{"url":"/sota/audio-classification-on-esc-50","task":{"name":"Audio Classification","url":"/task/audio-classification","note":null},"dataset":{"name":"ESC-50","url":"/dataset/esc-50"},"category":"Audio","categories":["Audio"],"category_note":null,"description":"**Audio Classification** is a machine learning task that involves identifying and tagging audio signals into different classes or categories. The goal of audio classification is to enable machines to automatically recognize and distinguish between different types of audio, such as music, speech, and environmental sounds.","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":["Top-1 Accuracy","PRE-TRAINING DATASET","Accuracy (5-fold)"],"metric_direction":{"note":"inferred from the metric name only (the archive records no direction); null = not inferred, chart draws points only","by_metric":{"Top-1 Accuracy":"higher","PRE-TRAINING DATASET":null,"Accuracy (5-fold)":"higher"}},"counts":{"rows":29,"rows_with_code":22,"rows_with_paper_page":27,"rows_dated":28,"rows_using_additional_data":16},"rows":[{"rank_in_archive_order":1,"model":"OmniVec2","metrics":{"Accuracy (5-fold)":"99.1","PRE-TRAINING DATASET":"Multiple","Top-1 Accuracy":"99.1"},"uses_additional_data":true,"paper_date":"2024-01-01","paper":"/paper/omnivec2-a-novel-transformer-based-network","paper_url":"http://openaccess.thecvf.com//content/CVPR2024/html/Srivastava_OmniVec2_-_A_Novel_Transformer_based_Network_for_Large_Scale_CVPR_2024_paper.html","paper_title":"OmniVec2 - A Novel Transformer based Network for Large Scale Multimodal and Multitask Learning","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":2,"model":"InternVideo2","metrics":{"Accuracy (5-fold)":"98.6","PRE-TRAINING DATASET":"Multiple","Top-1 Accuracy":"98.6"},"uses_additional_data":true,"paper_date":"2024-03-22","paper":"/paper/internvideo2-scaling-video-foundation-models","paper_url":"https://arxiv.org/abs/2403.15377v4","paper_title":"InternVideo2: Scaling Foundation Models for Multimodal Video Understanding","code":"https://github.com/opengvlab/internvideo","n_code_links":2,"syntology":null},{"rank_in_archive_order":3,"model":"M2D2 AS+","metrics":{"Accuracy (5-fold)":"98.5","PRE-TRAINING DATASET":"AudioSet,WavCaps","Top-1 Accuracy":"98.5"},"uses_additional_data":true,"paper_date":"2025-03-28","paper":"/paper/m2d2-exploring-general-purpose-audio-language","paper_url":"https://arxiv.org/abs/2503.22104v1","paper_title":"M2D2: Exploring General-purpose Audio-Language Representations Beyond CLAP","code":"https://github.com/nttcslab/m2d","n_code_links":2,"syntology":null},{"rank_in_archive_order":4,"model":"OmniVec","metrics":{"Accuracy (5-fold)":"98.4","PRE-TRAINING DATASET":"Multiple","Top-1 Accuracy":"98.4"},"uses_additional_data":true,"paper_date":"2023-11-07","paper":"/paper/omnivec-learning-robust-representations-with","paper_url":"https://arxiv.org/abs/2311.05709v1","paper_title":"OmniVec: Learning robust representations with cross modal sharing","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":5,"model":"BEATs","metrics":{"Accuracy (5-fold)":"98.1","PRE-TRAINING DATASET":"AudioSet","Top-1 Accuracy":"98.1"},"uses_additional_data":true,"paper_date":"2022-12-18","paper":"/paper/beats-audio-pre-training-with-acoustic","paper_url":"https://arxiv.org/abs/2212.09058v1","paper_title":"BEATs: Audio Pre-Training with Acoustic Tokenizers","code":"https://github.com/microsoft/unilm/tree/master/beats","n_code_links":4,"syntology":{"n_ran":3,"n_unverified":15,"n_samples":18,"n_pointer_only_licence":0}},{"rank_in_archive_order":6,"model":"mn40_as","metrics":{"Accuracy (5-fold)":"97.45","PRE-TRAINING DATASET":"AudioSet","Top-1 Accuracy":"97.45"},"uses_additional_data":true,"paper_date":"2022-11-09","paper":"/paper/efficient-large-scale-audio-tagging-via","paper_url":"https://arxiv.org/abs/2211.04772v3","paper_title":"Efficient Large-scale Audio Tagging via Transformer-to-CNN Knowledge Distillation","code":"https://github.com/fschmid56/efficientat","n_code_links":2,"syntology":{"n_ran":1,"n_unverified":1,"n_samples":2,"n_pointer_only_licence":0}},{"rank_in_archive_order":7,"model":"DyMN-L","metrics":{"Accuracy (5-fold)":"97.4","PRE-TRAINING DATASET":"AudioSet","Top-1 Accuracy":"97.4"},"uses_additional_data":true,"paper_date":"2023-10-24","paper":"/paper/dynamic-convolutional-neural-networks-as","paper_url":"https://arxiv.org/abs/2310.15648v1","paper_title":"Dynamic Convolutional Neural Networks as Efficient Pre-trained Audio Models","code":"https://github.com/fschmid56/efficientat","n_code_links":1,"syntology":null},{"rank_in_archive_order":8,"model":"M2D-CLAP/0.7","metrics":{"Accuracy (5-fold)":"97.4","PRE-TRAINING DATASET":"AudioSet","Top-1 Accuracy":"97.4"},"uses_additional_data":true,"paper_date":"2024-06-04","paper":"/paper/m2d-clap-masked-modeling-duo-meets-clap-for","paper_url":"https://arxiv.org/abs/2406.02032v1","paper_title":"M2D-CLAP: Masked Modeling Duo Meets CLAP for Learning General-purpose Audio-Language Representation","code":"https://github.com/nttcslab/m2d","n_code_links":2,"syntology":null},{"rank_in_archive_order":9,"model":"M2D-AS/0.7","metrics":{"Accuracy (5-fold)":"97.2","PRE-TRAINING DATASET":"AudioSet","Top-1 Accuracy":"97.2"},"uses_additional_data":true,"paper_date":"2024-04-09","paper":"/paper/masked-modeling-duo-towards-a-universal-audio","paper_url":"https://arxiv.org/abs/2404.06095v1","paper_title":"Masked Modeling Duo: Towards a Universal Audio Pre-training Framework","code":"https://github.com/nttcslab/m2d","n_code_links":2,"syntology":{"n_ran":7,"n_unverified":0,"n_samples":7,"n_pointer_only_licence":7}},{"rank_in_archive_order":10,"model":"HTS-AT","metrics":{"Accuracy (5-fold)":"97.0","PRE-TRAINING DATASET":"AudioSet","Top-1 Accuracy":"97.0"},"uses_additional_data":true,"paper_date":"2022-02-02","paper":"/paper/hts-at-a-hierarchical-token-semantic-audio","paper_url":"https://arxiv.org/abs/2202.00874v1","paper_title":"HTS-AT: A Hierarchical Token-Semantic Audio Transformer for Sound Classification and Detection","code":"https://github.com/retrocirce/hts-audio-transformer","n_code_links":1,"syntology":{"n_ran":4,"n_unverified":5,"n_samples":9,"n_pointer_only_licence":0}},{"rank_in_archive_order":11,"model":"EAT-M","metrics":{"Accuracy (5-fold)":"96.3","PRE-TRAINING DATASET":"AudioSet","Top-1 Accuracy":"96.3"},"uses_additional_data":true,"paper_date":"2022-04-25","paper":"/paper/end-to-end-audio-strikes-back-boosting","paper_url":"https://arxiv.org/abs/2204.11479v5","paper_title":"End-to-End Audio Strikes Back: Boosting Augmentations Towards An Efficient Audio Classification Network","code":"https://github.com/Alibaba-MIIL/AudioClassfication","n_code_links":1,"syntology":null},{"rank_in_archive_order":12,"model":"LHGNN","metrics":{"Top-1 Accuracy":"96.2"},"uses_additional_data":false,"paper_date":"2025-01-07","paper":"/paper/lhgnn-local-higher-order-graph-neural","paper_url":"https://arxiv.org/abs/2501.03464v2","paper_title":"LHGNN: Local-Higher Order Graph Neural Networks For Audio Classification and Tagging","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":13,"model":"ERANN-2-5","metrics":{"Accuracy (5-fold)":"96.1","PRE-TRAINING DATASET":"AudioSet","Top-1 Accuracy":"96.1"},"uses_additional_data":false,"paper_date":"2021-06-03","paper":null,"paper_url":"https://arxiv.org/abs/2106.01621v7","paper_title":"ERANNs: Efficient Residual Audio Neural Networks for Audio Pattern Recognition","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":14,"model":"M2D/0.7","metrics":{"Accuracy (5-fold)":"96.0","Top-1 Accuracy":"96.0"},"uses_additional_data":true,"paper_date":"2024-04-09","paper":"/paper/masked-modeling-duo-towards-a-universal-audio","paper_url":"https://arxiv.org/abs/2404.06095v1","paper_title":"Masked Modeling Duo: Towards a Universal Audio Pre-training Framework","code":"https://github.com/nttcslab/m2d","n_code_links":2,"syntology":{"n_ran":7,"n_unverified":0,"n_samples":7,"n_pointer_only_licence":7}},{"rank_in_archive_order":15,"model":"EAT","metrics":{"Accuracy (5-fold)":"96.0","PRE-TRAINING DATASET":"AudioSet","Top-1 Accuracy":"96.0"},"uses_additional_data":true,"paper_date":"2024-01-07","paper":"/paper/eat-self-supervised-pre-training-with","paper_url":"https://arxiv.org/abs/2401.03497v1","paper_title":"EAT: