{"url":"/sota/long-tail-learning-on-coco-mlt","task":{"name":"Long-tail Learning","url":"/task/long-tail-learning","note":null},"dataset":{"name":"COCO-MLT","url":"/dataset/coco-mlt"},"category":"Methodology","categories":["Methodology"],"category_note":null,"description":"Long-tailed learning, one of the most challenging problems in visual recognition, aims to train well-performing models from a large number of images that follow a long-tailed class distribution.","description_from":"task","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","rank":"the archive's row order at snapshot; not re-ranked","rows_end_at":"2025-07-28","rows_withheld_as_spam":0,"metric_values":"the archive's strings, untouched"},"metrics":["Average mAP"],"metric_direction":{"note":"inferred from the metric name only (the archive records no direction); null = not inferred, chart draws points only","by_metric":{"Average mAP":"higher"}},"counts":{"rows":13,"rows_with_code":11,"rows_with_paper_page":13,"rows_dated":13,"rows_using_additional_data":4},"rows":[{"rank_in_archive_order":1,"model":"LMPT(ViT-B/16)","metrics":{"Average mAP":"66.19"},"uses_additional_data":true,"paper_date":"2023-05-08","paper":"/paper/lmpt-prompt-tuning-with-class-specific","paper_url":"https://arxiv.org/abs/2305.04536v2","paper_title":"LMPT: Prompt Tuning with Class-Specific Embedding Loss for Long-tailed Multi-Label Visual Recognition","code":"https://github.com/richard-peng-xia/LMPT","n_code_links":1,"syntology":{"n_ran":0,"n_unverified":9,"n_samples":9,"n_pointer_only_licence":0}},{"rank_in_archive_order":2,"model":"CLIP(ViT-B/16)","metrics":{"Average mAP":"60.17"},"uses_additional_data":true,"paper_date":"2021-02-26","paper":"/paper/learning-transferable-visual-models-from","paper_url":"https://arxiv.org/abs/2103.00020v1","paper_title":"Learning Transferable Visual Models From Natural Language Supervision","code":"https://github.com/openai/CLIP","n_code_links":82,"syntology":{"n_ran":16,"n_unverified":4,"n_samples":20,"n_pointer_only_licence":16}},{"rank_in_archive_order":3,"model":"LMPT(ResNet-50)","metrics":{"Average mAP":"58.97"},"uses_additional_data":true,"paper_date":"2023-05-08","paper":"/paper/lmpt-prompt-tuning-with-class-specific","paper_url":"https://arxiv.org/abs/2305.04536v2","paper_title":"LMPT: Prompt Tuning with Class-Specific Embedding Loss for Long-tailed Multi-Label Visual Recognition","code":"https://github.com/richard-peng-xia/LMPT","n_code_links":1,"syntology":{"n_ran":0,"n_unverified":9,"n_samples":9,"n_pointer_only_licence":0}},{"rank_in_archive_order":4,"model":"LTML(ResNet-50)","metrics":{"Average mAP":"56.90"},"uses_additional_data":false,"paper_date":"2021-06-19","paper":"/paper/long-tailed-multi-label-visual-recognition-by","paper_url":"http://openaccess.thecvf.com//content/CVPR2021/html/Guo_Long-Tailed_Multi-Label_Visual_Recognition_by_Collaborative_Training_on_Uniform_and_CVPR_2021_paper.html","paper_title":"Long-Tailed Multi-Label Visual Recognition by Collaborative Training on Uniform and Re-Balanced Samplings","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":5,"model":"CLIP(ResNet-50)","metrics":{"Average mAP":"56.19"},"uses_additional_data":true,"paper_date":"2021-02-26","paper":"/paper/learning-transferable-visual-models-from","paper_url":"https://arxiv.org/abs/2103.00020v1","paper_title":"Learning Transferable Visual Models From Natural Language Supervision","code":"https://github.com/openai/CLIP","n_code_links":82,"syntology":{"n_ran":16,"n_unverified":4,"n_samples":20,"n_pointer_only_licence":16}},{"rank_in_archive_order":6,"model":"PG Loss(ResNet-50)","metrics":{"Average mAP":"54.43"},"uses_additional_data":false,"paper_date":"2023-06-26","paper":"/paper/probability-guided-loss-for-long-tailed-multi","paper_url":"https://ojs.aaai.org/index.php/AAAI/article/view/25244","paper_title":"Probability Guided Loss for Long-Tailed Multi-Label Image Classification","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":7,"model":"DB Focal(ResNet-50)","metrics":{"Average