder.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":"8baa827e7a093444"}},{"code_sha256_prefix":"1a80caf172692d16","entry":"main","repo":"moabitcoin/ig65m-pytorch","repo_kind":"listed","path":"ig65m/cli/server.py","file_url":"https://github.com/moabitcoin/ig65m-pytorch/blob/HEAD/ig65m/cli/server.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":"1a80caf172692d16"}},{"code_sha256_prefix":"cb39d2a70fde7d87","entry":"r2plus1d_34_32_ig65m","repo":"moabitcoin/ig65m-pytorch","repo_kind":"listed","path":"ig65m/models.py","file_url":"https://github.com/moabitcoin/ig65m-pytorch/blob/HEAD/ig65m/models.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":"cb39d2a70fde7d87"}},{"code_sha256_prefix":"21dfb37daa3d5a63","entry":"r2plus1d_34_8_ig65m","repo":"moabitcoin/ig65m-pytorch","repo_kind":"listed","path":"ig65m/models.py","file_url":"https://github.com/moabitcoin/ig65m-pytorch/blob/HEAD/ig65m/models.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":"21dfb37daa3d5a63"}},{"code_sha256_prefix":"ba55356a87900992","entry":"r2plus1d_34_8_kinetics","repo":"moabitcoin/ig65m-pytorch","repo_kind":"listed","path":"ig65m/models.py","file_url":"https://github.com/moabitcoin/ig65m-pytorch/blob/HEAD/ig65m/models.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":"ba55356a87900992"}},{"code_sha256_prefix":"c9d0ef61d7a0052a","entry":"ratio_correct","repo":"microsoft/computervision-recipes","repo_kind":"listed","path":"utils_cv/segmentation/model.py","file_url":"https://github.com/microsoft/computervision-recipes/blob/HEAD/utils_cv/segmentation/model.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":"c9d0ef61d7a0052a"}},{"code_sha256_prefix":"6da0ad792057efe4","entry":"res_net_block","repo":"microsoft/computervision-recipes","repo_kind":"listed","path":"contrib/document_cleanup/light_weight_document_cleanup_ICDAR2021/model.py","file_url":"https://github.com/microsoft/computervision-recipes/blob/HEAD/contrib/document_cleanup/light_weight_document_cleanup_ICDAR2021/model.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":"6da0ad792057efe4"}},{"code_sha256_prefix":"c9e8ae02d557b195","entry":"vgg_layers","repo":"microsoft/computervision-recipes","repo_kind":"listed","path":"contrib/document_cleanup/light_weight_document_cleanup_ICDAR2021/loss_function.py","file_url":"https://github.com/microsoft/computervision-recipes/blob/HEAD/contrib/document_cleanup/light_weight_document_cleanup_ICDAR2021/loss_function.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":"c9e8ae02d557b195"}},{"code_sha256_prefix":"19b949c8e8b6cc5a","entry":"zero_one_accuracy","repo":"microsoft/computervision-recipes","repo_kind":"listed","path":"utils_cv/classification/model.py","file_url":"https://github.com/microsoft/computervision-recipes/blob/HEAD/utils_cv/classification/model.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":"19b949c8e8b6cc5a"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}