Self-Supervised Pre-Training with Efficient Audio Transformer","code":"https://github.com/cwx-worst-one/eat","n_code_links":1,"syntology":{"n_ran":13,"n_unverified":4,"n_samples":17,"n_pointer_only_licence":0}},{"rank_in_archive_order":16,"model":"Audio Spectrogram Transformer","metrics":{"Accuracy (5-fold)":"95.7","PRE-TRAINING DATASET":"AudioSet, ImageNet","Top-1 Accuracy":"95.7"},"uses_additional_data":true,"paper_date":"2021-04-05","paper":"/paper/ast-audio-spectrogram-transformer","paper_url":"https://arxiv.org/abs/2104.01778v3","paper_title":"AST: Audio Spectrogram Transformer","code":"https://github.com/YuanGongND/ast","n_code_links":5,"syntology":{"n_ran":0,"n_unverified":3,"n_samples":3,"n_pointer_only_licence":0}},{"rank_in_archive_order":17,"model":"EAT-S","metrics":{"Accuracy (5-fold)":"95.25","PRE-TRAINING DATASET":"AudioSet","Top-1 Accuracy":"95.25"},"uses_additional_data":true,"paper_date":"2022-04-25","paper":"/paper/end-to-end-audio-strikes-back-boosting","paper_url":"https://arxiv.org/abs/2204.11479v5","paper_title":"End-to-End Audio Strikes Back: Boosting Augmentations Towards An Efficient Audio Classification Network","code":"https://github.com/Alibaba-MIIL/AudioClassfication","n_code_links":1,"syntology":null},{"rank_in_archive_order":18,"model":"MATPAC (SSL model, linear eval)","metrics":{"Accuracy (5-fold)":"93.5","PRE-TRAINING DATASET":"AudioSet","Top-1 Accuracy":"93.5"},"uses_additional_data":false,"paper_date":"2025-02-17","paper":"/paper/masked-latent-prediction-and-classification","paper_url":"https://arxiv.org/abs/2502.12031v1","paper_title":"Masked Latent Prediction and Classification for Self-Supervised Audio Representation Learning","code":"https://github.com/aurianworld/matpac","n_code_links":1,"syntology":null},{"rank_in_archive_order":19,"model":"EAT-S (scratch)","metrics":{"Accuracy (5-fold)":"92.15","Top-1 Accuracy":"92.15"},"uses_additional_data":false,"paper_date":"2022-04-25","paper":"/paper/end-to-end-audio-strikes-back-boosting","paper_url":"https://arxiv.org/abs/2204.11479v5","paper_title":"End-to-End Audio Strikes Back: Boosting Augmentations Towards An Efficient Audio Classification Network","code":"https://github.com/Alibaba-MIIL/AudioClassfication","n_code_links":1,"syntology":null},{"rank_in_archive_order":20,"model":"SepTr + LeRaC","metrics":{"Accuracy (5-fold)":"91.58","Top-1 Accuracy":"91.58"},"uses_additional_data":false,"paper_date":"2022-05-18","paper":"/paper/lerac-learning-rate-curriculum","paper_url":"https://arxiv.org/abs/2205.09180v4","paper_title":"Learning Rate Curriculum","code":"https://github.com/croitorualin/lerac","n_code_links":1,"syntology":null},{"rank_in_archive_order":21,"model":"SepTr","metrics":{"PRE-TRAINING DATASET":"-","Top-1 Accuracy":"91.13"},"uses_additional_data":false,"paper_date":"2022-03-17","paper":"/paper/septr-separable-transformer-for-audio","paper_url":"https://arxiv.org/abs/2203.09581v3","paper_title":"SepTr: Separable Transformer for Audio Spectrogram Processing","code":"https://github.com/ristea/septr","n_code_links":1,"syntology":null},{"rank_in_archive_order":22,"model":"Multi-Format Contrastive","metrics":{"Top-1 Accuracy":"90.5"},"uses_additional_data":true,"paper_date":"2021-03-11","paper":"/paper/multi-format-contrastive-learning-of-audio","paper_url":"https://arxiv.org/abs/2103.06508v3","paper_title":"Multi-Format Contrastive Learning of Audio Representations","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":23,"model":"Multi-Channel Audio Feature with CNN","metrics":{"Accuracy (5-fold)":"89.5","PRE-TRAINING DATASET":"EfficientNet","Top-1 Accuracy":"89.5"},"uses_additional_data":false,"paper_date":null,"paper":null,"paper_url":null,"paper_title":"","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":24,"model":"AVID","metrics":{"Top-1 Accuracy":"89.2"},"uses_additional_data":false,"paper_date":"2020-04-27","paper":"/paper/audio-visual-instance-discrimination-with","paper_url":"https://arxiv.org/abs/2004.12943v3","paper_title":"Audio-Visual Instance Discrimination with Cross-Modal