mAP":"53.55"},"uses_additional_data":false,"paper_date":"2020-07-19","paper":"/paper/distribution-balanced-loss-for-multi-label","paper_url":"https://arxiv.org/abs/2007.09654v4","paper_title":"Distribution-Balanced Loss for Multi-Label Classification in Long-Tailed Datasets","code":"https://github.com/wutong16/DistributionBalancedLoss","n_code_links":1,"syntology":null},{"rank_in_archive_order":8,"model":"Focal Loss(ResNet-50)","metrics":{"Average mAP":"49.46"},"uses_additional_data":false,"paper_date":"2017-08-07","paper":"/paper/focal-loss-for-dense-object-detection","paper_url":"http://arxiv.org/abs/1708.02002v2","paper_title":"Focal Loss for Dense Object Detection","code":"https://github.com/tensorflow/models","n_code_links":234,"syntology":{"n_ran":11,"n_unverified":0,"n_samples":11,"n_pointer_only_licence":6}},{"rank_in_archive_order":9,"model":"CB Loss(ResNet-50)","metrics":{"Average mAP":"49.06"},"uses_additional_data":false,"paper_date":"2019-01-16","paper":"/paper/class-balanced-loss-based-on-effective-number","paper_url":"http://arxiv.org/abs/1901.05555v1","paper_title":"Class-Balanced Loss Based on Effective Number of Samples","code":"https://github.com/vandit15/Class-balanced-loss-pytorch","n_code_links":11,"syntology":{"n_ran":7,"n_unverified":20,"n_samples":27,"n_pointer_only_licence":0}},{"rank_in_archive_order":10,"model":"RS(ResNet-50)","metrics":{"Average mAP":"46.97"},"uses_additional_data":false,"paper_date":"2015-12-18","paper":"/paper/relay-backpropagation-for-effective-learning","paper_url":"http://arxiv.org/abs/1512.05830v2","paper_title":"Relay Backpropagation for Effective Learning of Deep Convolutional Neural Networks","code":"https://github.com/craston/object_detection_cib","n_code_links":1,"syntology":null},{"rank_in_archive_order":11,"model":"OLTR(ResNet-50)","metrics":{"Average mAP":"45.83"},"uses_additional_data":false,"paper_date":"2019-04-10","paper":"/paper/large-scale-long-tailed-recognition-in-an","paper_url":"http://arxiv.org/abs/1904.05160v2","paper_title":"Large-Scale Long-Tailed Recognition in an Open World","code":"https://github.com/zhmiao/OpenLongTailRecognition-OLTR","n_code_links":2,"syntology":null},{"rank_in_archive_order":12,"model":"ML-GCN(ResNet-50)","metrics":{"Average mAP":"44.24"},"uses_additional_data":false,"paper_date":"2019-04-07","paper":"/paper/multi-label-image-recognition-with-graph","paper_url":"http://arxiv.org/abs/1904.03582v1","paper_title":"Multi-Label Image Recognition with Graph Convolutional Networks","code":"https://github.com/megvii-research/ml-gcn","n_code_links":2,"syntology":null},{"rank_in_archive_order":13,"model":"LDAM(ResNet-50)","metrics":{"Average mAP":"40.53"},"uses_additional_data":false,"paper_date":"2019-06-18","paper":"/paper/learning-imbalanced-datasets-with-label","paper_url":"https://arxiv.org/abs/1906.07413v2","paper_title":"Learning Imbalanced Datasets with Label-Distribution-Aware Margin Loss","code":"https://github.com/kaidic/LDAM-DRW","n_code_links":7,"syntology":{"n_ran":2,"n_unverified":9,"n_samples":11,"n_pointer_only_licence":3}}],"since_archive":{"present":false,"note":"No Syntology-extracted rows are published in this build."},"syntology":{"read_at":"2026-09-24T18:15:14+00:00","claim":"Per row: N of M harvested code samples from that row's paper executed on a synthesized fixture; the other M-N are unverified. Not a reproduction of the row's number; not a correctness claim. n_pointer_only_licence counts samples the site points at rather than redistributes (a licence axis, independent of ran/unverified).","rows_with_graph_line":7,"rows_with_any_sample_ran":5,"distinct_papers_with_graph_line":5,"distinct_papers_with_any_sample_ran":4,"samples_over_distinct_papers":{"n_ran":36,"n_unverified":42,"n_samples":78,"n_pointer_only_licence":25,"note":"each paper (arXiv id) counted once, however many rows it is behind; this is the page-level figure"},"samples_row_weighted":{"n_ran":52,"n_unverified":55,"n_samples":107,"n_pointer_only_licence":41,"note":"row-weighted: a paper behind several rows is counted once per row; inflated relative to samples_over_distinct_papers by design, kept for readers summing the per-row syntology blocks"}}}