Agreement","code":"https://github.com/facebookresearch/AVID-CMA","n_code_links":1,"syntology":null},{"rank_in_archive_order":25,"model":"ACDNet","metrics":{"Accuracy (5-fold)":"87.1","Top-1 Accuracy":"87.1"},"uses_additional_data":false,"paper_date":"2021-03-05","paper":"/paper/environmental-sound-classification-on-the","paper_url":"https://arxiv.org/abs/2103.03483v4","paper_title":"Environmental Sound Classification on the Edge: A Pipeline for Deep Acoustic Networks on Extremely Resource-Constrained Devices","code":"https://github.com/mohaimenz/acdnet","n_code_links":1,"syntology":null},{"rank_in_archive_order":26,"model":"XDC","metrics":{"PRE-TRAINING DATASET":"IG-Random","Top-1 Accuracy":"85.4"},"uses_additional_data":false,"paper_date":"2019-11-28","paper":"/paper/self-supervised-learning-by-cross-modal-audio","paper_url":"https://arxiv.org/abs/1911.12667v3","paper_title":"Self-Supervised Learning by Cross-Modal Audio-Video Clustering","code":"https://github.com/HumamAlwassel/XDC","n_code_links":1,"syntology":{"n_ran":0,"n_unverified":1,"n_samples":1,"n_pointer_only_licence":0}},{"rank_in_archive_order":27,"model":"XDC","metrics":{"PRE-TRAINING DATASET":"AudioSet","Top-1 Accuracy":"84.8"},"uses_additional_data":false,"paper_date":"2019-11-28","paper":"/paper/self-supervised-learning-by-cross-modal-audio","paper_url":"https://arxiv.org/abs/1911.12667v3","paper_title":"Self-Supervised Learning by Cross-Modal Audio-Video Clustering","code":"https://github.com/HumamAlwassel/XDC","n_code_links":1,"syntology":{"n_ran":0,"n_unverified":1,"n_samples":1,"n_pointer_only_licence":0}},{"rank_in_archive_order":28,"model":"AVTS","metrics":{"Top-1 Accuracy":"82.3"},"uses_additional_data":false,"paper_date":"2018-06-30","paper":"/paper/cooperative-learning-of-audio-and-video","paper_url":"http://arxiv.org/abs/1807.00230v2","paper_title":"Cooperative Learning of Audio and Video Models from Self-Supervised Synchronization","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":29,"model":"L3","metrics":{"Top-1 Accuracy":"79.3"},"uses_additional_data":false,"paper_date":"2017-05-23","paper":"/paper/look-listen-and-learn","paper_url":"http://arxiv.org/abs/1705.08168v2","paper_title":"Look, Listen and Learn","code":"https://github.com/marl/l3embedding","n_code_links":1,"syntology":null}],"since_archive":{"claim":"Results that newer papers report for their own method, placed here by Syntology. A model pointed at the cell in the paper's own table; the number was read from that cell and checked against this leaderboard's metric, dataset, split and scale; an independent check that saw this leaderboard's other rows and every other leaderboard on the same dataset accepted it. Not reviewed by the paper's authors or by the archive's editors, and not ranked against the archive rows.","extraction_file_present":true,"measurement":{"test_papers":883,"papers_with_output":881,"judged_true":108,"judged":110,"wilson95_lower":0.9361,"measured_on":"2026-09-24","frozen_commit":"0e3de0df94"},"measurement_note":"blind adjudication of accepted entries on a held-out split of archive papers, rules frozen before the test","coverage":{"sentence":"Syntology has checked 6,264 of the 9,581 papers on this site that are newer than the archive; results from the others appear after they are checked.","complete":false,"papers_newer_than_archive":9581,"papers_checked":6264,"papers_extracted_not_yet_verified":0,"boards_without_verdict":2,"papers_not_yet_extracted":3316},"order":"newest first by month (arXiv date, else the arXiv-id month), then arXiv id descending","columns":[],"entries":[]},"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":9,"rows_with_any_sample_ran":6,"distinct_papers_with_graph_line":7,"distinct_papers_with_any_sample_ran":5,"samples_over_distinct_papers":{"n_ran":28,"n_unverified":29,"n_samples":57,"n_pointer_only_licence":7,"note":"each paper (arXiv id) counted once, however many rows it is behind; this is the page-level figure"},"samples_row_weighted":{"n_ran":35,"n_unverified":30,"n_samples":65,"n_pointer_only_licence":14,"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"}}}